45
Coordination Chemistry Reviews 295 (2015) 1–45 Contents lists available at ScienceDirect Coordination Chemistry Reviews j ourna l h om epage: www.elsevier.com/locate/ccr Review Interpretation of europium(III) spectra Koen Binnemans KU Leuven, Department of Chemistry, Celestijnenlaan 200F, P.O. Box 2404, B-3001 Heverlee, Belgium Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2. Energy level structure of the [Xe]4f 6 configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 3. Luminescence spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 3.1. General features and selection rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 3.2. Transition 5 D 0 7 F 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.3. Transition 5 D 0 7 F 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.4. Transition 5 D 0 7 F 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.5. Transition 5 D 0 7 F 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.6. Transition 5 D 0 7 F 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.7. Transitions 5 D 0 7 F 5 and 5 D 0 7 F 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.8. Emission from higher excited states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.9. Polarized emission spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3.10. Sensitized luminescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 4. Absorption spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 4.1. General features and “hot” bands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 4.2. Transitions within the 7 F ground term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 4.3. Transition to the 5 D 0 level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 4.4. Transitions to the 5 D 1 level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.5. Transitions to the 5 D 2 level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.6. Transitions to higher energy levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.7. Charge-transfer bands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 5. Excitation spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 6. Other spectroscopic techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 6.1. Two-photon absorption (TPA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 6.2. Zeeman spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 6.3. Magnetic circular dichroism (MCD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 7. Eu 3+ as a spectroscopic probe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 8. Nephelauxetic effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 9. Judd–Ofelt parameterization of europium(III) spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 9.1. Determination of Judd–Ofelt intensity parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 9.2. Use of Judd–Ofelt parameters for calculation of photophysical quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 9.3. Hypersensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Abbreviations: A, probability for radiative decay (=Einstein coefficient A); acac, acetylacetonate; bipy, 2,2 -bipyridine; bmpyr, N-butyl-N-methylpyrrolidinium; CD, circular dichroism; C4mim, 1-butyl-3-methylimidazolium; C6mim, 1-hexyl-3-methylimidazolium; CNL , maximum ligand coordination number; CT, charge transfer; D, dipole strength; DED, dipole strength of an electric dipole transition; DMD, dipole strength of a magnetic dipole transition; dbm, dibenzoylmethanate; dmbipy, 4,4 -dimethyl-2,2 - bipyridine; DOTA, 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetate; DPA, 2,6-pyridinedicarboxylate (=dipicolinate); E eff , effective field; ED, (induced) electric dipole; EDTA, ethylenediaminetetraacetate; EPR, electron paramagnetic resonance; EXAFS, extended X-ray absorption fine structure; f, oscillator strength; H, Hamiltonian; I, intensity; I, nuclear spin; J, total angular momentum quantum number; L, total orbital angular momentum quantum number; LCP, left circularly polarized light; LMCT, ligand-to-metal charge transfer; MCD, magnetic circular dichroism; MCPE, magnetic circularly polarized emission; MD, magnetic dipole; MRI, magnetic resonance imaging; n, refractive index; NTA, nitrilotriacetate; ODA, oxydiacetate; p, formal charge; PCEM, point charge electrostatic model; q, hydration number; RCP, right circularly polarized light; S, total spin quantum number; S, singlet; T, triplet; terpy, 2,2 ;6 ,2 -terpyridine (=terpyridine); Tf2N , bis(trifluoromethylsulfonyl)imide (=bistriflimide); TMU, tetramethylurea; Tp, hydrotris(1-pyrazolyl)borate; TPA, two-photon absorption; TPP, tricapped trigonal prism; TTHA, triethylenetetraaminehexaacetate; W, probability for non-radiative decay; X A (T), fractional thermal population at temperature T; ˇR, branching ratio; opt (X), optical electronegativity of the ligand; uncorr (M), uncorrected optical electronegativity of the metal; sens , sensitization efficiency; , wavelength; ˜ v, wavenumber; , lifetime; obs , observed luminescence lifetime; rad , radiative lifetime; ˚, quantum yield; ˚ L Ln , overall quantum yield; ˚ Ln Ln , intrinsic quantum yield; ˝ , Judd–Ofelt intensity parameter. Tel.: +32 16 32 7446. E-mail address: [email protected] http://dx.doi.org/10.1016/j.ccr.2015.02.015 0010-8545/© 2015 Elsevier B.V. All rights reserved.

Coordination Chemistry Reviews - SOLVOMET · Coordination Chemistry Reviews 295 (2015) 1–45 Contents lists available at ScienceDirect Coordination Chemistry Reviews j ournal homepage:

  • Upload
    lydiep

  • View
    254

  • Download
    2

Embed Size (px)

Citation preview

  • R

    I

    KK

    C

    csbEIcishXoo

    h0

    Coordination Chemistry Reviews 295 (2015) 145

    Contents lists available at ScienceDirect

    Coordination Chemistry Reviews

    j ourna l h om epage: www.elsev ier .com/ locate /ccr

    eview

    nterpretation of europium(III) spectra

    oen Binnemans

    U Leuven, Department of Chemistry, Celestijnenlaan 200F, P.O. Box 2404, B-3001 Heverlee, Belgium

    ontents

    1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22. Energy level structure of the [Xe]4f6 configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33. Luminescence spectra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

    3.1. General features and selection rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.2. Transition 5D0 7F0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.3. Transition 5D0 7F1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103.4. Transition 5D0 7F2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.5. Transition 5D0 7F3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133.6. Transition 5D0 7F4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133.7. Transitions 5D0 7F5 and 5D0 7F6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.8. Emission from higher excited states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.9. Polarized emission spectra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153.10. Sensitized luminescence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

    4. Absorption spectra. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174.1. General features and hot bands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174.2. Transitions within the 7F ground term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194.3. Transition to the 5D0 level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194.4. Transitions to the 5D1 level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204.5. Transitions to the 5D2 level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204.6. Transitions to higher energy levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204.7. Charge-transfer bands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

    5. Excitation spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226. Other spectroscopic techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

    6.1. Two-photon absorption (TPA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236.2. Zeeman spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246.3. Magnetic circular dichroism (MCD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

    7. Eu3+ as a spectroscopic probe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258. Nephelauxetic effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299. JuddOfelt parameterization of europium(III) spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

    9.1. Determination of JuddOfelt intensity parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309.2. Use of JuddOfelt parameters for calculation of photophysical quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319.3. Hypersensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

    Abbreviations: A, probability for radiative decay (=Einstein coefficient A); acac, acetylacetonate; bipy, 2,2-bipyridine; bmpyr, N-butyl-N-methylpyrrolidinium; CD,ircular dichroism; C4mim, 1-butyl-3-methylimidazolium; C6mim, 1-hexyl-3-methylimidazolium; CNL , maximum ligand coordination number; CT, charge transfer; D, dipoletrength; DED , dipole strength of an electric dipole transition; DMD , dipole strength of a magnetic dipole transition; dbm, dibenzoylmethanate; dmbipy, 4,4-dimethyl-2,2-ipyridine; DOTA, 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetate; DPA, 2,6-pyridinedicarboxylate (=dipicolinate); Eeff , effective field; ED, (induced) electric dipole;DTA, ethylenediaminetetraacetate; EPR, electron paramagnetic resonance; EXAFS, extended X-ray absorption fine structure; f, oscillator strength; H, Hamiltonian; I, intensity;, nuclear spin; J, total angular momentum quantum number; L, total orbital angular momentum quantum number; LCP, left circularly polarized light; LMCT, ligand-to-metalharge transfer; MCD, magnetic circular dichroism; MCPE, magnetic circularly polarized emission; MD, magnetic dipole; MRI, magnetic resonance imaging; n, refractivendex; NTA, nitrilotriacetate; ODA, oxydiacetate; p, formal charge; PCEM, point charge electrostatic model; q, hydration number; RCP, right circularly polarized light; S, totalpin quantum number; S, singlet; T, triplet; terpy, 2,2 ;6 ,2-terpyridine (=terpyridine); Tf2N , bis(trifluoromethylsulfonyl)imide (=bistriflimide); TMU, tetramethylurea; Tp,ydrotris(1-pyrazolyl)borate; TPA, two-photon absorption; TPP, tricapped trigonal prism; TTHA, triethylenetetraaminehexaacetate; W, probability for non-radiative decay;A(T), fractional thermal population at temperature T; R , branching ratio; opt(X), optical electronegativity of the ligand; uncorr(M), uncorrected optical electronegativityf the metal; sens , sensitization efficiency; , wavelength; v, wavenumber; , lifetime; obs , observed luminescence lifetime; rad , radiative lifetime; , quantum yield; LLn ,verall quantum yield; Ln

    Ln, intrinsic quantum yield; , JuddOfelt intensity parameter.

    Tel.: +32 16 32 7446.E-mail address: [email protected]

    ttp://dx.doi.org/10.1016/j.ccr.2015.02.015010-8545/ 2015 Elsevier B.V. All rights reserved.

    dx.doi.org/10.1016/j.ccr.2015.02.015http://www.sciencedirect.com/science/journal/00108545http://www.elsevier.com/locate/ccrhttp://crossmark.crossref.org/dialog/?doi=10.1016/j.ccr.2015.02.015&domain=pdfmailto:[email protected]/10.1016/j.ccr.2015.02.015

  • 2

    a

    ARAA

    KELLLRS

    1

    ptccatwtst[aepectTt((nmlic[etb

    itfsfi[

    K. Binnemans / Coordination Chemistry Reviews 295 (2015) 145

    10. Dynamics of excited states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3310.1. Decay processes and lifetimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3310.2. Determination of hydration numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

    11. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

    r t i c l e i n f o

    rticle history:eceived 30 December 2014ccepted 16 February 2015vailable online 23 February 2015

    eywords:uropiumanthanidesuminescenceuminescent materialsare earthspectroscopy

    a b s t r a c t

    The trivalent europium ion (Eu3+) is well known for its strong luminescence in the red spectral region,but this ion is also interesting from a theoretical point of view. Due to the even number of electrons inthe 4f shell (4f6 configuration), the crystal-field perturbation by the crystalline host matrix lifts partlyor completely the degeneracies of the 2S+1LJ levels. The Eu3+ ion has the great advantage over otherlanthanide ions with an even number of 4f electrons that the starting levels of the transitions in boththe absorption and the luminescence spectrum are non-degenerate (J = 0). Moreover, the interpretationof the spectra is facilitated by the small total angular momentum J of the end levels in the transitions.The number of lines observed for the 5D0 7FJ transitions in the luminescence spectrum or the 5DJ 7F0transitions in the absorption spectrum allows determining the site symmetry of the Eu3+ ion. This reviewdescribes the spectroscopic properties of the trivalent europium ion, with emphasis on the energy levelstructure, the intensities of the ff transitions (including the JuddOfelt theory), the decay times of theexcited states and the use of the Eu3+ ion as a spectroscopic probe for site symmetry determination. Itis illustrated how the maximum amount of information can be extracted from optical absorption andluminescence spectra of europium(III) compounds, and how pitfalls in the interpretation of these spectracan be avoided.

    2015 Elsevier B.V. All rights reserved.

    . Introduction

    The trivalent europium ion (Eu3+) exhibits an intense redhotoluminescence upon irradiation with UV radiation. This pho-oluminescence is observed not only for Eu3+ ions doped intorystalline host matrices or glasses, but also for europium(III)omplexes with organic ligands. These ligands can act as anntenna to absorb the excitation light and to transfer the exci-ation energy to the higher energy levels of the Eu3+ ion, fromhich the emitting excited levels can be populated. The pho-

    oluminescence of europium(III) complexes has been studied inolutions [1,2], polymer matrices [3,4], solgel glasses [5,6], func-ionalized solgel glasses [711], ionogels [12,13], liquid crystals1416], encapsulated into inorganic hosts such as zeolites [1720]nd in metalorganic frameworks (MOFs) [21,22]. The design ofuropium(III)-containing inorganicorganic hybrid materials is aopular research field [2326]. Europium(III) complexes can bexcellent luminescent probes for biochemical or biomedical appli-ations [2733]. The most important application of europium ishe red phosphor Y2O3:Eu3+ (YOX) in fluorescent lamps [3436].he red emission of Eu3+ can be achieved not only by excita-ion with UV light, but also by irradiation with an electron beamcathodoluminescence) [37,38], X-rays, -rays, - or -particlesradioluminescence) [3942], strong electric fields (electrolumi-escence) [43,44], mechanical agitation (triboluminescence orechanoluminescence) [4547] or by chemical reactions (chemi-

    uminescence) [48]. A well-known cathodoluminescent phosphors Y2O2S:Eu3+, which is the red phosphor used in the old-fashionedathode-ray tubes of color television screens or computer monitors37,38,49]. This compound replaced the older cathodoluminescenturopium(III) phosphor YVO4:Eu3+ [50,51]. It is worth mentioninghat europium is present in the anti-counterfeiting ink of EUROanknotes [52].

    Not only its red luminescence, but also the narrow transitionsn the absorption and luminescence spectra are typical features ofhe Eu3+ ion, and these spectroscopic properties have been known

    Prandtl was the first to publish a picture of an absorption spectrumof Eu3+ [55]. In 1906, Urbain reported on the red luminescence ofEu2O3 diluted in lime [56,57]. However, two years earlier in 1904,Urbain had already noticed that crystals of europium(III) sulfateoctahydrate, Eu2(SO4)38H2O, had a faint pink color [58], but he didnot realize that this color was caused by the photoluminescence ofEu3+ ions excited by the UV part of sunlight [59]. In the absence ofthis luminescence, europium(III) compounds are colorless. In 1909,Urbain described the cathodoluminescence of Gd2O3:Eu3+ [60].

    The fine structure and the relative intensities of the transi-tions in the absorption and luminescence spectra of Eu3+ can beused to probe the local environment of the Eu3+ ion. The spec-troscopic data give information on the point group symmetry ofthe Eu3+ site and sometimes also information on the coordinationpolyhedron. However, a rigorous interpretation of europium(III)spectra can be a daunting task for newcomers in the field of lan-thanide coordination chemistry. Chemists who have been trainedin the synthesis and characterization of luminescent lanthanidecomplexes are often lacking a sound theoretical background inlanthanide spectroscopy. The classical books or reviews on spec-troscopy of rare earths are often too theoretical or put littleemphasis on the relationship between features observed in spectraand structural properties [6170]. There exist several reviews onthe luminescence of lanthanide-based molecular materials or pho-tophysics of lanthanides, but only few of them focus on the Eu3+

    ion in detail [23,24,27,28,33,64,71100]. In general, these worksdo not give a detailed description of the transitions in europium(III)spectra. As a consequence, many authors who describe luminescenteuropium(III) complexes do not go beyond reporting general state-ments with little information content, such as mentioning that thetransitions observed in the luminescence spectra are the 5D0 7FJ(J = 06) transitions or that a very intense hypersensitive transition5D0 7F2 indicates that the Eu3+ is not at a site with a center ofsymmetry.

    The aim of this review is to give a sound introduction to the spec-troscopic properties of the trivalent europium ion, with emphasis

    rom the earliest history of the chemical element europium. Theharp lines in the absorption spectra of Eu3+ in solution wererst described in 1900 by Demarc ay, the discoverer of europium53], and his observations were confirmed by Prandtl in 1920 [54].

    on the energy level structure, the intensities of the ff transitions,

    the decay times of the excited states and the use of the Eu3+ ionas a spectroscopic probe for site symmetry determination. It isshown how the maximum amount of information can be extracted

  • K. Binnemans / Coordination Chemistry Reviews 295 (2015) 145 3

    N-OOC COO-N NN N

    O

    H3C CH3O-

    O O-O

    F3C

    O-S

    N NCOO-

    COO-

    -OOC

    -OOC N

    N

    N

    NCOO-

    COO-

    -OOC

    -OOC

    -OOC O COO-

    NN N

    NNO

    CH3

    CH3

    N+

    CH3

    O-

    attcaca dbmODA

    bpy phe n terpy DPA

    EDTA DOTA antipyrene 4-picNO

    F acac, aO 2-tert , 2,3-

    fctEitmsm

    2

    acioidEi(

    Heotcetcmmieee1i

    singlet, doublet, triplet, quartet, quintet, sextet, septet for 2S + 1 = 1, 2,3, 4, 5, 6, 7, respectively. The term with the highest spin multiplic-ity for the 4f6 configuration is a septet, which corresponds to six

    ig. 1. Selection of ligands of luminescent europium(III) complexes. Abbreviations: DA, oxydiacetate; bpy, 2,2-bipyridine; phen, 1,10-phenanthroline; terpy, 2,2;6 ,

    raacetate; DOTA, 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetate; antipyrene

    rom optical absorption and luminescence spectra of europium(III)ompounds, and how pitfalls in the interpretation of these spec-ra can be avoided. In this review, europium(III) is represented asu3+ rather than Eu(III). In principle Eu3+ is the trivalent europiumon in the gas phase. However, it is common practice among spec-roscopists to use the symbol Eu3+ for europium(III)-doped solid

    aterials and even for europium(III) in solutions. Fig. 1 shows aelection of ligands that are found in the europium(III) complexesentioned in this review paper.

    . Energy level structure of the [Xe]4f6 configuration

    Eu3+ has 60 electrons: 54 electrons in the same closed shellss the xenon atom and 6 electrons in the 4f shell. This electroniconfiguration can be written as [Xe]4f6, or 4f6 for short. The 4f shells well shielded from its environment by the closed 5s2 and 5p6

    uter shells [101]. The six electrons in the 4f shell can be arrangedn 3003 different ways into the seven 4f orbitals, so that the totalegeneracy of the [Xe]4f6 electronic configuration of the trivalentu3+ ion is 3003. The degeneracy of a 4fn electronic configurations given by the binomial coefficient:

    14

    n

    )= 14!

    n!(14 n)! (1)

    ere n is the number of 4f electrons (n = 6 for Eu3+). Each differ-nt electronic arrangement is called a microstate. The degeneracyf the 4f6 configuration is partly or totally lifted by several per-urbations acting on the Eu3+ ion: electron repulsion, spinorbitoupling, the crystal-field perturbation and eventually the Zeemanffect (Fig. 2). The electron repulsion is the electrostatic interac-ion between the different electrons in the 4f shell. The spinorbitoupling results from the interaction between the spin magneticoment of the electron and the magnetic field created by the move-ent of the electron around the nucleus. The crystal-field effect

    s caused by the interactions between the 4f electrons and thelectrons of the ligands. The Zeeman effect is the splitting of the

    nergy levels by an external magnetic field. After introduction oflectron repulsion, the [Xe]4f6 configuration is characterized by19 2S+1L() terms (Table 1) [102]. The degeneracy of each term

    s (2S + 1)(2L + 1). S is the total spin quantum number and L is the

    cetylacetonate; tta, 2-thenoyltrifluoroacetylacetonate; dbm, dibenzoylmethanate;pyridine; DPA, 2-pyridinedicarboxylate (=dipicolinate); EDTA, ethylenediaminete-dimethyl-1-phenyl-3-pyrazolin-5-one; 4-PicNO, 4-picoline-N-oxide.

    total orbital angular momentum quantum number. is an additionalquantum number to differentiate between terms with identical Sand L quantum numbers [103]. Terms are denoted by capital lettersof the Latin alphabet: S (L = 0), P (L = 1), D (L = 2), F (L = 3), G (L = 4), H(L = 5), I (L = 6), K (L = 7), L (L = 8), M (L = 9),. . . Notice that the letterJ is not used as a term label. The term with the highest L value ofthe 4f6 configuration has L = 12, giving a 1Q term. 2S + 1 is the spinmultiplicity of the term. The nomenclature for spin multiplicity is

    Fig. 2. Partial energy diagram of Eu3+ (4f6) showing the relative magnitude ofthe interelectronic repulsion (terms), spinorbit coupling (levels) and crystal-fieldeffects (sublevels). The downward arrows indicate the excited states 5D0 and 5D1from which luminescence occurs.

    Reprinted with permission from reference [105]. Copyright 1987 Elsevier.

  • 4 K. Binnemans / Coordination Chemis

    Table 1The 119 2S+1L() terms of the 4f6 configuration of Eu3+ [102].

    7F 3K(6) 3F(8) 1I(6)5L 3I(1) 3F(9) 1I(7)5K 3I(2) 3D(1) 1H(1)5I(1) 3I(3) 3D(2) 1H(2)5I(2) 3I(4) 3D(3) 1H(3)5H(1) 3I(5) 3D(4) 1H(4)5H(2) 3I(6) 3D(5) 1G(1)5G(1) 3H(1) 3P(1) 1G(2)5G(2) 3H(2) 3P(2) 1G(3)5G(3) 3H(3) 3P(3) 1G(4)5F(1) 3H(4) 3P(4) 1G(5)5F(2) 3H(5) 3P(5) 1G(6)5D(1) 3H(6) 3P(6) 1G(7)5D(2) 3H(7) 1Q 1G(8)5D(3) 3H(8) 1N(1) 1F(1)5P 3H(9) 1N(2) 1F(2)5S 3G(1) 1M(1) 1F(3)3O 3G(2) 1M(2) 1F(4)3N 3G(3) 1L(1) 1D(1)3M(1) 3G(4) 1L(2) 1D(2)3M(2) 3G(5) 1L(3) 1D(3)3M(3) 3G(6) 1L(4) 1D(4)3L(1) 3G(7) 1K(1) 1D(5)3L(2) 3F(1) 1K(2) 1D(6)3L(3) 3F(2) 1K(3) 1P3K(1) 3F(3) 1I(1) 1S(1)3K(2) 3F(4) 1I(2) 1S(2)3K(3) 3F(5) 1I(3) 1S(3)

    uvscoiEt2

    2

    t

    sqaLJlfpi7

    cdlsl5

    tott(clct

    F4 and F6 parameters can be expressed approximately in func-

    3K(4) 3F(6) 1I(4) 1S(4)3K(5) 3F(7) 1I(5)

    npaired electrons: S = [ + + + + + ] = 3 or 2S + 1 = 7. The Lalue of this septet is 3 (or an F term), which corresponds to theum of the ml values: L = [(+3) + (+2) + (+1) + 0 + (1) + (2)] = 3. Foronfigurations with an even number of electrons, all terms havedd multiplicity. Only singlets, triplets, quintets and septets occurn the 4f6 configuration with six electrons, such as in the case ofu3+. The 2S+1L() terms can be rigorously classified by the groupheoretical labels introduced by Racah [104], but in practice theS+1L() labels are preferred. The separation between the differentS+1L() terms is of the order of 10,000 cm1 for the lower terms ofhe 4f6 configuration.

    The 2S+1L() terms of the [Xe]4f6 configuration are split by thepinorbit interaction in 295 2S+1L()J levels. J is the total angularuantum number and it indicates the relative orientation of the spinnd the orbital momenta. The possible values for J are L + S, L + S-1,

    + S-2, . . ., |L-S|. For the 7F term, L = 3 and S = 3, so that the possible values are: 6, 5, 4, 3, 2, 1, 0. The degeneracy of each spinorbitevel is 2J + 1. The quantum number is often omitted, so that theree-ion levels are labeled as 2S+1LJ. In the RussellSaunders cou-ling scheme (also called LS coupling scheme), each free-ion levels characterized by a 2S+1LJ label. For Eu3+, only the J levels of theF and 5D terms are adequately described by the RussellSaundersoupling scheme. A better description of the free-ion levels can beone by applying the intermediate coupling scheme, in which each

    evel is a linear combination of different 2S+1LJ states, but with theame J quantum level. For example, the wave functions of the 7F0evel in the intermediate coupling scheme is: 0.9680 7F0 + 0.0016D(2)0 + 0.1659 5D(3)0 0.1815 5D(1)0 [106]. The splitting of theerms into J states by the spinorbit coupling interaction is of therder of 1000 cm1. The 2J + 1 degeneracy of the energy levels inhe free ion is further lifted by the crystal-field effect, after whichhe energy levels are characterized by the irreducible representationirreps) of the point group of the Eu3+ site [95]. These levels arealled crystal-field levels (or Stark levels). The splitting of the energy

    evels by the crystal-field effect is of the order of a few hundredm1 or less. In systems with an orthorhombic or lower symme-ry, all degeneracy is lifted by the crystal field. In systems with a

    try Reviews 295 (2015) 145

    higher symmetry, all degeneracy can be lifted by an external mag-netic field, via the so-called Zeeman effect. Even in strong magneticfields, the splitting of the energy levels by the Zeeman effect is onlya few cm1. The J quantum numbers are well defined in the freeEu3+ ion, but J-mixing occurs when the Eu3+ is located in a non-spherically symmetric ligand environment (vide infra) [107,108].J-mixing is induced by the even-parity components of the crystal-field potential.

    Hunds rules explain why 7F0 is the ground state of the 4f6 elec-tronic configuration: Rule 1: the spin multiplicity has to be as largeas possible; Rule 2: in case there is more than one term with thesame spin multiplicity, the term with the highest total orbital angu-lar momentum (or L value) is the ground state; Rule 3: For electronicshells that are less than half-filled, the ground state has the lowestpossible J value. For electronic shells that are more than half-filled,the ground state has the highest possible J value. Since the high-est multiplicity of the terms of the 4f6 electronic configuration is aseptet and since there is only one septet, 7F is the ground term. The4f6 shell is less than half filled and, as explained above, the possibleJ values for the 7F term are 0, 1, 2, 3, 4, 5, 6, so that the ground stateof Eu3+ is 7F0. The order of energies of the levels within the 7F termis therefore: 7F0 < 7F1 < < 7F6. However, the relative positions ofthe energy levels of the excited states can be determined only bycalculations.

    The energy levels and wave functions of the Eu3+ ion can beobtained by diagonalization of the energy matrix [95]. The matrixelements are of the type lnSLJM |H|ln SLJM , where H is theeffective-operator Hamiltonian, and lnSLJM and ln SLJM are

    basis functions of the 4fn configuration (n = 6 for Eu3+). The angularparts of the matrix elements can be calculated exactly, whereas theradial parts are treated as adjustable parameters. A parameter set isobtained by optimizing a start set of parameters by a general least-squares fitting process in which the energy differences between thecalculated and experimental energy levels are minimized. The bestknown fitting programs are those written by Crosswhite [109], andby Reid [110]. The total Hamiltonian can be written as the sum ofa free-ion and a crystal-field part:

    H = Hfree ion + Hcrystal field (2)The free-ion Hamiltonian is characterized by a set of three electronrepulsion parameters (F2, F4, F6), by the spinorbit coupling con-stant 4f, the Trees configuration interaction parameters (, , ),the three-body configuration interaction parameters (T2, T3, T4, T6,T7, T8) and parameters which describe magnetic interactions (M0,M2, M4, P2, P4, P6). An additional parameter Eave (ave stands foraverage) takes into account the kinetic energy of the electronsand their interaction with the nucleus. It only shifts the barycen-ter of the whole 4f6 configuration. The free-ion Hamiltonian can bewritten as [109,111]:

    Hfree ion = Eave +

    k=2,4,6Fkfk + 4fASO + L(L + 1) + G(G2) + G(R7)

    +

    i=2,3,4,6,7,8Titi +

    j=0,2,4

    Mlml +

    k=2,4,6Pkpk (3)

    Here fk and ASO represent the angular part of the electrostatic andspinorbit interaction, respectively. L is the total orbital angularmomentum. G(G2) and G(R7) are the so-called Casimir operatorsfor the groups G2 and R7, respectively. The ti are the three-particleoperators. The ml and pk represent the operators for the mag-netic corrections. The Fk parameters decrease if k increases. The

    tion of F2: F4/F2 = 0.668 and F6/F2 = 0.495 [112]. These ratios arethose of the hydrogenic wave functions and are applied if thenumber of experimental data is insufficient to vary the three

  • hemistry Reviews 295 (2015) 145 5

    erftFfpoctsirieriutt1astetFHttosIoicattHMi

    eStotopa(c4

    ptfi

    H

    Htat

    Table 2Average free-ion parameters for Eu3+ [95].

    Parameter Value (cm1)

    EAVE 63,736F2 82,786F4 59,401F6 42,644 19.80 617 1460T2 370T3 40T4 40T6 330T7 380T8 3704f 1332M0 2.38M2 1.33M4 0.90P2 303P4 227P6 152

    Table 3Calculated energies of free-ion levels for Eu3+ between 0 and 40,000 cm1, calculatedwith the parameters listed in Table 2 [117].

    2S+1LJ Ecalc (cm1)

    7F0 07F1 3797F2 10437F3 18967F4 28697F5 39127F6 49925D0 17,2275D1 18,9735D2 21,4455D3 24,3355L6 25,1255L7 26,1775G2 26,2695G3 26,4935G4 26,6115G5, 5G6 26,6425L8 27,0955D4 27,5835L9 27,8445L10 28,3415H3 30,8705H7 31,0705H4 31,2925H6, 5H5 31,5113P0 32,7905F2 33,0555F3 33,0925F1 33,3665F4 33,5135F5 34,0405I4 34,0575I5 34,3885I6 34,9665I7 35,4295I8 35,4535K5 36,1685K6 37,3203P1 38,1325K7 38,2475G2 38,6165K8 38,6673K6, 3I6 38,780

    K. Binnemans / Coordination C

    lectrostatic parameters independently. For instance, if data areestricted to the energy levels of the 7F and 5D terms are availableor Eu3+, only one parameter can be varied. This does not implyhat the f orbitals are hydrogenic, but that the ratios F4/F2 and6/F2 are rather insensitive to the exact composition of the waveunctions [61]. Although electron repulsion and spinorbit cou-ling can explain the free-ion level structure in a qualitative way,ther minor interactions have to be taken into account for detailedalculations of the free-ion energy levels. These weak interac-ions include configuration interactions, electrostatic correlatedpinorbit interaction, spinspin, spinother-orbit and relativisticnteractions. Diagonalization of the energy matrix which incorpo-ates only the electrostatic and spinorbit interaction, often resultsn discrepancies between experimental and calculated levels of sev-ral hundred cm1 [61]. Additional parameters and operators areequired to describe the configuration interaction. Configurationnteraction is the spin-independent interaction between config-rations of equal parity. The new operators are two-particle andhree-particle operators working within the 4fn configuration. Thewo-particle correction term in the free-ion Hamiltonian is L(L +) + G(G2) + G(R7) [113]. The values of the parameters , and re rather constant across the lanthanide series, because processesuch as excitation of one or two particles to the high energy con-inuum states have large contributions to the parameters and thenergies of these continuum states relative to the 4fn configura-ions do not change significantly with the atomic number Z [111].or 4fn configurations with three or more f electrons, the free-ionamiltonian is expanded with the term

    i=2,3,4,6,7,8T

    iti to take thehree-particle configuration interaction into account [114]. Noticehat t5 and the corresponding parameter T5 do not exist. Variationf the Ti parameters in a fitting procedure has to be done carefully,ince these parameters are only sensitive to particular 2S+1LJ levels.f the level for which a Ti parameter shows a great sensitivity is notbserved in the spectra, a variation of that Ti parameter will resultn a meaningless parameter value [115]. The parameter has to beonstrained in that case. Magnetically correlated corrections suchs spinspin and spinother-orbit interactions are represented byhe term

    j=0,2,4M

    lmj in the Hamiltonian. In the calculations,hese parameters are mostly maintained by the pseudo-relativisticartreeFock ratios M2/M0 = 0.56 and M4/M0 = 0.38, allowing only0 to vary freely [109]. The electrostatic correlated spinorbit

    nteractions are described by the term

    k=2,4,6Pkpk. The param-

    ters Pk can be varied in the ratios P4/P2 = 0.75, P6/P2 = 0.5 [109].ince the introduction of new parameters may alter the values ofhe parameters already fitted, Judd and Crosswhite have introducedrthogonalized operators [116]. These operators yield parametershat more precisely defined and more stable than the conventionalnes. Grller-Walrand and Binnemans reported a set of free-ionarameters for Eu3+, by averaging different parameter sets that arevailable in the literature for Eu3+ ions doped into single crystalsTable 2) [95], and this set of parameters has been used to cal-ulate the free-ion levels of the 4f6 configuration between 0 and0,000 cm1 (Table 3) [117].

    The terms in the Hamiltonian that represent the non-sphericalart of the interactions with the host matrix are described by usinghe crystal-field Hamiltonian. According to Wybourne, the crystal-eld Hamiltonian can be written as [61,95]:

    crystal field =n

    i=0

    k=0

    kq=k

    BkqCkq (i) (4)

    ere Ckq (i) are tensor operators of rank k, with components q. Theseensor operators transform like the spherical harmonics. The Bkqre the crystal-field parameters, n is the number of electrons (6 inhe case of Eu3+) and i represents the ith electron. For f electrons,

    5G3 39,1435K9 39,5185G4 39,726

  • 6 hemistry Reviews 295 (2015) 145

    ktootgmncs

    H

    H

    Dbisnlotpf

    tc2

    crTratgleipnlapaoom

    elietsa

    3

    3

    blstp

    575 600 62 5 650 675 700 72 5

    7F07F1

    7F2

    7F47F3

    Inte

    nsity

    (a.u

    )

    Wavelength (nm)

    Fig. 3. Luminescence spectrum [Eu(tta)3(phen)] at 77 K. The excitation wavelength

    K. Binnemans / Coordination C

    = 2, 4, 6. The number of non-zero parameters is determined byhe point-group site symmetry of the lanthanide ion. The numberf parameters increases if the site symmetry is lowered. Whereasnly 2 parameters are required to describe the crystal-field split-ing in Oh symmetry, 27 parameters are required in C1 symmetry. Ineneral, the Bkq parameters are complex numbers, but in some sym-etry the imaginary part of the parameters is zero. The increase in

    umber of crystal-field parameters upon a lowering of symmetryan be illustrated by the crystal-field Hamiltonians for D2d and S4ymmetry [118]:

    D2d = B20C20 + B20C40 + B44(C44 + C44 ) + B60C60 + B64(C64 + C64 ) (5)

    S4 = HD2d + iB44(C44 C44 ) + iB64(C64 C64 ) (6)

    2d is the symmetry of an undistorted dodecahedron, which shoulde more correctly be called triangular dodecahedron, to distinguish

    t from the conventional dodecahedron with pentagonal faces. Amall distortion lowers the symmetry from D2d to S4. The determi-ation of a reliable set of crystal-field parameters for sites with a

    ow symmetry is very challenging [119122]. A problem with sitesf a low symmetry is that a large number of parameters is requiredo describe the crystal-field perturbation and that some of thesearameters can take unrealistic values, since they will compensateor wrong values of other parameters [123].

    The crystal-field perturbation destroys the spherical symme-ry of the free-ion and the 2S+1LJ terms split up in a number ofrystal-field levels. The extent to which the 2J + 1 degeneracy of aS+1LJ term is removed depends on the symmetry class (icosahedral,ubic, octagonal, hexagonal, pentagonal, tetragonal, trigonal, ortho-hombic, monoclinic, triclinic) and not on the point itself (Table 4).he splitting pattern of the J levels can be derived from full-otational group compatibility tables. For all point groups within

    symmetry class, the splitting of a J term is identical. For instance,he splitting of the 2S+1LJ terms is the same for all tetrahedralroups (D4h, D4, C4v, C4h, C4, D2d, S4). All the 2J + 1 degeneracy isifted in orthorhombic symmetry, so that a further symmetry low-ring will not result in an additional splitting of the 2S+1LJ termsn more crystal-field levels. The differences between the differentoint groups are reflected in different selection rules or in differentumbers of transitions that are allowed between two 2S+1LJ terms. A

    owering in symmetry results in a relaxation of the selection rulesnd to an increase in the number of allowed transitions. For theoint group C1, no transitions are forbidden by the selection rulesnd transitions are allowed between all the crystal-field sublevelsf two 2S+1LJ terms. It should be noticed that in spectroscopy, notnly the 32 crystallographic point groups are considered, but alsoolecular point groups such as Ih or D4d.The convention to describe a transition between two 2S+1LJ lev-

    ls is to write the high energy state at the left hand side and theow energy state at the right hand side. The arrow points from thenitial to the final state. For instance, the transition from the 5D0xcited state to the 7F1 state in the luminescence spectrum is writ-en as 5D0 7F1. The same convention is used for the absorptionpectra. The transition from the 7F1 state to the 5D0 state is writtens 5D0 7F1 (not 7F1 5D0).

    . Luminescence spectra

    .1. General features and selection rules

    A luminescence spectrum (or emission spectrum) is recordedy fixing the excitation wavelength, while the detection wave-

    ength of the spectrofluorimeter is scanned. The luminescencepectra of europium(III) compounds are more informative thanhe corresponding absorption spectra. Many europium(III) com-ounds show an intense photoluminescence, due to the 5D0 7FJ

    is 396 nm. All the transitions start from the 5D0 state.

    Adapted by permission of The Royal Society of Chemistry from reference [15]. Copy-right 2002 The Royal Society of Chemistry.

    transitions (J = 06) from the 5D0 excited state to the J levels ofthe ground term 7F. An overview of the transitions is given inTable 5. Very often the transitions to the 7F5 and 7F6 levels arenot observed, because they are outside the wavelength range ofthe detectors of spectrofluorimeters (vide infra). In Fig. 3, theluminescence spectrum of the europium -diketonate complex[Eu(tta)3(phen)] is shown (tta = 2-thenoyltrifluoroacetylacetonate,phen = 1,10-phenanthroline). Transitions from higher excitedstates (5D1, 5D2, 5D3) are much less common (see Section 3.8).

    An observation that can be made from the inspection of thepositions of the different 5D0 7FJ transitions is that the distancebetween a J and the J + 1 line increases with increasing J value,i.e. the 5D0 7F1 transition is very close to the 5D0 7F0 transi-tion, but the 5D0 7F6 transition is lying more than 50 nm furtherto the infrared than the 5D0 7F5 transition [124]. This behaviorcan be explained by the fact that the splitting of the 7FJ multi-plet corresponds quite well to the Land interval rule: the intervalbetween successive energy levels is proportional to the larger oftheir total angular momentum values J (i.e. the splitting increaseswith increasing J values). The majority of the transitions observedin the luminescence spectrum are induced electric dipole transitions(ED transitions). An electric dipole transition is the consequenceof the interaction of the lanthanide ion with the electric field vec-tor through an electric dipole. The creation of an electric dipolesupposes a linear movement of charge. Such a transition has oddparity. Therefore, the electric dipole operator has odd transforma-tion properties under inversion with respect to an inversion center.Intraconfigurational electric dipole transitions (e.g. ss, pp, dd,or ff transitions) are forbidden by the Laporte selection rule. TheLaporte selection rule strictly applies to a lanthanide ion in the gasphase (i.e., a centrosymmetric environment); however, it is relaxedfor lanthanide ions embedded in a medium, since the transitionscan be partly allowed by vibronic coupling or via mixing of higherconfigurations into the 4f wavefunctions by the crystal-field effect.The observed transitions are much weaker than ordinary electricdipole transitions. Therefore, they are often called induced electricdipole transitions (or forced electric dipole transitions), rather thanjust electric dipole transitions. The intensities of the ED transitionscan be described by the JuddOfelt theory (JO-theory; see Section9) [94,125129]. Some transitions such as the 5D0 7F1 transition

    have magnetic dipole character. Magnetic dipole transitions (MDtransitions) are allowed by the Laporte selection rule, but theirintensities are weak and comparable to those of the induced electricdipole transitions [94]. The intensity of a magnetic dipole transition

  • K. Binnemans / Coordination Chemistry Reviews 295 (2015) 145 7

    Table 4Number of sublevels of a 2S+1LJ term for the different symmetry classes.

    Symmetry class Point groups J = 0 J = 1 J = 2 J = 3 J = 4 J = 5 J = 6

    Icosahedral Ih, I 1 1 1 2 2 3 4Cubic Oh, O, Td, Th, T 1 1 2 3 4 4 6Octagonal D8, C8v, S8, D4d 1 2 3 4 6 7 8Hexagonal D6h, D6, C6v, C6h, C6, D3h, C3h 1 2 3 5 6 7 9Pentagonal D5h, D5, C5v, C5h, C5 1 2 3 4 5 7 8Tetragonal D4h, D4, C4v, C4h, C4, S4, D2d 1 2 4 5 7 8 10Trigonal D3d, D3, C3v, C3i (=S6), C3 1 2 3 5 6 7 9Orthorhombic D2h, D2, C2v 1 3 5 7 9 11 13Monoclinic C2h, C2, Cs 1 3 5 7 9 11 13Triclinic C1, Ci 1 3 5 7 9 11 13

    Table 5Overview of the transitions observed in luminescence spectra of europium(III) compounds.

    Transitiona Dipole characterb Wavelength range (nm) Relative intensityc Remarks

    5D0 7F0 ED 570585 vw to s Only observed in Cn, Cnv and Cs symmetry5D0 7F1 MD 585600 s Intensity largely independent of environment5D0 7F2 ED 610630 s to vs Hypersensitive transition; intensity very strongly dependent on environment5D0 7F3 ED 640660 vw to w Forbidden transition5D0 7F4 ED 680710 m to s Intensity dependent on environment, but no hypersensitivity5D0 7F5 ED 740770 vw Forbidden transition5D0 7F6 ED 810840 vw to m Rarely measured and observeda Only transitions starting from the 5D0 level are shown.

    iitaifippmeaainoacrcqqccmtada

    TS

    b ED = induced magnetic dipole transition, MD = magnetic dipole transition.c vw = very weak, w = weak, m = medium, s = strong, vs = very strong.

    s largely independent of the environment and can be consideredn a first approximation to be constant [130]. For the calculation ofhe intensities of MD transitions, only the free-ion wave functionsre needed, not the crystal-field wave functions. A MD transitions caused by interaction of the lanthanide ion with the magneticeld component of the light via a magnetic dipole. If charge is dis-laced over a curved path during the transition, the transition willossess magnetic dipole character. The curvature of the displace-ent will only be weakly apparent in a volume as small as the

    xtent of a lanthanide ion, so that magnetic dipole transitions have weak intensity. Magnetic dipole radiation can also be considereds a rotational displacement of charge. Since the sense of a rotations not reversed under inversion through an inversion center, a mag-etic dipole transition has even parity. Therefore, a magnetic dipoleperator possesses even transformation properties under inversionnd allows transitions between states with even parity (i.e. intra-onfigurational transitions such as 4f4f transitions). The selectionules for ED and MD transitions are summarized in Table 6. In prin-iple, also electric quadrupole transitions could occur. An electricuadrupole transition arises from a displacement of charge that hasuadrupolar character. An electric quadrupole consists of four pointharges with overall zero charge and zero dipole moment. It can beonsidered as two dipoles arranged in such a way that their dipoleoments cancel out. An electric quadrupole has even parity. Elec-

    ric quadrupole transitions are much weaker than magnetic dipole

    nd induced electric dipole transitions. There is no convincing evi-ence for electric quadrupole transitions in lanthanide spectra,lthough hypersensitive induced electric dipole transitions obey

    able 6election rules for intraconfigurational ff transitions.

    Induced electric dipole transitions(ED)

    Magnetic dipole transitions(MD)

    |S| = 0 S = 0|L| 6 L = 0|J| 6 and |J| = 2, 4, 6 if J = 0 or

    J = 0 (as in the case of Eu3+)J = 0, 1, but 0 0 isforbidden

    the selection rules for electric quadrupole transitions (see Section9.3).

    The selection rules on S and L are only strictly valid inthe RussellSaunders coupling scheme. They are relaxed in theintermediate coupling scheme, so S and L are not good quantumnumbers in that scheme. Since J remains a good quantum numberin the intermediate coupling scheme, the selection rule for J ismore rigorous. It can be relaxed only by J-mixing. For these reasons,the 5D0 7FJ (J = 0, 3, 5) transitions have very weak intensities. J-mixing involves the mixing of the wave functions of sublevels ofdifferent J levels, when their irreducible representations are thesame. Thus, wave functions with the same symmetry can mix underthe influence of the crystal field. The degree of J-mixing betweentwo multiplets J and J is inversely proportional to the energy dif-ference between the J and J states.

    In general, luminescence spectra are recorded in a wavelengthscale (expressed in nanometers, nm). To facilitate the comparisonbetween different spectra, it is recommended to plot the spectrawith the shortest wavelength at the left hand side and the longestwavelength at the right hand side. In the older literature, the oppo-site convention is often used. The frequency of light is physicallymore significant than its wavelength, since the frequency remainsunchanged when the light wave propagates through various media.Moreover, the frequency is directly proportional to the energy E ofthe transition, via the formula E = h, where h is Plancks constant.Most spectroscopists prefer to deal with wavenumbers (numberof waves per cm) rather than frequencies. The wavenumber v isdefined as:

    v = vc= 1

    vac= 1

    nairair(7)

    To find the wavenumber for a transition measured in air, one hasto correct in principle for the refractive index of air, which is wave-

    length dependent. Except for high accuracy spectroscopic work,one can assume that nair = 1. In other cases, nair has to be calculated,for instance via the empirical Edln formula [131]. Wavenumbersare expressed in units of reciprocal centimeters (cm1). A spectrum

  • 8 hemistry Reviews 295 (2015) 145

    rw

    v

    Ihb

    3

    swJrimJtdsetdJomastc

    tmavtggi1tihEs

    t[mJpnrcstbtb5

    cmw[

    Fig. 4. Corrected luminescence spectra of Eu3+ in different ligand environments: (A)Eu3+ in water; (B) [Eu(DPA)]+ in water; (C) [Eu(DPA)3]3 in water. The spectra havebeen scaled so that the respective 5D0 7F1 bands have identical areas. Note thedifferent scales of the Y-axes.

    K. Binnemans / Coordination C

    ecorded in wavelength scale (in nanometers) can be converted toavenumber scale (in cm1) by applying the following formula:

    (cm1) = 107

    (nm)(8)

    t is recommended to plot spectra in wavenumber scale with theighest wavenumber at the left hand side and the lowest wavenum-er at the right hand side of the spectrum.

    .2. Transition 5D0 7F0

    The 5D0 7F0 transition is strictly forbidden according to thetandard JuddOfelt theory. The occurrence of this transition is aell-known example of the breakdown of the selection rules of the

    uddOfelt theory (a 00 transition is forbidden by the J selectionule of the JuddOfelt theory). Several authors have tried to theoret-cally explain why this transition is observed [132142]. Theoretical

    odels include the breakdown of the closure approximation in theuddOfelt theory and third order perturbation theory. However,he most obvious explanation is to assume that this transition isue to J-mixing [143147] or to mixing of low-lying charge-transfertates into the wavefunctions of the 4f6 configuration [148]. Asxplained in Section 3.1, J-mixing is due to the crystal-field per-urbation and causes mixing of the wavefunctions of terms withifferent J values. The wavefunction of the 7F0 state contains after

    -mixing also contributions from the J = 2, 4, 6 states. The mixingf the charge-transfer states is described in Section 4.6. The twoechanisms are not independent, since it has been noticed that

    n inverse relationship exists between the energy of the transfertate and the crystal-field strength: low energies for the charge-ransfer states result in strong crystal-field effects [149,150]. Strongrystal-field effects enhance J-mixing.

    The 5D0 7F0 transition belongs to the 4f4f transitions withhe smallest line width ever observed. The half width at half maxi-

    um of the 5D0 7F0 transition in EuCl36H2O is about 0.12 cm1t 4.2 K and about 1 cm1 at 250 K [151]. For Eu(NO3)36H2O, thealues are 0.18 and 2 cm1, respectively [151]. For Eu(BrO3)39H2O,he values are 1.1 cm1 at 77 K and 2.3 cm1 at 295 K [152]. Inlasses, the 5D0 7F0 transition is much broader due to inhomo-eneous line broadening. For example, the line width is 105 cm1

    n calcium diborate glass [153], 119 cm1 in phosphate glasses and49 cm1 in silicate and germanate glasses [154]. It was concludedhat about 50 slightly different sites of Cs symmetry are presentn phosphate, silicate and germinate glasses, by comparison of thealf width at half maximum in glasses with the value of 2 cm1 inu2O3. The slight differences in the environment are the result ofmall differences in metalligand angles and distances.

    The observation of the 5D0 7F0 transition is an indicationhat the Eu3+ ion occupies a site with Cnv, Cn or Cs symmetry155]. The occurrence of the 5D0 7F0 transition in these sym-

    etries can be understood by considering the selection rules. A = 0 state must transform as the identity representation of theoint symmetry group and this requires that some of the compo-ents of the electric dipole operator also transform as the identityepresentation. This is the case for point groups for which therystal-field potential contains C1q spherical harmonics, i.e. theymmetry groups Cnv, Cn or Cs. Nieuwpoort and Blasse noticed thathe 5D0 7F0 transition always appears whenever it is allowedy the observed site symmetry [139]. The fact that this transi-ion is observed only for certain symmetries is nicely illustratedy the consecutive formation of dipicolinate (DPA) complexes. TheD0 7F0 transition occurs for the intermediate low symmetry

    omplexes [Eu(DPA)]+ and [Eu(DPA)2], but not for the high sym-etry complexes [Eu(H2O)9]3+ (D3h) and [Eu(DPA)3]3 (D3). Thisas first observed in the absorption spectra of these complexes

    156], and later in the luminescence spectra (Fig. 4) [157].

    Reprinted with permission of The Royal Society of Chemistry from reference [157].Copyright 2002 The Royal Society of Chemistry.

    In most europium(III) spectra, the 5D0 7F0 transition is veryweak, even for complexes with Cnv, Cn or Cs symmetry. However,the 5D0 7F0 transition is unusually intense in the -diketonatecomplex [Eu(dbm)3(H2O)], (dbm = dibenzoylmethanate), with theEu3+ ion at a site with C3 symmetry [158]. In the luminescence spec-trum of this complex, the 5D0 7F0 transition has a higher peakheight than the 5D0 7F1 transition, although the latter transitionhas the largest integrated peak area, due to the extreme narrow-ness of the 5D0 7F0 transition. Other examples of europium(III)complexes with intense 5D0 7F0 transitions are the complexes ofnitrilotriacetate (NTA) and of the macrocyclic ligand DOTA [159]. InSr2TiO4:Eu3+, the transition 5D0 7F0 is 1.65 times more intensethan the 5D0 7F1 transition [160]. The high intensity was ascribedto the ordered crystal structure of Sr2TiO4, which leads to large lin-ear terms in the crystal-field potential. Unusually high intensitiesfor the 5D0 7F0 transition are also observed for Eu3+ in fluo-rapatite Ca5(PO4)3F:Eu3+ [161], hydroxyapatite [162], oxysulfatesLn2O2SO4:Eu3+ (Ln = La, Gd, Y) [163], LaOCl:Eu3+ [164], -cordierite[165,166], mullite [167]. La2Si2O7 [167], La2O3:Eu3+ [168], C-type oxides (Gd2O3, Lu2O3, Lu2O3, Y2O3, In2O3, Sc2O3) [169] andBa4Ln2ZrWO12:Eu3+ (Ln = La, Gd, Y) [170]. In Sr5(PO4)3F:Eu3+, withthe Eu3+ ion in the Sr2+ site with a charge-compensating oxideion substituting a nearest-neighbor fluoride ion in the lattice, the5D0 7F0 transition dominates the spectrum: the intensity ratioI(5D0 7F0)/I(5D0 7F1) is larger than 20 [171]. This intensity ratio

    shows the following trend for Eu3+ doped in oxybromides: YOBr(

  • K. Binnemans / Coordination Chemis

    Fig. 5. Part of the luminescence spectra of Eu3+ doped in oxybromides at 77 K:(a) YOBr; (b) GdOBr and (c) LaOBr. The increase in the intensity of the 5D0 7F0transition in the order YOBr < GdOBr < LaOBr is clearly visible.

    Reprinted with permission from reference [149] Copyright 1982 AIP Publishing LLC.

    try Reviews 295 (2015) 145 9

    LaOBr:Eu3+. In layered crystal structures, the intensity of the tran-sition strongly depends on the details of the layer packing and theinterionic distances [172]. The most intense 5D0 7F0 transitionever reported is that of the Cs(O2) site of BaFCl:Eu3+, with a charge-compensating oxide ion substituting a nearest-neighbor fluorideion in the lattice [148]. This transition is 25 times more intensethan the 5D0 7F1 magnetic dipole transition! Such an extremeintensity cannot be explained by J-mixing. A possible explana-tion is mixing of charge-transfer states into the 4f6 levels of Eu3+.As will be explained in Section 4.6, the charge-transfer states arelying at a much lower energy in Eu3+ than in the other lanthanideions. As a consequence, a much stronger interaction between thecharge-transfer states and the lower levels of the 4fn configura-tion is expected. This mixing also gives rise to other anomalies inthe crystal-field spectra, as described by Chen and Liu [148]. Thereis a correlation between the solvent basicity and the intensity ofthe 5D0 7F0 transition in a tetrakis -diketonate complex with atetraalkylphosphonium counter ion [173]. The higher intensity inthe more basic solvents was attributed to a higher nucleophilicityof the solvent and a resulting change in the coordination sphere byinteraction between the solvent and the Eu3+ ion.

    The 5D0 7F0 transition is also useful for the determinationof the presence of non-equivalent sites in a host crystal or fordetermination of the number of different europium(III) species insolution, because maximum one peak is expected for a single siteor species, due to the non-degeneracy of the 7F0 and 5D0 levels[80,98,174,175]. The observation of more than one peak in the spec-troscopic region where the 5D0 7F0 transition is expected, showsthat more than one site or species is present, but it does not allowthe determination of the exact number of sites or species, becausesites or species with a symmetry other than Cnv, Cn or Cs do notgive an observable 5D0 7F0 transition. The luminescence spectraof the individual sites of emitting species can be observed sepa-rately by site-selective excitation via a tunable laser source. Thiswas nicely illustrated by Bnzli and coworkers for europium(III)crown ether complexes (Fig. 6) [176,177]. Four luminescence cen-ters could be detected in the garnet Y3Al5O12:Eu3+ by site-selectiveexcitation [178]. Site selective excitation has often been used toprobe the local structure in Eu3+-doped glasses and glass ceramics(Fig. 7) [179186].

    If the structural difference between two sites is small, theenergy differences between the different peaks in the 5D0 7F0region are small as well and the presence of more than onesite is only revealed by an asymmetric shape of the 5D0 7F0line or as a shoulder. However, the presence of two dis-tinct geometrical isomers in a crystal structure can result inquite a large energy difference between the transitions in the5D0 7F0 region. This is illustrated by the splitting of 35 cm1in the luminescence spectrum of tris(dipivaloylmethanato)(2,9-dimethyl-1,10-phenanthroline)europium(III) [187]. It is evidentthat when mixtures of different complexes are present in solu-tion the energy differences between the different transitions inthe 5D0 7F0 region can be large as well. Monitoring the intensityof the 5D0 7F0 transition as a function of the ligand concentra-tion has been used to determine stability constants of complexes[188191]. It must be mentioned that often luminescence exci-tation spectra rather than emission spectra are being used formeasurement of stability constants (see Section 5). The quadraticshift of the 5D0 7F0 transition energy of Eu3+ with temperaturehas been used to determine the operating temperature of phos-phor screens in cathode-ray tubes. The method is reliable thanmeasurement of the relative intensities of the transitions in the

    luminescence spectrum [192].

    If the 5D0 7F0 transition is strictly forbidden by the selectionrules, the determination of the energy of the 5D0 state becomes

  • 10 K. Binnemans / Coordination Chemistry Reviews 295 (2015) 145

    Fig. 6. Luminescence spectra of the europium(III) 21-crown-7 complex[Eu(NO3)2(21C7)]3[Eu(NO3)6] at 77 K: (a) excitation at 395 nm, (bd) site-selectiveexcitation of the 5D0 7F0 transitions of the different sites.Reprinted with permission from reference [177]. Copyright 1989 American ChemicalSociety.

    Fig. 7. Time-resolved line-narrowed emission spectra of 5D0 7F0,1,2 transitions ofEu3+ ions in 60NaPO315BaF224.5YF30.5EuF3 fluorophosphate glass. The lumi-nescence was measured at 4.2 K at a time delay of 1 ms after the laser pulse and atdifferent excitation wavelengths. Different sites can be recognized.

    Reprinted with permission from reference [186]. Copyright 1996 The AmericanPhysical Society.

    less straightforward. However, an accurate location of the 5D0 stateis required for a precise determination of the 7FJ levels, becausethe 5D0 state is the initial level for the 5D0 7FJ transitions. In theabsence of the 5D0 7F0 transition, the position of the 7F1 level canbe determined from the wavenumber of the 5D1 7F1 transition,and the wavenumber of the 5D0 7F1 transition can then be usedto determine the position of the 5D0 level. In case the 5D1 7F1transition is not observed in the luminescence spectrum, the cor-responding absorption spectrum can be measured to determine theposition of the 5D1 7F1 transition.

    At low temperatures and by using a spectroscopic setup with anextremely high resolution, it is possible to observe fine structurefor the 5D0 7F0 transition due to the hyperfine interactions withthe nuclear momenta I of the nuclei of the 151Eu and 153Eu isotopes.By this interaction, the 7F0 and 5D0 levels split each into three sub-levels. Hyperfine splitting has been observed in the high resolutionspectra of EuCl36H2O [193,194]. Further hyperfine structure wasobserved due to interactions with the nuclear momenta of H, Cl andO isotopes in the sample [193].

    3.3. Transition 5D0 7F1

    The 5D0 7F1 transition is a magnetic dipole (MD) transi-tion. Although the intensity of a magnetic dipole transition islargely independent of the environment of the Eu3+ ion, it must

  • hemis

    bttctotnts5

    titetsss[prs5C8tfilts7

    bwspsnclatosGEm7

    pNELonpctnspbhwsi

    K. Binnemans / Coordination C

    e noticed that the invariability of the intensity of the 5D0 7F1ransition only applies to the total integrated intensity of thisransition and not to the individual intensities of the crystal-fieldomponents [130]. The total intensity of the 5D0 7F1 transi-ion can be influenced by J-mixing. Nevertheless, the intensityf this transition is often considered to be constant and thisransition is used to calibrate the intensity of europium(III) lumi-escence spectra. For comparison of two luminescence spectra,he intensities are scaled in such a way that the 5D0 7F1 tran-ition has the same (integrated) intensity in the two spectra. TheD0 7F1 transition directly reflects the crystal-field splitting ofhe 7F1 level. In cubic or icosahedral crystal fields, the 7F1 levels not split. In hexagonal, tetragonal and trigonal crystal fields,he 7F1 level is split into a non-degenerate and a twofold degen-rate crystal-field level. In orthorhombic or lower symmetries,he total removal of crystal field degeneracies results in threeublevels for 7F1. The total splitting of the 7F1 level in highlyymmetric compounds ranges between 0 cm1 for the cubic elpa-olites (e.g. Cs2NaEuCl6) [195,196] to 346 cm1 for LaOBr:Eu3+

    149]. For Eu3+ compounds with a low site symmetry, exam-les of an even larger total splitting of the 7F1 level have beeneported: 392 cm1 for LaMgB5O10:Eu3+ [197], 456 cm1 for the Aite in Gd2(SiO4)O:Eu3+ [198], 476 cm1 for LaBGeO5:Eu3+ [199],53 cm1 for Y6WO16:Eu3+ [200], 653 cm1 for the A site ina10xEux(PO4)6O1+x/2 [201], 724 cm1 for cordierite [166], and87 cm1 for hydroxyapatite [167]. If the crystal-field splitting ofhe 7F1 level is very large, there will be an overlap with the crystal-eld sublevels of the 7F2 state. As a consequence, the crystal-field

    ines of the 5D0 7F1 transition overlap with those of the 5D0 7F2ransition. This is a very exceptional situation. The crystal-fieldublevels of the 7F1 level can be discriminated from those of theF2 level, by relying on the empirical correlations between thearycenter of the 7F1 state and the position of the 5D0 level, asell as between the barycenters of the 7F1 and 7F2 levels [202]. A

    pectrum with very large splitting of the 7F1 level into three com-onents and a missing 5D0 7F0 transition can be mistaken for apectrum consisting of a splitting of the 7F1 level in two compo-ents and the 5D0 7F0 transition present. Here too, the empiricalorrelation between the barycenter of the 7F1 state and the 5D0evel is helpful, as illustrated for the luminescence spectrum ofn ionic europium(III) complex with Schiff base ligands [203]. Onhe other hand, a small splitting of the 7F1 level is observed notnly for systems with a cubic or approximately cubic symmetry,uch as the elpasolites [196,204,205] and oxyfluorides (LaOF:Eu3+,dOF:Eu3+, YOF:Eu3+) [206,207], but also for the double nitratesu2M3(NO3)1224H2O (M = Mg, Zn) [208,209], which have a sym-etry close to that of an icosahedron [210]. A small splitting of the

    F1 level is also present in many systems with a tricapped trigonalrism as the coordination polyhedron around the Eu3+ ion, such asa3[Eu(ODA)3]2NaClO46H2O (also called EuODA) [112,211,212],u(BrO3)39H2O [152,213,214], Eu(C2H5SO4)39H2O [215,216], andaCl3:Eu3+ [217]. It is not totally unexpected that small splittingsf the 7F1 level are found for systems with a tricapped trigo-al prism or an icosahedron as coordination polyhedron. In theseolyhedra with a high coordination number, the atoms in the firstoordination sphere have a fairly equal spatial distribution andhis distribution is mimicking a spherical distribution for whicho splitting of the 7F1 level occurs [218]. This small crystal-fieldplitting can also be explained by simple calculations based on aoint charge electrostatic model (PCEM) [219]. The PCEM model haseen used to study the splitting of the 7F1 level in a series of oxideost matrices [220]. Due to the presence of many different sites

    ith different crystal-field strengths, a large range of 7F1 splitting

    izes can be observed in one glass host. For instance, by the laser-nduced fluorescence line narrowing technique, a variation for the

    try Reviews 295 (2015) 145 11

    7F1 splitting between 150 and 550 cm1 was observed for Eu3+ ina silicate glass [221]. Similar results were observed for Eu3+-doped40Bi2O340PbO10Ga2O310GeO2 and 60GeO225PbO15Nb2O5glasses [222], as well as TeO2TiO2Nb2O5 glass [223].

    The 5D0 7F1 transition is the most intense transition inthe spectra of solids with a centrosymmetric crystal struc-ture. This is nicely illustrated by the luminescence spectraof Ba2GdNbO6:Eu3+ (perovskite structure, Oh symmetry) andGd2Ti2O7:Eu3+ (pyrochlore structure, approximate symmetry D3d)[224]. This 7F1 level is not split for Ba2GdNbO6:Eu3+ (Oh), whereasit is split for Gd2Ti2O7:Eu3+ (D3d), as predicted by theory. For thesetwo centrosymmetric host crystals, the transition 5D0 7F4 wasnot observed. Besides Gd2Ti2O7:Eu3+, other compounds with apyrochlore structure such as Gd2Sn2O7:Eu3+ and Gd2TiSnO7:Eu3+

    have a luminescence spectrum that is dominated by the 5D0 7F1transition [225]. Also in perovskites other than Ba2GdNbO6:Eu3+,such as Ba2GdTaO6:Eu3+ and Ba2GdNbO6:Eu3+, the 5D0 7F1 tran-sition is dominant [225]. The influence of the cation size onthe structure of the host matrix and hence on the luminescencespectra of Eu3+ in these host matrices is nicely illustrated for aseries of borate compounds [226]. The low-temperature lumines-cence spectra of Ba2LnNbO6:Eu3+ (Ln = Gd, Y) are dominated bythe 5D0 7F1 transition [227]. The coordination polyhedron canbe described as a distorted octahedron. The analysis of the split-ting pattern reveals that the actual symmetry is C2h or Ci. Thefact that the coordination polyhedron is close to an ideal octa-hedron is evident from the very small splitting of the 7F1 level(13 cm1). Interestingly, the low-temperature luminescence spec-trum of the related compound Ba2LaNbO6:Eu3+ is dominated bythe 5D0 7F2 transition [228]. The symmetry of this compoundis low: C2 or C2v. In SrTiO3 with the cubic perovskite structure, theEu3+ enters the centrosymmetric Sr2+ site and is twelve-coordinate[229]. This results in the typical spectrum of a centrosymmetriceuropium(III) compound with an intense 5D0 7F1 transition. Asimilar situation is found for SrSnO3:Eu3+, where up to 2 at.% ofEu3+ can enter the Sr2+ sites [230,231]. The 5D0 7F1 transitionis the most intense transition in the cathodoluminescence spectraof InBO3:Eu3+ and ScBO3:Eu3+with the centrosymmetric rhom-bohedral calcite structure (C3i symmetry). LuBO3:Eu3+ occurs astwo polymorphs, one with the calcite structure and one withthe pseudovaterite structure (D3 symmetry, no center of symme-try). The 5D0 7F1 transition is the most intense transition forthe two structures, but the 5D0 7F2 transition is more intensein the pseudo-vaterite polymorph than in the calcite polymorph.YBO3:Eu3+ and GdBO3:Eu3+ have a pseudo-vaterite structure (D3symmetry). In these compounds, the 5D0 7F1 transition stilldominates the luminescence spectrum, but the intensity of the5D0 7F2 transition cannot be neglected. LaBO3:Eu3+ with thelargest host cation has the orthorhombic aragonite structure (withEu3+ in an asymmetric site with Cs symmetry), and in this casethe 5D0 7F1 transition is no longer the most intense transition.The 5D0 7F2 transition is the most intense, but the 5D0 7F4transition has a remarkably high intensity. The 5D0 7F1 transi-tion dominates the luminescence spectrum of [Eu(TMU)6(AsF6)3](TMU = tetramethylurea), where the Eu3+ is at an octahedral sitewith Oh symmetry [232]. Dominance of the 5D0 7F1 transitionis also seen in the room-temperature luminescence spectra of thecubic site of ThO2:Eu3+ [233,234]. In many fluoride-containingcompounds, the 5D0 7F1 transition is the most intense transi-tion, for instance LaF3:Eu3+ [235241], EuF3 [242,243], GdF3:Eu3+

    [236,244,245] and KGdF4:Eu3+ [246,247], but not LiGdF4:Eu3+

    [248,249], in hexagonal or cubic NaGdF4:Eu3+ [250253]. In com-

    pounds with the delafossite structure, e.g. CuLa1xEuxO2, Eu3+ is at acentrosymmetric site and the luminescence has an orange color dueto the strong 5D0 7F1 and 5D0 7F0 transitions [254256]. The

  • 1 hemistry Reviews 295 (2015) 145

    lttAttps

    tedewototllaite

    3

    tbls|rtmtpitoiccYtiitit

    avm5

    I3ddtsai

    000210005100081

    600 70 0 80 0

    Inte

    nsity

    (a.u

    .)

    wavenu mber (c m-1)

    waveleng th (n m)

    a b

    c

    d e f g

    Fig. 8. Luminescence spectrum of the tetakis -diketonate complex[C6mim][Eu(tta)4], C6mim = 1-hexyl-3-methylimidazolium, tta = 2-thenoyltrifluor-acetylacetonate (room temperature, exc = 340 nm). The assignment of the linesis: (a) 5D0 7F0; (b) 5D0 7F1; (c) 5D0 7F2; (d) 5D0 7F3; (e) 5D0 7F4; (f)5D0 7F5 and (g) 5D0 7F6. The dominance of the spectrum by the hypersensitive

    2 K. Binnemans / Coordination C

    uminescence spectrum of Eu(ClO4)3 in water shows the 5D0 7F1ransition as the most intense transition in the spectrum, indicatinghat the Eu3+ aquo ion probably possesses an inversion center [257].nother observation is that the relative intensities of the transi-

    ions and the shapes of the luminescence bands do not depend onhe concentration of the perchlorate ion. These data show that theerchlorate ion does not coordinate to the Eu3+ ion, even not at highalt concentrations.

    The presence of more than three lines for the 5D0 7F1ransition is an indication for the presence of more than one non-quivalent site for the Eu3+ ion. This transition can be used toetect multiple sites if the 5D0 7F0 transition is forbidden. How-ver, one has to be cautious not to confuse vibronic transitionsith purely electronic transitions. The splitting of the 7F1 level

    bserved by the 5D0 7F1 transition in the luminescence spec-rum of a europium(III) can be used as a direct measure of the valuef the second rank crystal-field parameter B20 [258]. This parame-er is directly proportional to the magnetic anisotropy of theanthanide complex. Therefore, the splitting of the 7F1 level in theuminescence spectrum can be used as a probe for the magneticnisotropy of lanthanide complexes. The magnetic anisotropy is ofmportance to explain the lanthanide-induced shift in NMR spec-ra and the alignment of lanthanide-containing liquid crystals in anxternal magnetic field [259262].

    .4. Transition 5D0 7F2

    The 5D0 7F2 transition is a so-called hypersensitive transi-ion, which means that its intensity is much more influencedy the local symmetry of the Eu3+ ion and the nature of the

    igands than the intensities of the other ED transitions. Hyper-ensitive transitions obey the selection rules |S| = 0, |L| 2 andJ| 2 [94]. These selection rules are the same as the selectionules for a quadrupole transition, but calculations have shown thathe intensities of hypersensitive transitions are several orders of

    agnitude larger than the values expected for quadrupole transi-ions. Therefore, hypersensitive transitions have been labeled alsoseudo-quadrupole transitions [263]. Hypersensitivity is discussedn more detail in Section 9.3. The intensity of the hypersensitiveransition 5D0 7F2 is often used as a measure for the asymmetryf the Eu3+ site (see Section 7). Large variations are observed for thentensity of this transition, depending on the type of europium(III)ompound. The 5D0 7F2 transition is responsible for the typi-al red luminescence observed in europium(III) phosphors such as2O3:Eu3+ or Y2O2S:Eu3+ [49,264]. The intensity of the 5D0 7F2ransition is directly proportional to the value of the JuddOfeltntensity parameter 2 (see Section 9.1). Instead of the absolutentensity of the 5D0 7F2 transition, the ratio R of the intensities ofhe transitions 5D0 7F2 and 5D0 7F1, I(5D0 7F2)/I(5D0 7F1)s also often used to compare the intensities of the hypersensitiveransition in different europium(III) compounds.

    Europium(III) -diketonate complexes, either Lewis basedducts of tris complexes or tetrakis complexes, have typically aery intense hypersensitive 5D0 7F2 transition. It is not uncom-on that the 5D0 7F2 transition is 10 times more intense than the

    D0 7F1 transition in this type of complexes [12,13,15,265,266].n Fig. 8, the luminescence spectrum of the europium(III) tetrakis-diketonate complex [C6mim][Eu(tta)4] (where C6mim = 1-hexyl--methylimidazolium and tta = 2-thenoyltrifluoracetylacetonate)oped into an ionogel is shown [13]. The 5D0 7F2 transitionominates the spectrum. The high intensity is often attributed

    o the low symmetry of the Eu3+, but it is more realistic to con-ider the high polarizability of the chelating -diketonate ligandss the intensity enhancing mechanism [158]. A dramatic increasen intensity of the hypersensitive transition 5D0 7F2 is observed

    transition 5D0 7F2 is evident.Reprinted with permission from reference [13]. Copyright 2009 American ChemicalSociety.

    for the luminescence spectrum of Eu3+ in an aqueous solution ofK2CO3 in comparison with the spectrum of the europium(III) aquoion [267,268]. This intensity enhancement is due to the forma-tion of the anionic carbonato complex [Eu(CO3)4]5 in solution.The intensification finds applications in analytical chemistry: Sinhadeveloped a spectrofluorimetric method to detect Eu3+ concen-trations as low as 107 M using a 3 M aqueous solution of K2CO3[269]. A sharp decrease in the intensity of the 5D0 7F2 transi-tion was observed when water was added to Eu(Tf2N)3 dissolvedin the hydrophobic ionic liquids N,N-diethyl-N-methyl-N-(2-methoxyethyl)ammonium bis(trifluoromethylsulfonyl)imide or1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide[270]. Addition of dipicolinate ions to an aqueous solution of Eu3+

    led to a very strong increase in the intensity of the 5D0 7F2 tran-sition, reaching a maximum when the [Eu(DPA)3]3 complex wasformed [157].

    If the 5D0 7F2 transition is very weak, the luminescencespectrum is dominated by the 5D0 7F1 transition and anorange luminescence color is observed [271]. Examples ofeuropium(III) compounds with an orange photoluminescenceare Na9EuW10O3618H2O (D4d symmetry) [272], YF3:Eu3+ (D4d)[273], GdB3O6:Eu3+ (D4d), CeO2:Eu3+ (Oh) [274], [Eu(4-picoline-N-oxide)8](PF6)3 (D4d) [275], [Eu(pyridine-N-oxide)8](ClO4)3(D4d) [276], Eu(antipyrene)6I3 (S6) (antipyrene = 1-phenyl-2,3-dimethyl-5-pyrazolone) [277,278], compounds withthe hexakis(nitrito)europate(III) ion [Eu(NO2)6]3 (Th)[279282], SnO2:Eu3+ (D2h) [283], Gd2Sn2O7:Eu3+ (D3d)[284], Na3[Eu(oxydiacetato)3]2NaClO46H2O (D3) [271,285],[Eu(terpy)3](ClO4)3 (D3) [286], [Eu(H2O)9](BrO3)3 (D3h) [152,287],and [Eu(H2O)9](EtSO4)3 (C3h) [215]. A pink luminescence isobserved for Cs2NaEuCl6 (Oh) at room temperature, but an orangeluminescence at 77 K, due to a decrease of the vibronic intensity of

    the 5D0 7F2 transition [271]. These examples show that correlat-ing the luminescence color with a particular symmetry point groupis difficult. The list contains compounds with different symmetries,and both centrosymmetric and non-centrosymmetric point groups

  • K. Binnemans / Coordination Chemis

    Fig. 9. Luminescence spectrum of tris(hydrotris(1-pyrazolyl)borato)europium(III),a

    R

    o5

    5

    tEttbhtpOaay(Ets

    mocpctt

    3

    fctpip5

    cbhohitT

    complex with D3h symmetry.

    eprinted with permission from reference [289]. Copyright 2002 Wiley-VCH.

    ccur. One could conclude from an orange luminescence that theD0 7F2 transition must be weak and much less intense than theD0 7F1 transition, but one has to be cautious for compoundshat also show emission from higher excited states (5D1, 5D2, 5D3).mission from higher excited states can shift the luminescenceoward orange and yellow emission colors [288]. The relative con-ribution of emission from the higher excited states can be tunedy variation of the Eu3+ concentration in the host matrix, becauseigher doping concentrations favor emission from the 5D0 level athe expense of emission from the higher excited states. Not all phos-hors show a strong color shift as a function of Eu3+ concentrations.nly phosphors with a large contribution of 5D1 and 5D2 emissiont low Eu3+ concentrations exhibit strong color shifts. Examplesre the white to orange to red emission with (Y1xEux)2O2S and theellow to red emission with (Y1xEux)2O3 [288]. On the other hand,Y1xEux)VO4 shows very little color change upon variation of theu3+ concentration. Also compounds with an intense 5D0 7F2ransition shifted to higher energies (shorter wavelengths) canhow an orange photoluminescence is expected.

    A typical feature of europium(III) complexes with a D3h sym-etry is the narrowness of the 5D0 7F2 transition, because

    nly one crystal-field line is allowed in this symmetry. Thisan be seen in the luminescence spectra of tris(hydrotris(1-yrazolyl)borato)europium(III) (Fig. 9) [289]. For D3 symmetry, twoomponents are expected for the 5D0 7F2 transition. This split-ing is sometimes not resolved, as in the case of the europium(III)ris dipicolinate complex [Eu(DPA)3]3 [157].

    .5. Transition 5D0 7F3

    The 5D0 7F3 transition is in general very weak, because it isorbidden according to the JuddOfelt theory, and this transitionan only gain intensity via J-mixing [290]. An intense 5D0 7F3ransition is a sign of strong J-mixing and a strong crystal-fielderturbation. This transition is not considered when the Eu3+

    on is used as a spectroscopic probe. The -diketonate com-lex [Eu(dbm)3(H2O)] is one of the rare examples of an intenseD0 7F3 transition [158]. In fact, the 5D0 7F3 transition of thisompound is more intense than its 5D0 7F4 transition. It shoulde noted that also the 5D0 7F0 transition and the 5D0 7F2ypersensitive transition are very intense in this compound. Thisbservation can be explained by strong crystal-field effects and

    5 7

    ence strong J-mixing. On the other hand, the D0 F3 transitions totally absent in BaEu(CO3)2F and Na3La2(CO3)4F:Eu3+, althoughhese compound give fairly intense luminescence spectra [291].he absence of this transition was attributed to weak J-mixing,

    try Reviews 295 (2015) 145 13

    which was also evident from the small values of the second andfourth rank crystal-field parameters (B2q and B

    4q). Another remark-

    able feature in the luminescence spectra of these compounds is theabsence of luminescence from excited states higher than 5D0. Thisis attributed to the high phonon energies of the carbonate groupswhich efficiently depopulate the excited states. The 5D0 7F3 tran-sition of the C2v site in BaFCl:Eu3+ is more intense than the strongestline of the 5D0 7F4 transition [148]. This anomalous behavior wasexplained by J-mixing induced by the large fourth rank crystal fieldparameters (B40 = 1489 cm1 and B44 = 1266 cm1). The extent ofJ-mixing was estimated to be about 6.5% (which means that the7F3 state has 93.5% 7F3 character and 6.5% 7F2 character). Inter-estingly, the extent of J-mixing of 7F2 into 7F0 was in this compoundonly about 2%, due to the small second rank crystal field parameters(B20 = 72 cm1 and B22 = 290 cm1). As a result, the 5D0 7F0transition has a weak intensity.

    In the luminescence spectrum of Mg3F3BO3:Eu3+, a very intensetransition is observed in the 5D0 7F3 transition region at 658.3 nm[292]. This transition is much more intense than the 5D0 7F4 tran-sition. The compound has also other remarkable properties, such asa very intense 5D0 7F0 transition situated at a very high energy(17,615 cm1 or 567.7 nm) and a very large splitting of the 7F1 level(700 cm1). This very large splitting causes an overlap between theenergy levels of the 7F1 and 7F2 levels. The fact that a very strongcrystal-field effect is present inspired the authors to give an alterna-tive explanation for the transition at 658.3 nm instead of attributingthis line to the 5D0 7F3 transition. The authors suggest that theline could also be a crystal-field component of the 5D0 7F2 tran-sition. In that case, a very large crystal-field splitting of the 7F2level would occur (1750 cm1). Further research on this interestingcompound is recommended.

    3.6. Transition 5D0 7F4

    One must be careful with the interpretation of the intensityof the 5D0 7F4 ED transition. The transition lies in a spectro-scopic region in which most photomultiplier tubes have a lowsensitivity. Correction of the luminescence spectra is required,because otherwise erroneous conclusions could be drawn. Inan uncorrected luminescence spectrum, the intensity of the5D0 7F4 transition is too low compared to the other transi-tions, whereas the intensity of this transition is exaggerated inan over-corrected spectrum. The intensity of the 5D0 7F4 tran-sition should not be considered in terms of absolute values, butrather compared to the intensity of the 5D0 7F1 magnetic dipoletransition. In many europium luminescence spectra, the 5D0 7F4transition is weaker than the 5D0 7F2 transition, but severalexceptions are known. The luminescence spectra of compoundswith D4d symmetry are often dominated by the 5D0 7F4 tran-sition. In D4d symmetry, the 5D0 7F2 transition is forbidden,but the 5D0 7F4 transition is intense because a center of sym-metry is absent [271,293]. Examples of such compounds are:Na9[EuW10O36]14H2O (Eu3+ decatungstate) [272,294], YF3:Eu3+,GdB3O6:Eu3+ [293], [Eu(4-picoline-N-oxide)8](PF6)3 [295] and[Eu(4-picoline-N-oxide)8](ClO4)3 [296]. An undistorted squareantiprism has D4d symmetry, so that for compounds with alower symmetry than D4d, but with a coordination polyhedronclose to a square antiprism, have an intense 5D0 7F4 transition(and a weak 5D0 7F2 transition). In the macrocyclic complex[Eu(DOTA)(H2O)], the Eu3+ is nine-coordinate, with a coordina-tion polyhedron that can be described as a monocapped squareantiprism [159,297]. A very intense 5D0 7F4 transition has been

    observed for the alkalimetal europium dinitrosalicylates (espe-cially for the sodium complex), but the crystal structure of thesecompounds is not known yet [298]. In these compounds, the5D0 7F4 transition is less intense than the 5D0 7F2 transition,

  • 14 K. Binnemans / Coordination Chemis

    Fig. 10. Luminescence spectrum of Ca Sc Si O :Eu3+, with an intense 5D 7Ft

    R

    btamotgestiTttotEfbSoahtiddtt3tttfsv

    3

    5

    csPt

    5

    3 2 3 12 0 4

    ransition.

    eprinted with permission from reference [299]. Copyright 2011 Elsevier.

    ut much more intense than the 5D0 7F1 magnetic dipoleransition. The same remark can be made for LaBO3:Eu3+ withn orthorhombic aragonite structure (with Eu3+ in an asym-etric site with Cs symmetry) [226]. The very high intensity

    f the 5D0 7F4 transition in Ca3Sc2Si3O12:Eu3+ was attributedo a distortion of the cubic geometry of the Eu3+ site in thisarnet host toward the actual D2 symmetry (Fig. 10) [299]. How-ver, an alternative explanation is a distortion of the cube to aquare antiprism. The 5D0 7F4 transition dominates the spec-rum of GdOBr:Eu3+, whereas the 5D0 7F2 transition is the mostntense transition in the isostructural GdOCl:Eu3+ compound [172].his clearly shows that the intensity of the 5D0 7F4 transi-ion is determined not only by symmetry factors, but also byhe chemical composition of the host matrix. Other examplesf europium(III)-containing systems with an intense 5D0 7F4ransition are: Eu(Tp)3 (Tp = hydrotris(pyrazol-1-yl)borate) [300],u(Tp)3 in PMMA polymer matrix [301] and the two-dimensionalrameworks of the formula 2[Eu2Cl6(4, 4 bipy)3] 2(4, 4 ipy), where 4,4-bipy = 4,4-bipyridine [302]. In a recent paper,kaudzius et al. have made a systematic study of the intensityf the 5D0 7F4 transition of Eu3+ in different orthophosphatend garnet host matrices and investigated the influence of theost material, in particular of the electronegativity, the radius ofhe rare earth and of other trivalent cations [303]. An increasen the average electronegativity of the trivalent cations, i.e. aecrease of the optical basicity, in the octahedral and tetrahe-ral sites in the structure of the garnets and orthophosphates ledo an increase of the relative intensity of the 5D0 7F4 transi-ion. In Y3Al5O12:Eu3+(1%), the 5D0 7F4 transition accounts for9.5% of the total intensity of the 5D0 7FJ transitions, whereashis value increases to 49.8% in LuPO4:Eu3+(1%). The 5D0 7F4ransition is sometimes considered as a hypersensitive one, buthis it is not correct, since it does not obey the selection rulesor quadrupole transitions (J /= 2). The variations in the inten-ity ratios I(5D0 7F4)/I(5D0 7F1) are much less pronounced thanariations in the ratio I(5D0 7F2)/I(5D0 7F1).

    .7. Transitions 5D0 7F5 and 5D0 7F6

    In many studies, the 5D0 7F5 transition (740770 nm) and theD0 7F6 transitio