Synthesis, X-ray Structure, Spectroscopic Characterization, and Theoretical Prediction of the Structure and Electronic Spectrum of Eu(btfa)3·bipy and an Assessment of the Effect of Fluorine as a β-Diketone Substituent on the Ligand-Metal Energy Transfer Process

Article abstract of DOI:10.1021/ic971602v

The 2,2′-dipyridyl adducts of two europium β-diketonate complexes, Eu(btfa)₃*bipy [btfa = 4,4,4-trifluoro-1-phenyl-2,4-butanedione, bipy = 2,2′-dipyridyl] and Eu(bzac)₃·bipy [bzac = 1-phenyl-2,4-butanedione], have been prepared. The crystal structure of the former with chemical formula EuC₄₀H₂₆O₆F₉ has been solved by single-crystal X-ray diffraction methods. The complex crystallizes in the monoclinic space group P2₁/n with a = 11.122-(5) A?, b = 22.860(8) A?, c = 15.870(6) A?, β= 102.62(3)°, V = 3937(5) A?₃, and Z = 4. A single, eight-coordinate environment, which approximates a square antiprism, is found for the europium(III). The UV absorption spectra of both complexes were obtained from ethanol solutions and, in the case of Eu(btfa)₃·bipy, from a thin film. In both cases the absorption spectra are reasonably well predicted by the INDO/S-CI method using, for Eu(btfa)₃·bipy, both the X-ray data and that obtained through the SMLC/AM1 method as input geometry and, for Eu(bzac)₃·bipy, that obtained through the SMLC/AM1 method. There is a blue shift of the calculated spectra relative to the solution spectra and a slightly larger blue shift compared to the spectrum of the thin film. Both complexes are luminescent under near-UV excitation, and the spectra are in accord with the existence of a single emitting site in each. The increased quantum yield in the fluorinated complex is correlated with a decrease in the bipy-europium(III) distance, a closer match of the lowest ligand-centered triplet state (that level which is primarily responsible for the energy transfer from the ligands to the europium(III)), and the lower vibrational energy of the C-F bonds relative to the C-H bonds. In the fluorinated complex the calculations show that the lowest triplet level is primarily localized on the 2,2′-dipyridyI whereas in the nonfluorinated complex this is the second lowest triplet level.

Full text of DOI:10.1021/ic971602v

3542
Inorg. Chem. 1998, 37, 3542-3547
Synthesis, X-ray Structure, Spectroscopic Characterization, and Theoretical Prediction of
the Structure and Electronic Spectrum of Eu(btfa)3‚bipy and an Assessment of the Effect of
Fluorine as a â-Diketone Substituent on the Ligand-Metal Energy Transfer Process
He´lcio J. Batista,† Antoˆnio V. M. de Andrade,‡ Ricardo L. Longo,† Alfredo M. Simas,†
Gilberto F. de Sa´,*,† Nao K. Ito,§ and Larry C. Thompson§
Departamento de Qu´ımica Fundamental and Departamento de Engenharia Qu´ımica, Universidade
Federal de Pernambuco, CEP 50740-540, Recife, PE, Brazil, and Department of Chemistry,
University of MinnesotasDuluth, Duluth, Minnesota
ReceiVed December 26, 1997
The 2,2′-dipyridyl adducts of two europium â-diketonate complexes, Eu(btfa)3‚bipy [btfa ) 4,4,4-trifluoro-1-
phenyl-2,4-butanedione, bipy ) 2,2′-dipyridyl] and Eu(bzac)3‚bipy [bzac ) 1-phenyl-2,4-butanedione], have been
prepared. The crystal structure of the former with chemical formula EuC40H26O6N2F9 has been solved by single-
crystal X-ray diffraction methods. The complex crystallizes in the monoclinic space group P21/n with a ) 11.122-
(5) Å, b ) 22.860(8) Å, c ) 15.870(6) Å, â ) 102.62(3)o, V ) 3937(5) Å3, and Z ) 4. A single, eight-
coordinate environment, which approximates a square antiprism, is found for the europium(III). The UV absorption
spectra of both complexes were obtained from ethanol solutions and, in the case of Eu(btfa)3‚bipy, from a thin
film. In both cases the absorption spectra are reasonably well predicted by the INDO/S-CI method using, for
Eu(btfa)3‚bipy, both the X-ray data and that obtained through the SMLC/AM1 method as input geometry and, for
Eu(bzac)3‚bipy, that obtained through the SMLC/AM1 method. There is a blue shift of the calculated spectra
relative to the solution spectra and a slightly larger blue shift compared to the spectrum of the thin film. Both
complexes are luminescent under near-UV excitation, and the spectra are in accord with the existence of a single
emitting site in each. The increased quantum yield in the fluorinated complex is correlated with a decrease in the
bipy-europium(III) distance, a closer match of the lowest ligand-centered triplet state (that level which is primarily
responsible for the energy transfer from the ligands to the europium(III)), and the lower vibrational energy of the
C-F bonds relative to the C-H bonds. In the fluorinated complex the calculations show that the lowest triplet
level is primarily localized on the 2,2′-dipyridyl whereas in the nonfluorinated complex this is the second lowest
triplet level.
Introduction
reagents”, and during the 1970s a number of solid adducts were
studied by X-ray methods in order to obtain structural informa-
tion in support of various ideas relative to the mechanism of
this effect.4
The rare earth â-diketone complexes have had a long and
useful history.1 The first serious, intensive study of these
complexes was primarily involved in considering their use as
complexing agents for separations although their interesting
luminescence properties had been reported previously. For
several years beginning about 1960 there was an intense effort
directed toward finding suitable complexes that could be utilized
in lasers. Although much of this work involved the study of
materials of dubious purity and formulation, the period culmi-
nated in two definitive publications that demonstrated the types
of complexes that could be formed.2,3 The most useful of these
are the tris complexes, the tetrakis complexes, and, perhaps the
most useful of all, the adducts of the tris complexes. In solution
these latter are most commonly manifested in the so-called “shift
Since 1985 there has been renewed interest in these tris
adducts because of a number of characteristics that make them
quite useful and the fact that they can be used as precursors to
other materials.5 The number of types of molecules that can
form adducts with the tris-â-diketonates is quite large and
includes heterocyclic nitrogen donors such as pyridine, 2,2′-
dipyridyl, 1,10-phenanthroline, and 2,2′:6′,2′′-terpyridine as well
as oxygen donors such as water, dimethyl sulfoxide, dimethyl
formamide, and triphenylphosphine oxide, among others. In
general the complexes with europium(III) and terbium(III) are
strongly luminescent when excited with near-UV radiation, and
the energy transfer process is highly efficient. Moreover, these
adducts can generally be prepared in good purity, and it is often
possible to isolate X-ray quality crystals for structure determina-
tions. Although the luminescence spectra are definitive for the
identification of a particular compound, the low symmetry does
† Departamento de Qu´ımica Fundamental, Universidade Federal de
Pernambuco.
‡ Departamento de Engenharia Qu´ımica, Universidade Federal de Per-
nambuco.
§ Department of Chemistry, University of MinnesotasDuluth.
(1) Thompson, L. C. In Handbook on the Physics and Chemistry of Rare
Earths; Gschneider, K. A., Eyring, L., Eds.; North-Holland: Amster-
dam, 1979; Chapter 25.
(4) Sievers, R. E., Ed. Nuclear Magnetic Shift Reagents; Academic
Press: New York 1973.
(2) Melby, L. R.; Rose, N. J.; Abramson, E.; Caris, J. C. J. Am. Chem.
Soc. 1964, 86, 5117.
(3) Bauer, H.; Blanc, J.; Ross, D. L. J. Am. Chem. Soc. 1964, 86, 5125.
(5) Baxter, I.; Draker, S. R.; Hursthouse, M. B.; Abdyl Malik, K. M.;
McAleese, J.; Otway, D. J.; Plakatouras, J. C. Inorg. Chem. 1995, 34,
1384.
S0020-1669(97)01602-9 CCC: $15.00 © 1998 American Chemical Society
Published on Web 06/10/1998

Eu(btfa)3‚bipy Structure and Electronic Spectrum
Inorganic Chemistry, Vol. 37, No. 14, 1998 3543
not permit conclusions to be drawn about the geometry of the
complex.
spectroscopic-configuration interaction) method,20,21 has pro-
vided a valuable theoretical tool to study the effects of the
ligands on the luminescent properties. The energy levels and
transition moments, as well as other electronic properties
determined by these combined methodologies, have been used
to estimate the rate of energy transfer between the ligands and
the Ln(III) ion, allowing the calculation of the quantum yield
for the luminescence processes.22
The variety of â-diketones and adducting molecules that are
available permits the study of varying steric and electronic
effects on the structure, luminescence, and efficiency of
luminescence, which are of particular importance in the context
of connecting modern theoretical ideas to discrete complexes.
For example, the availability of complexes of known structure
has enabled the theories related to band intensities and crystal
field parameters to be tested and extended.6
The ultimate aim of our work is to use these powerful
predictive tools to design ligands which will form stable
complexes that will function as efficient light conversion
molecular devices (LCMD). Such complexes should have
strong ligand-centered absorption in the UV region, efficient
ligand to metal energy transfer rates, and intense metal-centered
emission in the visible range, red for Eu(III) or green for
Tb(III). Toward this end it is important to test further the utility
of the predictive theoretical methods on a series of compounds
in which it is possible to vary both steric and electronic effects.
As pointed out above, the adducts of the tris-â-diketone
complexes are an excellent choice for these tests and, conse-
quently, in this communication the synthesis, structure, and
absorption and luminescence spectra of Eu(btfa)3‚bipy are
reported together with the results of both types of calculations.
Similar calculations are made for the corresponding benzoyl-
acetone complex in order to compare the efficiency of emission
between the fluorinated and nonfluorinated compounds. This
fluorinated complex was chosen for these studies, in part,
because it is sufficiently volatile that it can be deposited as thin
films and is a strong candidate for a number of applications.
For example, it can be employed as an antireflection coating
(ARC) on a silicon solar cell with a resultant increase in cell
efficiency of approximately 21% or as a UV dosimeter.23
Given the importance of lanthanide complexes as shift
reagents and their subsequent development as contrast agents
for use in magnetic resonance imaging, several molecular
mechanics methods have been developed7-10 to predict the
molecular structure of such compounds. Despite the fact that
these methods are computationally efficient, allowing them to
be applied in molecular dynamics simulations,10 their param-
etrization is very demanding and specific, limiting their ap-
plication to selected groups of ligands. In addition, these
methods cannot provide reliable results for bond formation/
breaking in the ligand part of the complex nor of their electron
density. In order to solve some of these drawbacks a molecular
orbital model (sparkle model for lanthanide complexes based
on Austin model 1:SMLC/AM1)11,12 has been developed to
determine the molecular and electronic structure of lanthanide
compounds. This SMLC/AM1 method has proven to yield
excellent geometry for large europium compounds,11-14 includ-
ing cryptate and cagelike ligands.15 This approach has been
extended and generalized for other lanthanide ions16 as well as
for the prediction of properties other than geometry, such as
vibrational frequencies and intensities, enthalpy, entropy, heat
capacity, etc.17,18 In this semiempirical method the lanthanide-
(III) ion is replaced by a sparkle19 with charge +3, and thus
the method considers the ligand-lanthanide interaction as
essentially electrostatic, keeping, however, the excellent per-
formance of the AM1 method for the ligands and their
interactions.
Experimental Section
Reagents. The benzoyltrifluoroacetone, benzoylacetone, and 2,2′-
dipyridyl were obtained from Eastman Chemicals and were used without
further purification. The Eu(NO3)3‚6H2O (99.9% Eu) was purchased
from Rhoˆne-Poulenc Basic Chemicals Co. All solvents used were of
reagent quality.
The molecular structure determination is the first step in the
rationalization and prediction of the luminescent properties of
these lanthanide compounds. The combination of the SMLC/
AM1 method for obtaining molecular structure with semiem-
pirical methods for electronic spectra calculations, such as the
INDO/S-CI (intermediate neglect of differential overlap/
Preparation of Complexes. The syntheses of Eu(btfa)3‚bipy and
Eu(bzac)3‚bipy were accomplished by adding a stoichiometric quantity
of an ethanolic solution of Eu(NO3)3‚6H2O (1 mmol) dropwise with
stirring to an ethanolic solution containing the anion of the â-diketone
(3 mmol) (btfa or bzac, prepared by neutralization with an aqueous
solution of NaOH) and bipy (1 mmol) over a 2 h period. The
compounds precipitated and were separated by filtration and washed
with a small amount of ethanol. The white powders had melting points
of 193-4 °C [Eu(btfa)3‚bipy] and 173-4 °C [Eu(bzac)3‚bipy]. Chemi-
cal analysis of the fluorinated compound confirmed its formulation (C
calcd 50.58, found 50.26; H calcd 2.75, found 2.60; N calcd 2.94, found
2.92). Crystallization of Eu(btfa)3‚bipy from ethyl acetate/methanol
gave crystals suitable for X-ray structure determination. The melting
point and luminescence spectra of the powder and crystals were
identical.
(6) Malta, O. L.; Couto dos Santos, M. A.; Thompson, L. C.; Ito, N. K.
J. Lumin. 1996, 69, 77.
(7) Brecknell, D. J.; Raber, D. J.; Ferguson, D. M. J. Mol. Struct.
(THEOCHEM) 1985, 124, 343.
(8) Ferguson, D. M.; Raber, D. J. J. Comput. Chem. 1990, 11, 1061.
(9) Hay, B. P. Inorg. Chem. 1991, 30, 2876.
(10) Fossheim, R.; Dugstad, H.; Dahl, G. J. Med. Chem. 1991, 34, 819.
(11) de Andrade, A. V. M.; da Costa, N. V., Jr.; Simas, A. M.; de Sa´, G.
F. Chem. Phys. Lett. 1994, 227, 349.
(12) de Andrade, A. V. M.; da Costa, N. V., Jr.; Simas, A. M.; de Sa´, G.
F. J. Alloys Compd. 1995, 225, 55.
(13) de Andrade, A. V. M.; Longo, R. L.; Simas, A. M.; de Sa´, G. F. J.
Chem. Soc., Faraday Trans. 1996, 92, 1835.
(14) Batista, H. J.; de Andrade, A. V. M.; Longo, R. L.; Simas, A. M.; de
Sa´, G. F.; Thompson, L. C. J. Lumin. 1997, 72-74, 159.
(15) Longo, R. L. Presented at the 17th Brazilian Symposium of Theoretical
Chemistry, Caxambu, Nov 17-19, 1997.
(16) Benson, M. T.; Cundari, T. R.; Lutz, M. L.; Sommerer, S. O. In
ReViews in Computational Chemistry; Boyd, D., Lipkowitz, K., Eds.;
VCH: New York, 1996, Vol. 8, pp 145-202.
(17) de Andrade, A. V. M.; da Costa, N. V., Jr.; Simas, A. M.; de Sa´, G.
F. Submitted for publication.
(18) Rocha, G. B.; Simas, A. M. Presented at the 17th Brazilian Symposium
of Theoretical Chemistry, Caxambu, Nov 17-19, 1997.
(19) Stewart, J. J. J. P. MOPAC 93.00 Manual; Fujitsu Limited: Tokyo,
Japan, 1993; J. Comput.-Aided Mol. Des. 1990, 4, 1.
Spectral Measurements. Luminescence spectra were obtained on
solid samples at room and liquid nitrogen temperatures in Duluth with
a McPherson RS-10 spectrophotometer as described previously.24 In
(20) Ridley, J. E.; Zerner, M. C. Theor. Chim. Acta 1973, 32, 111; 1976,
42, 223.
(21) Zerner, M. C.; Loew, G. H.; Kirchner, R. F.; Mueller-Westerhoff, U.
T. J. Am. Chem. Soc. 1980, 102, 589.
(22) de Andrade, A. V. M.; da Costa, N. V., Jr.; Malta, O. L.; Longo, R.
L.; Simas, A. M.; de Sa´, G. F. J. Alloys Compd. 1997, 250, 412.
(23) (a) de Sa´, G. F.; Alves, S., Jr.; da Silva, E. F., Jr. Opt. Mater.
(Amsterdam), in press. (b) Gameiro, C. G.; da Silva, E. F., Jr.; Alves,
S., Jr.; de Sa´, G. F.; Santa-Cruz, P. A. Submitted to Materials Science
Forum.

3544 Inorganic Chemistry, Vol. 37, No. 14, 1998
Batista et al.
Table 1. Crystallographic Data for Eu(btfa)3‚bipy
empirical formula: EuC40H26O6N2F9 fw: 953.60
cryst syst: monoclinic
a ) 11.122(5) Å
b ) 22.860(8) Å
c ) 15.870(6) Å
â ) 102.62(3)°
V ) 3937(5) Å3
Z ) 4
space group: P21/n (No. 14)
T ) 24 ( 1 °C
λ ) Mo KR (0.710 69 Å)
F
calcd ) 1.609 g/cm3
µ ) 16.81 mm-1
R ) 0.052a
Rw ) 0.052b
2
a R ) ∑(|Fo| - |Fc|)/∑|Fo|. b Rw ) [(∑w(|Fo| - |Fc|)2/∑w|Fo| )]1/2
.
Recife, measurements of luminescence were made at both temperatures
on a Jobin Yvon U1000 spectrophotometer and measurements of
lifetime were made as described previously.25 Absorption spectra in
the UV region were obtained from ethanol solutions of the complexes
(≈10-4-10-5 M) with a Perkin-Elmer Lambda 6 spectrophotometer.
A UV spectrum of the fluorinated complex was also obtained in the
solid state from a film of 400 Å thickness that was prepared by the
thermal deposition of the complex on a quartz disk that had been
previously cleaned and degreased to ensure proper film adhesion. The
film was prepared by thermally evaporating the compound from an
alumina crucible onto the substrate and defined by photolithography
to form a rectangular structure with an area of about 1 cm2. The
thickness of the film was monitored during deposition by a quartz crystal
thickness meter and, after deposition, ellipsometry measurements at
several wavelengths to ensure film quality and uniformity. The
luminescence spectrum and lifetime of the film were the same as those
of analytically pure samples of the complex.
Structure Determination. The determination of the structure was
carried out in the X-ray Crystallography Laboratory of the University
of MinnesotasTwin Cities. Crystal data are summarized in Table 1,
and full details of the structure determination are given in the Supporting
Information. All intensity data were collected on an Enraf-Nonius
CAD4 diffractometer using monochromated Mo KR radiation. The
intensities of three representative reflections that were measured after
each 80 min period of exposure remained constant, indicating crystal
and electronic stability. The structure was solved by direct methods
as described previously24 using the programs MITHRIL26 and DIRDIF,27
and all calculations were performed with the TEXSAN crystallographic
software package of Molecular Structure Corporation.
Theoretical Calculations. The ground state geometries of both the
fluorinated and nonfluorinated complexes were obtained using the
SMLC/AM1 method with an improved parametrization.17 In this
model, the lanthanide ion is represented, within AM1,28 by a sparkle
which is a point of charge +3e in the center of a repulsive spherical
potential of the form exp(-Rr). As such, the sparkle model simulates
well the essentially electrostatic interaction between the lanthanide ion
and the ligands and the geometry of the entire complex (sparkle plus
ligands) is optimized. In the first parametrization of the SMLC/AM1
model,11-13 chemometric approaches, such as principal component
analysis and factorial analysis, were used to obtain the optimum values
for the parameter R (ALPAM1) and also for the monopole-monopole
interaction parameter (AMAM1) involved in the core-core repulsion
integrals. However, to improve the model and at the same time keep
it consistent with the AM1 Hamiltonian, SMLC/AM1 has been
generalized so that Gaussian functions similar to the ones used in AM1
to improve the core-core interaction were introduced and all six
Gaussian parameters were optimized for several europium(III) com-
pounds.17 It should be noted that a proper description of complexes
with cryptate and cagelike ligands was obtained only with the latter
parametrization.
Figure 1. ORTEP drawing of Eu(btfa)3‚bipy (50.0% probability
ellipsoids).
stable solution is considered the appropriate structure. When the
crystallographic structure is known for a similar compound, it is used
as the initial guess for the geometry optimization procedure.
The energy levels and transition moments were calculated with the
INDO/S-CI method20,21 implemented within the ZINDO program29
allowing the comparison and prediction of the electronic absorption
spectra. The molecular structure used was either from SMLC/AM1
model version 2 or from X-ray crystallography where the sparkle or
the metal ion is replaced by a point of charge +3e.13 Only the singly
excited configurations (CIS) with respect to the closed shell reference
are included in the CI matrix. These configurations are generated by
all single substitutions, within a chosen set of occupied and unoccupied
orbitals, on the reference determinant. The choice of this orbital set
used to generate the CI configurations is made by saturating the CIS
space in a given range of energy. That is, a preliminary electronic
spectrum is generated from a small set of orbitals. This configuration
space is then gradually increased until there are no visual differences
between the spectra generated by two consecutive sets of configurations.
The range of energy chosen for performing this visual comparison
corresponds to 200-600 nm. Once the configuration space is selected
for one complex or a set of ligands, it is used for similar complexes
and/or ligands. In order to simulate the experimental electronic
absorption spectra it is necessary to take into account the line broadening
of the calculated transition energies and moments. This can be
accomplished by considering that binary collisions are the main reason
for this broadening in the condensed phase. Assuming that the
collisions are strong and short enough that the Boltzmann distribution
is restored after the collision, the Van Vleck-Weisskopf line shape
function can be obtained,30 of which the Lorentzian form is a particular
case.31 It should be noted that for frequencies in the UV-vis region
there is no practical distinction between the Van Vleck-Weisskopf
and Lorentzian line shapes. The calculated transition energies and
oscillator strengths are then fitted and normalized to a Lorentzian line
shape with a given half-height bandwidth (28 nm, in the present case).
It should be noted that when the line shape function is transformed
from the frequency to the wavelength domain, the intensities (oscillator
strengths) become explicitly frequency dependent and the relative
intensities can be very different when comparing the spectrum in these
domains.
Results
Experimental Results. The compound Eu(btfa)3‚bipy crys-
tallizes in a monoclinic space group (P21/n) with four molecules
in the unit cell. As expected, the complex is eight-coordinate
with all six â-diketone oxygen atoms and the two nitrogen atoms
from the dipyridyl bound to the metal (Figure 1). The average
For the determination of the molecular structure of a new compound,
several initial guesses for the starting geometry are used and the most
(24) Holtz, R. C.; Thompson, L. C. Inorg. Chem. 1993, 32, 5251.
(25) de Sa´, G. F.; de Azevedo, W. M.; Gomes, A. S. L. J. Chem. Res.
1994, 234.
(26) Gilmore, C. J. J. Appl. Crystallogr. 1984, 17, 42.
(27) Beurskens, P. T. Technical Report 1984/1, Crystallography Laboratory,
Toernooiveld, 6525 Ed.: Nijmegen, The Netherlands.
(28) Dewar, M. J. S.; Zoebish, E. G.; Healy E. F.; Stewart, J. J. P. J. Am.
Chem. Soc. 1985, 107, 3902.
(29) Zerner, M. C. ZINDO Manual; QTPsUniversity of Florida: Gaines-
ville, FL, 1990.
(30) Van Vleck, J. H.; Weisskopf, V. F. ReV. Mod. Phys. 1945, 17, 227.
(31) Townes, C. H.; Schawlow, A. L. MicrowaVe Spectroscopy; Dover:
New York, 1975.

Eu(btfa)3‚bipy Structure and Electronic Spectrum
Inorganic Chemistry, Vol. 37, No. 14, 1998 3545
Figure 4. Luminescence spectrum of Eu(btfa)3‚bipy at 77 K upon
ligand excitation (370 nm). The labels refer to the J values of the final
level of the emission transition D0 f FJ. The two insets show the
5D0 f 7F0,1 and the 5D0 f 7F3,4 regions magnified but not in the same
scale.
Figure 2. Calculated electronic absorption spectrum of Eu(btfa)3‚bipy
(SMLC input geometry) (dotted line); ethanolic solution (solid line);
thin film on a quartz disk (dashed line).
5
7
Table 2. Interatomic Distances and Angles between the Atoms of
the Coordination Polyhedron and the Eu(III) Ion for the
Eu(btfa)3‚bipy (FL) and Eu(bzac)3‚bipy (NFL)
calcd
exptl
FL
FL
NFL
Distances (Å)
2.369(5)
2.351(5)
2.399(5)
2.349(6)
2.369(5)
2.321(5)
2.586(6)
2.577(5)
Eu-O1A
Eu-O2A
Eu-O1B
Eu-O2B
Eu-O1C
Eu-O2C
Eu-N1D
Eu-N12D
2.407
2.388
2.362
2.431
2.350
2.388
2.560
2.531
2.383
2.357
2.353
2.406
2.354
2.376
2.633
2.582
Figure 3. Calculated electronic spectrum of Eu(bzac)3‚bipy (dotted
line); ethanolic solution (solid line).
Eu-O distance of 2.360(19) Å and the average Eu-N distance
of 2.582(5) Å are typical for complexes of this type. The CF3
groups are commonly disordered in substituted â-diketone
complexes, and such is the case in this complex. However,
this did not affect the solution of the structure. Visual inspection
suggests that the coordination polyhedron is a square antiprism.
This was confirmed by an analysis using the method developed
by Porai-Koshits and Aslanov,32 although there are significant
deviations from the ideal values. The two “square” planes are
defined by O2C, O1C, O2A, and O1A and by O1B, O2B,
N12D, and N1D, respectively.
The electronic absorption spectra of the fluorinated complex
in the region 200-400 nm, both in ethanol solution and as a
thin film, are shown in Figure 2, and that of the nonfluorinated
complex in ethanol solution is shown in Figure 3. The most
intense peak occurs in all three cases between 325 and 350 nm.
The calculated gas phase spectrum for each complex is also
shown in Figure 2 and Figure 3. The spectra are normalized
to the same arbitrary scale, but the molar absorptivities are 2.6
× 104 l mol-1 cm-1 for the fluorinated complex and 1.2 × 104
l mol-1 cm-1 for the nonfluorinated complex.
Angles (deg)
71.1(2)
140.0(2)
141.3(2)
123.9(2)
79.4(2)
71.3(2)
76.4(2)
148.8(2)
82.8(2)
78.8(2)
115.2(2)
133.6(2)
83.0(2)
71.4(2)
78.3(2)
77.0(2)
74.0(2)
104.6(2)
75.8(2)
138.7(2)
111.9(2)
72.4(2)
72.3(2)
146.4(2)
145.0(2)
83.1(2)
142.7(2)
62.5(2)
O1A-Eu-O2A
O1A-Eu-O1B
O1A-Eu-O2B
O1A-Eu-O1C
O1A-Eu-O2C
O1A-Eu-N1D
O1A-Eu-N12D
O2A-Eu-O1B
O2A-Eu-O2B
O2A-Eu-O1C
O2A-Eu-O2C
O2A-Eu-N1D
O2A-Eu-N12D
O1B-Eu-O2B
O1B-Eu-O1C
O1B-Eu-O2C
O1B-Eu-N1D
O1B-Eu-N12D
O2B-Eu-O1C
O2B-Eu-O2C
O2B-Eu-N1D
O2B-Eu-N12D
O1C-Eu-O2C
O1C-Eu-N1D
O1C-Eu-N12D
O2C-Eu-N1D
O2C-Eu-N12D
N1D-Eu-N12D
65.8
66.5
144.9
141.1
117.5
82.5
73.0
78.5
149.2
85.7
82.3
118.2
131.8
81.6
66.2
81.6
78.6
74.4
99.8
81.9
135.9
114.2
71.3
145.6
134.9
113.8
81.2
73.1
74.9
146.2
80.3
80.9
120.3
132.6
81.8
67.4
88.4
83.9
73.2
96.1
88.7
143.6
114.8
70.7
The complex Eu(btfa)3‚bipy is very strongly luminescent in
the red when excited with near-ultraviolet radiation, and
transitions were measured from the first excited level to the
first five components of the ground term (Figure 4). The main
66.9
67.8
141.0
149.6
78.5
143.4
66.2
139.8
155.1
74.8
137.0
64.3
emission occurs, as expected, in the 5D0
Although the D0 f F0 and D0 f F1 transitions also are
f
7F2 transition.
5
7
5
7
present, they are 50 and 10 times less intense, respectively. The
geometry was used as the input to the SMLC/AM1 method.
The output geometry conformed very well to the experimental
geometry, and the values determined for the Eu-O and Eu-N
distances along with the angles about the europium(III) ion are
presented in Table 2. The average predicted values for the
Eu-O and Eu-N distances are 2.38 and 2.54 Å, and the average
deviation of all eight values from the experimental values in
the solid state is 0.05 Å. The data for the bond angles show a
similarly good agreement. No experimental structure is avail-
able for Eu(bzac)3‚bipy, and its ground state geometry was
predicted by means of the SMLC/AM1 method using the
geometry of the fluorinated complex as the input. The two
structures predicted by the SMLC/AM1 method are compared
7
5D0 f F0 transition is a singlet, which indicates the presence
of only a single emitting species, and the degeneracy of the J
) 1 and J ) 2 levels is completely removed, in agreement with
the structural results. There is no indication of emission from
5
the higher excited levels such as D1. The emission spectrum
5
of Eu(bzac)3‚bipy is similar except that the transitions D0 f
5
7
7F0 and D0 f F1 are now both about 10 times less intense
5
7
than the D0 f F2 transition.
Theoretical Results and Discussion. Because the structure
of Eu(btfa)3‚bipy had been determined by X-ray diffraction, this
(32) Porai-Koshits, M. A.; Aslanov, L. A. J. Struct. Chem. 1972, 13, 244.

3546 Inorganic Chemistry, Vol. 37, No. 14, 1998
Batista et al.
also in Table 2, and it can be seen that they are very similar,
suggesting that the effect of fluorination on the ground state
structure is minimal. With the exception of the Eu-O3A
distance, the Eu-O distances are slightly shorter in the
nonfluorinated complex whereas the Eu-N distances are
calculated to be slightly longer. This slightly shorter Eu-N
distance in compounds containing the CF3 group, compared to
compounds that do not, has been observed in the X-ray
structures of 10 related molecules of this same type, six of which
contain the CF3 group.33 As can be seen in Table 2, this
shortening of the Eu-N distance in Eu(btfa)3‚bipy is 0.05-
0.08 Å whereas the lengthening of the Eu-O distances is ∼
0.02 Å.
atomic compositions. Moreover, fluorination seems to lower
the triplet state energies overall, of an amount of approximately
1000 cm-1, yielding donor states in closer resonance with the
metal ion acceptor states (5D2: 21 483 cm-1); see the Sup-
porting Information for a complete table of energy levels.
According to the Fermi golden rule the energy transfer rate
(WET) can be written in the form35
2π
p
2π
p
2
2
WET
)
| Ψfφf|H|Ψiφi | F ) |Hda| F
where the donor-acceptor matrix element (Hda) has the fol-
lowing dependence with the donor-acceptor distance (RL):
The UV absorption spectra of the fluorinated and nonfluori-
nated complexes calculated via INDO/S-CI (Figures 2 and 3)
are in satisfactory agreement with the experimental spectra
obtained from ethanol solution. It has been shown previously14
that the use of the calculated molecular structure (SMLC/AM1)
as the input geometry for the spectroscopic calculation yields
better agreement with the experimental spectra than the spectra
calculated with the crystallographic structure. Both qualitative
(number of maxima and shoulders) and quantitative features of
the absorption spectra are reproduced, and a systematic blue
shift of 20 nm in the calculated spectra is found when compared
to the experimental one. This shift might be due to several
effects, but mainly from the neglect of the solvent effects in
the calculated spectra.
Also in Figure 2 is the absorption spectrum obtained from
the thin film. This is quite similar to the solution spectrum
and, thus, to that calculated from the SMLC/AM1 geometry.
The number of peaks and their relative intensities are again
reproduced fairly well. The experimental spectrum is shifted
toward the red relative to that of the isolated gaseous molecule
even more than in the case of the solvated molecules. This
additional crystal lattice effect is quite reasonable. It would be
desirable to have the spectrum of the molecule in the gas phase
for direct comparison with the theoretical spectrum, but this is
not presently available. The volatility of some adducts of the
tris-â-diketonates has made it possible to determine their gas
phase luminescence spectra, and in the future we hope to be
able to measure the corresponding absorption spectra of
molecules of this type.
The fluorinated complex is visually considerably more
luminescent than the nonfluorinated complex. This is quanti-
tatively reflected in the quantum yield of 65% for Eu(btfa)3‚
bipy, both in the solid state and in the thin film, compared to
16% for solid Eu(bzac)3‚bipy. Although our ultimate intention
is to analyze the quantum yields in these compounds using the
method described in ref 34 we can make a number of qualitative
observations from the calculated spectra that are in agreement
with the experimental values.
The origin of the triplet donor states of the ligands is very
important, because it can have a profound influence on the
energy transfer rate. The lowest triplet state in the fluorinated
complex (23 938 cm-1) is basically centered at the bipy ligand
whereas in the nonfluorinated complex this (25 260 cm-1) is
the second lowest level. Both of these levels have similar
atomic contributions. In both complexes, almost all of the other
triplet states are due to the â-diketone ligands and have similar
1
2
|Hda|
l+2 2
(RL
)
with l varying from 1 up to the highest angular momentum value
allowed by the selection rules for the 4f electrons, usually 6.
The energy and temperature dependence is included in the F
term, which contains a summation over the Franck-Condon
factor and the appropriate energy mismatch conditions. As-
suming that the ligand bandwidth (γL) is much larger than that
for the rare earth ion, it can be shown35 that
F )
exp -
2 ln 2
1
ln 2
π
∆
( )
γ x
[
]
γL
L
where ∆ ) EL - ERE is the energy difference between the donor
and acceptor band peaks.
Upon fluorination, the donor-acceptor distance changes
(decreases approximately 0.07 Å). The ratio between the
electronic factors of the fluorinated (FL) and nonfluorinated (NF)
complexes can then be written approximately as
2
2l+4
|Hda|FL
RNLF
≈
2
FL
[
]
|Hda|NF
RL
For a typical value of 5 Å for the bipy-Ln(III) distance in the
nonfluorinated complex, the same distance in the fluorinated
complex is shortened to 4.93 Å, and this ratio is 1.09 and 1.12
for l ) 1 and 2, respectively. Similarly, the ratio for the
mismatch energy factor is given by
F
ln 2
FL ) exp
[(∆NF)2 - (∆FL)2]
2
FNF
[
]
γL
which for a typical value of 5000 cm-1 for γL yields 1.41 and
1.21 for the lowest triplet donor state to the 5D1 and 5D2 acceptor
states, respectively. The ratio between the energy transfer rates
should then be in the range 1.32-1.58, showing that those two
effects, namely, lowering the donor state energy and shortening
the donor-acceptor distance, are important in explaining the
significant increase in the luminescence quantum yield (16%
to 65%) when the CH3 groups are replaced by CF3 in the
europium complex.
It is well accepted that the nonradiative deactivation of the
ligand and metal ion excited states by vibronic coupling with
O-H modes is effective in quenching the fluorescence, thus
(33) Thompson, L. C.; Young, V. G.; Berry, S.; Atchison, F. W.; Maser,
D. Unpublished results.
(34) (a) de Mello Donega´, C.; Ribeiro, S. J. L.; Gonc¸alves, R. R.; Blasse,
G. J. Phys. Chem. Solids 1996, 57, 1727. (b) de Mello Donega´, C.;
Alves, S., Jr.; de Sa´, G. F. J. Chem. Soc., Chem. Commun. 1996, 1199.
(35) Malta, O. L. J. Lumin. 1997, 71, 229.

Eu(btfa)3‚bipy Structure and Electronic Spectrum
Inorganic Chemistry, Vol. 37, No. 14, 1998 3547
decreasing the quantum yield. It has been proposed36 that the
vibronic coupling with C-H modes from CH3 groups would
also be effective, because the C-H stretching mode has a
vibrational frequency close to that of the O-H. Therefore, it
is very likely that the fluorination of the CH3 groups will lead
to a less effective vibronic coupling because of the significant
difference between the vibrational frequencies of the C-F and
C-H stretching modes.
of the significant increase of the luminescence quantum yield
when the CH3 groups are replaced by CF3 groups in the
Eu(bzac)3‚bipy complex. Indeed, there seem to be three factors
acting in concert to increase the efficiency of the luminescence
process, namely, a decrease in the donor-acceptor energy
difference, a decrease in the donor-acceptor distance, and a
decrease in the vibronic coupling with the C-F vibrational
modes.
Conclusions
Acknowledgment. The authors are grateful to CNPq,
PADCT, CAPES, FACEPE, and FINEP (Brazilian agencies)
and to the Graduate School of the University of Minnesota for
financial support, to Prof. Oscar Malta for useful suggestions,
to Prof. Eronides Felisberto for the preparation of the thin film
sample, to Mr. Severino Alves for spectroscopic measurements,
and to Prof. Doyle Britton for the structure determination.
The complex Eu(btfa)3‚bipy has been shown by X-ray
diffraction to consist of a single species in agreement with the
presence of a single 5D0
f
7F0 line in the luminescence
spectrum. The sparkle model, SMLC/AM1, gives a reasonably
good prediction of the structure of the molecule, taking into
account the differences between the gas and solid phases.
Moreover, the method yields a similar structure for the
analogous molecule Eu(bzac)3‚bipy. For both complexes the
UV absorption spectrum determined by means of the INDO/
S-CI technique is in satisfactory agreement with experiment.
The theoretical calculations of the molecular structure and
the electronic energy levels have cast some light onto the origin
Supporting Information Available: Text describing data collec-
tion, data reduction, and structure solution and refinements and tables
listing crystal data, intensity measurements, structure refinement details,
positional and thermal parameters, general temperature factor expres-
sions, bond distances, bond angles, torsional angles, intermolecula
contacts, least-squares planes, and theoretical electronic energy levels
for the complexes (45 pages). Ordering information is given on any
current masthead page.
(36) (a) Bu¨nzli, J.-C. G. In Lanthanide Probes in Life, Chemical and Earth
Sciences. Theory and Practice; Bu¨nzli, J.-C. G., Choppin, G. R., Eds.;
Elsevier-North Holland: Amsterdam, 1989; Chapter 7. (b) Wang, Z.-
M.; Choppin, G. R.; di Bernardo, P.; Zanonato, P.-L.; Portanova, R.;
Tollazzi, M. J. Chem. Soc., Dalton Trans. 1993, 2791.
IC971602V