Highly luminescent Cu(I)-phenanthroline complexes in rigid matrix and temperature dependence of the photophysical properties

Article abstract of DOI:10.1021/ja0043439

We synthesized new [Cu(NN)₂]+-type complexes where NN = 2-5 and denotes a 2,9-disubstituted-1,10-phenanthroline ligand (related complexes of 1 and 6 ligands are used for reference purposes). For 2, 3, and 4 the ligand substituents are long alkyl-type fragments, whereas in 5 a phenyl ring is directly attached to the chelating unit. At 298 K the four complexes display relatively intense metal-to-ligand-charge-transfer (MLCT) emission bands with maxima around 720 nm, Φ_em ≈ 1 × 10^-3 and τ > 100 ns in deaerated CH₂Cl₂. The emission behavior at 77 K in a CH₂Cl₂/MeOH matrix is quite different for complexes of alkyl- (2-4) versus phenyl-substituted (5) ligands. The former exhibit very intense emission bands centered around 642 nm and hypsochromically shifted with respect to 298 K, whereas the luminescence band of [Cu(5)₂]⁺ is faint and shifted toward the infrared side. These results prompted us to study in detail the temperature dependence of luminescence properties of [Cu(2)₂]⁺ and [Cu(5)₂]⁺ in the 300-96 K range. For both complexes the excited state lifetimes increase monotonically by decreasing temperatures, and the trend is well described by an Arrhenius-type treatment involving two equilibrated MLCT excited levels. The emission bands show a similar behavior for the two compounds (intensity decrease and red-shift) only in the 300-120 K range, when the solvent is fluid. In the frozen regime (T ≤ 120 K), the emission intensity of [Cu(5)₂]⁺ continues to drop, whereas that of [Cu(2)₂]⁺ exhibits a dramatic intensity increase. We interpret this different behavior in terms of structural factors, suggesting that long alkyl-chains in the 2,9-phenanthroline positions are optimal to prevent significant ground- and excited-state distortions in rigid matrix. We show that our results do not contradict current models describing the photophysics of [Cu(NN)₂]⁺ but, instead, bring further evidence to support their validity. They also suggest guidelines for the design of Cu(I)-phenanthroline complexes showing optimized luminescence performances both in fluid and in rigid matrix, an elusive goal for over two decades.

Full text of DOI:10.1021/ja0043439

J. Am. Chem. Soc. 2001, 123, 6291-6299
6291
Highly Luminescent Cu(I)-Phenanthroline Complexes in Rigid
Matrix and Temperature Dependence of the Photophysical Properties
Delphine Felder,‡ Jean-Franc¸ois Nierengarten,*,‡ Francesco Barigelletti,*,†
Barbara Ventura,† and Nicola Armaroli*,†
Contribution from Institut de Physique et Chimie des Mate´riaux de Strasbourg, Groupe des Mate´riaux
Organiques, UniVersite´ Louis Pasteur et CNRS, 23 rue du Loess, 67037 Strasbourg, France, and
Istituto di Fotochimica e Radiazioni d’Alta Energia del CNR, Via P. Gobetti 101, 40129 Bologna, Italy
ReceiVed December 27, 2000
Abstract: We synthesized new [Cu(NN)2]+-type complexes where NN ) 2-5 and denotes a 2,9-disubstituted-
1,10-phenanthroline ligand (related complexes of 1 and 6 ligands are used for reference purposes). For 2, 3,
and 4 the ligand substituents are long alkyl-type fragments, whereas in 5 a phenyl ring is directly attached to
the chelating unit. At 298 K the four complexes display relatively intense metal-to-ligand-charge-transfer (MLCT)
emission bands with maxima around 720 nm, Φem ≈ 1 × 10-3 and τ > 100 ns in deaerated CH2Cl2. The
emission behavior at 77 K in a CH2Cl2/MeOH matrix is quite different for complexes of alkyl- (2-4) versus
phenyl-substituted (5) ligands. The former exhibit very intense emission bands centered around 642 nm and
hypsochromically shifted with respect to 298 K, whereas the luminescence band of [Cu(5)2]+ is faint and
shifted toward the infrared side. These results prompted us to study in detail the temperature dependence of
luminescence properties of [Cu(2)2]+ and [Cu(5)2]+ in the 300-96 K range. For both complexes the excited
state lifetimes increase monotonically by decreasing temperatures, and the trend is well described by an
Arrhenius-type treatment involving two equilibrated MLCT excited levels. The emission bands show a similar
behavior for the two compounds (intensity decrease and red-shift) only in the 300-120 K range, when the
solvent is fluid. In the frozen regime (T e 120 K), the emission intensity of [Cu(5)2]+ continues to drop,
whereas that of [Cu(2)2]+ exhibits a dramatic intensity increase. We interpret this different behavior in terms
of structural factors, suggesting that long alkyl-chains in the 2,9-phenanthroline positions are optimal to prevent
significant ground- and excited-state distortions in rigid matrix. We show that our results do not contradict
current models describing the photophysics of [Cu(NN)2]+ but, instead, bring further evidence to support their
validity. They also suggest guidelines for the design of Cu(I)-phenanthroline complexes showing optimized
luminescence performances both in fluid and in rigid matrix, an elusive goal for over two decades.
Introduction
sensing abilities based on luminescence.12 In general, in view
of practical uses, some requirements related to the capability
of light absorption and formation of excited states should be
met.13 Indeed, while several Cu(I)-phenanthrolines with strong
absorption in the visible spectral region (vis) have been
prepared,1,2 long-lived (τ > 100 ns) and luminescent excited
states have not been always observed. Reasons have mainly to
do with the fact that in [Cu(NN)]2+-type complexes, the lowest-
lying excited levels are of metal-to-ligand-charge-transfer
(MLCT) nature14 and their formation is accompanied by
extended structural rearrangements.3 These can be described in
terms of a release of the ground-state pseudotetrahedral geom-
etry (D2d), to a more open geometry, mainly occurring via
“flattening” and “rocking” distortions.15 In this way, the
thermally equilibrated MLCT excited states16 undergo a con-
The investigation of the luminescence properties of bisphenan-
throline complexes of copper(I), herein abbreviated [Cu(NN)2]+,
has been the focus of many research reports that have been
recently reviewed.1-3 The underlying motivation for this type
of study has to do with possible practical applications, for
instance in the field of energy conversion and storage.4-7 In
the area of inorganic photochemistry other metal complexes,
for instance those of the ruthenium(II)-polypyridine family,8
have attracted a great deal of attention since they proved to be
very useful for the design of excitation energy transduction
schemes and storage9-11 or for the development of various
‡ IPCMS, Strasbourg.
† Istituto FRAE/CNR, Bologna. E-mail: armaroli@frae.bo.cnr.it. Fax
+39-051-6399844.
(1) Armaroli, N. Chem. Soc. ReV. 2001, 30, 113.
(2) Scaltrito, D. V.; Thompson, D. W.; O’Callaghan, J. A.; Meyer, G. J.
Coord. Chem. ReV. 2000, 208, 243.
(9) Barigelletti, F.; Flamigni, L. Chem. Soc. ReV. 2000, 29, 1.
(10) Harriman, A.; Ziessel, R. Chem. Commun. 1996, 1707.
(11) Balzani, V.; Juris, A.; Venturi, M.; Campagna, S.; Serroni, S. Chem.
ReV. 1996, 96, 759.
(12) For a series of papers on luminescent chemical sensors based on
transition metal complexes, see: Coord. Chem. ReV. 2000, Vol. 205.
(13) Roundhill, D. M. Photochemistry and Photophysics of Metal
Complexes; Plenum Press: New York, 1994.
(3) McMillin, D. R.; McNett, K. M. Chem. ReV. 1998, 98, 1201.
(4) Bignozzi, C. A.; Argazzi, R.; Kleverlaan, C. J. Chem. Soc. ReV. 2000,
29, 87.
(5) Schwab, P. F. H.; Levin, M. D.; Michl, J. Chem. ReV. 1999, 99, 1863.
(6) Balzani, V.; Campagna, S.; Denti, G.; Juris, A.; Serroni, S.; Venturi,
M. Acc. Chem. Res. 1998, 31, 26.
(7) Breddels, P. A.; Berdowski, P. A. M.; Blasse, G. Recl. TraV. Chim.
Pays-Bas 1981, 100, 439.
(14) McMillin, D. R.; Buckner, M. T.; Ahn, B. T. Inorg. Chem. 1977,
16, 943.
(8) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; von
Zelewsky, A. Coord. Chem. ReV. 1988, 84, 85.
(15) Dobson, J. F.; Green, B. E.; Healy, P. C.; Kennard, C. H. L.;
Pakawatchai, C.; White, A. H. Aust. J. Chem. 1984, 37, 649.
10.1021/ja0043439 CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/08/2001

6292 J. Am. Chem. Soc., Vol. 123, No. 26, 2001
Felder et al.
Chart 1
sistent energetic stabilization, which is facilitated by an easy
access of the donating solvent toward the formally d9 Cu(II)
center.17 Relevant consequences, nowadays well documented
by the investigations of McMillin,3,18-25 Karpishin,26-29 G. J.
Meyer,2,30-32 and others,33-38 are related to the fact that
deactivation of the excited level is subject to strongly enhanced
nonradiative paths via effects concerned with the “energy-gap
law”39-41 or even with direct quenching by solvent molecules.17
The strongly appealing possibility of using cheap copper(I)
complexes for replacing the more expensive compounds based
on ruthenium(II) or other metal ions, along with the need for a
deeper understanding of the correlation between structural
processes and photophysical properties, have pushed a continu-
ous progress in the design of photoluminescent [Cu(NN)2]+
complexes.1 In these, a common motif has been that of using
appended units to the phenanthroline frame so as to block as
much as possible the excited-state distortions. Nowadays,
complexes are available that feature remarkable luminescence
properties. For instance, [Cu(dbtmp)2]+ exhibits τ ) 920 ns and
φ ) 0.0062 in degassed CH2Cl2 (dbtmp ) 2,9-dibutyl-3,4,7,8-
tetramethyl-1,10-phenathroline).24 Here, steric effects exerted
by n-butyl groups appended at 2,9 and buttressing methyl groups
at 3, 4, 7, and 8 phenanthroline positions cooperate to effectively
block the excited state close to the ground-state geometry. A
highly emissive compound where blocking of the geometry was
also successfully obtained is the heteroleptic [Cu(dbp)(dmp)]+
complex where τ ) 730 ns and φ ) 0.01 in degassed CH2Cl2
(dbp ) 2,9-di-tert-butyl-1,10-phenanthroline, dmp ) 2,9-
dimethyl-1,10-phenanthroline).29 A further approach for block-
ing structural rearrangements at the excited state has been based
on the use of aryl groups appended to the phen unit.22 In this
case, one exploits interligand π-staking interactions (involving
the aryl group of one ligand and a ring of the opposite ligand)
which cause deviations from the D2d symmetry already for the
ground state. Remarkably, upon light absorption the excited-
state rearrangement is somewhat restrained as an effect of this
type of interactions.
In line with the above-mentioned approaches, we have
prepared the [Cu(NN)2]+ complexes of the ligands 1-5 il-
lustrated in Chart 1, and we have then studied their photophysi-
cal properties in detail. In some of these compounds the basic
photoactive [Cu(phen)2]+ unit is equipped with alkyl chains that
might be long enough to wrap it. We have thus observed good
luminescence properties at room temperature, as a consequence
of geometric effects leading to protection of the cation center
from interactions with the environment.17 Both alkyl- ([Cu(2)2]+,
[Cu(3)2]+, [Cu(4)2]+) and aryl-phenanthroline ([Cu(5)2]+) cases
have been worked out. The complexes [Cu(1)2]+1 and [Cu-
(6)2]+,34 have been employed for reference purposes. On the
basis of the well-known two-level model for the lowest MLCT
excited states of [Cu(NN)2]+ (vide infra),16,19 a detailed inves-
tigation of the temperature dependence of the luminescence for
two representative cases, [Cu(2)2]+ and [Cu(5)2]+, has been
(16) Everly, R. M.; McMillin, D. R. J. Phys. Chem. 1991, 95, 9071.
(17) Everly, R. M.; McMillin, D. R. Photochem. Photobiol. 1989, 50,
711.
(18) Buckner, M. T.; McMillin, D. R. J. Chem. Soc., Chem. Commun.
1978, 759.
(19) Kirchhoff, J. R.; Gamache, R. E.; Blaskie, M. W.; Del Paggio, A.
A.; Lengel, R. K.; McMillin, D. R. Inorg. Chem. 1983, 22, 2380.
(20) Phifer, C. C.; McMillin, D. R. Inorg. Chem. 1986, 25, 1329.
(21) Ichinaga, A. K.; Kirchhoff, J. R.; McMillin, D. R.; Dietrich-
buchecker, C. O.; Marnot, P. A.; Sauvage, J. P. Inorg. Chem. 1987, 26,
4290.
(22) Gushurst, A. K. I.; McMillin, D. R.; Dietrich-Buchecker, C. O.;
Sauvage, J. P. Inorg. Chem. 1989, 28, 4070.
(23) Eggleston, M. K.; McMillin, D. R.; Koenig, K. S.; Pallenberg, A.
J. Inorg. Chem. 1997, 36, 172.
(24) Cunningham, C. T.; Cunningham, K. L. H.; Michalec, J. F.;
McMillin, D. R. Inorg. Chem. 1999, 38, 4388.
(25) Cunningham, C. T.; Moore, J. J.; Cunningham, K. L. H.; Fanwick,
P. E.; McMillin, D. R. Inorg. Chem. 2000, 39, 3638.
(26) Miller, M. T.; Gantzel, P. K.; Karpishin, T. B. Angew. Chem., Int.
Ed. 1998, 37, 1556.
(27) Miller, M. T.; Karpishin, T. B. Inorg. Chem. 1999, 38, 5246.
(28) Miller, M. T.; Gantzel, P. K.; Karpishin, T. B. Inorg. Chem. 1999,
38, 3414.
(29) Miller, M. T.; Gantzel, P. K.; Karpishin, T. B. J. Am. Chem. Soc.
1999, 121, 4292.
(30) Castellano, F. N.; Ruthkosky, M.; Meyer, G. J. Inorg. Chem. 1995,
34, 3.
(31) Ruthkosky, M.; Kelly, C. A.; Zaros, M. C.; Meyer, G. J. J. Am.
Chem. Soc. 1997, 119, 12004.
(32) Ruthkosky, M.; Kelly, C. A.; Castellano, F. N.; Meyer, G. J. Coord.
Chem. ReV. 1998, 171, 309.
(33) Vo¨gtle, F.; Lu¨er, I.; Balzani, V.; Armaroli, N. Angew. Chem., Int.
Ed. Engl. 1991, 30, 1333.
(34) Dietrich-Buchecker, C. O.; Nierengarten, J. F.; Sauvage, J. P.;
Armaroli, N.; Balzani, V.; De Cola, L. J. Am. Chem. Soc. 1993, 115, 11237.
(35) Armaroli, N.; Balzani, V.; Barigelletti, F.; De Cola, L.; Flamigni,
L.; Sauvage, J. P.; Hemmert, C. J. Am. Chem. Soc. 1994, 116, 5211.
(36) Dietrich-Buchecker, C. O.; Sauvage, J. P.; Armaroli, N.; Ceroni,
P.; Balzani, V. Angew. Chem., Int. Ed. Engl. 1996, 35, 1119.
(37) Armaroli, N.; Diederich, F.; Dietrich-Buchecker, C. O.; Flamigni,
L.; Marconi, G.; Nierengarten, J. F.; Sauvage, J. P. Chem.sEur. J. 1998,
4, 406.
(38) Armaroli, N.; Boudon, C.; Felder, D.; Gisselbrecht, J. P.; Gross,
M.; Marconi, G.; Nicoud, J. F.; Nierengarten, J. F.; Vicinelli, V. Angew.
Chem., Int. Ed. 1999, 38, 3730.
(39) Caspar, J. V.; Kober, E. M.; Sullivan, B. P.; Meyer, T. J. J. Am.
Chem. Soc. 1982, 104, 630.
(40) Englman, R.; Jortner, J. J. Mol. Phys. 1970, 18, 145.
(41) Freed, K.; Jortner, J. J. Chem. Phys. 1970, 52, 6272.

Luminescent Cu(I)-Phenanthroline Complexes
J. Am. Chem. Soc., Vol. 123, No. 26, 2001 6293
Preparation of 2-(4-n-Butylphenyl)-1,10-phenanthroline. A 1.6
M n-BuLi solution (6.9 mL, 11.04 mmol) was added by syringe at
room temperature to an argon-flushed, stirred solution of 4-bromo-n-
butylbenzene (2.34 g, 11.00 mmol) in dry Et2O (40 mL). After 1.5 h,
the resulting yellow solution was added under argon to a degassed
suspension of 1,10-phenanthroline (2.0 g, 10.09 mmol) in dry Et2O
(100 mL) cooled to 0 °C (ice-water bath). The resulting dark red
mixture was stirred for 5 h and then hydrolyzed with water. The bright
yellow ether layer was decanted and the aqueous layer extracted with
CH2Cl2 (3×). The combined organic layers were thereafter rearomatized
by addition of MnO2 (30 g), dried (MgSO4), and filtered; the filtrate
was then evaporated to dryness. Column chromatography on SiO2 (CH2-
Cl2) yielded 2-(4-n-butylphenyl)-1,10-phenanthroline as a colorless
glassy product (1.89 g, 60%). 1H NMR (CDCl3, 200 MHz): δ ) 0.96
(t, J ) 6 Hz, 3H), 1.40 (m, 2H), 1.63 (m, 2H), 2.71 (t, J ) 6 Hz, 2H),
7.36 (d, J ) 8 Hz, 2H), 7.65 (dd, J ) 4 and 8 Hz, 1H), 7.80 (AB, J )
9 Hz, 2H), 8.10 (d, J ) 8 Hz, 1H), 8.29 (m, 4H), 9.26 (dd, J ) 4 and
1 Hz, 1H); C22H20N2 (312.41): calcd C 84.58, H 6.45, N 8.97; found
C 84.31, H 6.66, N 8.90.
carried out in the 300-96 K range. We have found that the
change of the luminescence properties (intensity and lifetime)
across the glass-to-fluid transition region of the solvent42,43
provides useful hints for understanding the role of the structural
and electronic factors that affect the excited levels of [Cu(NN)2]+
complexes. Importantly, for the first time, it is reported that in
some cases (i.e., complexes with long alkyl residues on the
phenanthroline ligands) the luminescence intensity pattern is
oscillating as a function of temperature: a decrease is found in
the range 300-120 K, then a dramatic increase is recorded by
further cooling. This yields strongly luminescent [Cu(NN)2]+
complexes below 100 K, quite rarely observed earlier.44
Experimental Section
General Synthetic Procedure. Reagents and solvents were pur-
chased as reagent grade and used without further purification. Com-
mercially available 1,10-phenanthroline monohydrate was dried by three
successive azeotropic distillation in toluene/ethanol (51:35). Compounds
Preparation of 2,9-Bis(4-n-Butylphenyl)-1,10-phenanthroline (5).
A 1.6 M n-BuLi solution (6.9 mL, 11.04 mmol) was added by syringe
at room temperature to an argon-flushed, stirred solution of 4-bromo-
n-butylbenzene (2.34 g, 11.00 mmol) in dry Et2O (40 mL). After 1.5
h, the resulting yellow solution was added under argon to a degassed
solution of 2-(4-n-butylphenyl)-1,10-phenanthroline (1.80 g, 5.76 mmol)
in dry Et2O (100 mL) at room temperature. The resulting dark red
mixture was stirred for 5 h and then hydrolyzed with water. The bright
yellow ether layer was decanted and the aqueous layer extracted with
CH2Cl2 (3×). The combined organic layers were thereafter rearomatized
by addition of MnO2 (30 g), dried (MgSO4), and filtered; the filtrate
was then evaporated to dryness. Column chromatography on SiO2 (CH2-
345 and (1)2Cu‚BF4 were prepared according to the literature. All
14
reactions were performed in standard glassware under an inert Ar
atmosphere. Evaporation and concentration were done at water aspirator
pressure and drying in vacuo at 10-2 Torr. Thin layer chromatography
(TLC) was performed on glass sheets coated with silica gel 60 F254
purchased from E. Merck, visualization by UV light. Melting points
were measured on an electrothermal digital melting point apparatus
and are uncorrected. IR spectra (cm-1) were measured on an ATI
Mattson Genesis Series FTIR instrument. NMR spectra were recorded
on a Bruker AC 200 (200 MHz) with solvent peaks as reference.
Elemental analysis were performed by the analytical service at the
Institut Charles Sadron (Strasbourg, France).
1
Cl2) yielded 5 as a colorless glassy product (1.95 g, 76%). H NMR
Preparation of Di-2,9-hexyl-1,10-phenanthroline (2). A 2.5 M
n-hexyllithium solution (10 mL, 25 mmol) was added by syringe to an
argon-flushed, stirred suspension of 1,10-phenanthroline (2.0 g, 10.09
mmol) in dry Et2O (100 mL) at room temperature. The resulting dark
red mixture was stirred for 48 h and then hydrolyzed with water. The
bright yellow ether layer was decanted and the aqueous layer extracted
with CH2Cl2 (3×). The combined organic layers were thereafter
rearomatized by addition of MnO2 (30 g), dried (MgSO4), and filtered;
the filtrate was then evaporated to dryness. Column chromatography
on SiO2 (CH2Cl2/hexane 5:1) yielded 2 as a colorless glassy product
(CDCl3, 200 MHz): δ ) 0.99 (t, J ) 6 Hz, 6H), 1.44 (m, 4H), 1.70
(m, 4H), 2.74 (t, J ) 6 Hz, 4H), 7.42 (d, J ) 8 Hz, 4H), 7.79 (s, 2H),
8.14 (d, J ) 8 Hz, 2H), 8.30 (d, J ) 8 Hz, 2H), 8.39 (d, J ) 8 Hz,
4H); 13C NMR (CDCl3 50 MHz): δ ) 13.96, 22.33, 33.49, 35.48,
119.67, 125.73, 127.53, 127.65, 128.87, 136.70, 136.91, 144.38, 146.02,
156.73; C32H32N2 (444.62): calcd C 86.45, H 7.25, N 6.30; found C
86.31, H 7.42, N 6.36.
General Procedure for the Preparation of the Copper(I) com-
plexes. A solution of Cu(CH3CN)4‚BF4 (0.6 equiv) in CH3CN was
added under an argon atmosphere at room temperature to a stirred,
degassed solution of the phenanthroline derivative (1 equiv) in CH2-
Cl2. The mixtures turned dark red instantaneously, indicating the
formation of the complexes. After 1 h, the solvents were evaporated.
The resulting complexes were purified by column chromatography on
Al2O3 (CH2Cl2/5% MeOH) followed by recrystallization from CH2-
Cl2/pentane [(2)2Cu‚BF4, (4)2Cu‚BF4, and(5)2Cu‚BF4] or CH2Cl2/MeOH
[(3)2Cu‚BF4]. The copper(I) complexes were thus obtained in 80-95%
yields.
(2)2Cu‚BF4: 1H NMR (CDCl3, 200 MHz): δ ) 0.56 (t, J ) 6 Hz,
12H), 0.58-0.90 (m, 24H), 1.39 (m, 8H), 2.71 (t, J ) 6 Hz, 8H), 7.80
(d, J ) 8 Hz, 4H), 8.08 (s, 4H), 8.58 (d, J ) 8 Hz, 4H); 13C NMR
(CDCl3, 50 MHz): δ ) 13.61, 22.00, 28.85, 29.60, 30.76, 40.23,
124.61, 126.15, 127.74, 137.57, 142.97, 161.51; C48H64N4CuBF4
(847.41): calcd C 68.03, H 7.61; found C 67.68, H 7.60.
(3)2Cu‚BF4: 1H NMR (acetone-d6, 200 MHz): δ ) 0.80 (m, 16H),
1.48 (m, 8H), 2.82 (t, J ) 6 Hz, 8H), 3.05 (m, 8H), 8.05 (d, J ) 8 Hz,
4H), 8.26 (s, 4H), 8.80 (d, J ) 8 HZ, 4H); 13C NMR (acetone-d6, 50
MHz): δ ) 24.97, 26.54, 32.84, 41.09, 61.81, 61.93, 126.13, 127.12,
128.93, 138.55, 138.75, 162.83; IR (CH2Cl2): 3077 cm-1 (O-H);
C44H58N4O4CuBF4 (857.32): calcd C 61.64, H 6.82; found C 61.55, H
6.67.
1
(2.47 g, 70%). H NMR (CDCl3, 200 MHz): δ ) 0.91 (t, J ) 6 Hz,
6H), 1.30-1.60 (m, 12H), 1.92 (m, 4H), 3.22 (t, J ) 6 Hz, 4H), 7.53
(d, J ) 8 Hz, 2H), 7.72 (s, 2H), 8.16 (d, J ) 8 Hz, 2H); 13C NMR
(CDCl3 50 MHz): δ ) 13.95, 22.45, 29.34, 29.57, 31.64, 39.37, 122.12,
125.23, 126.85, 135.96, 145.24, 163.05; C24H32N2 (348.53): calcd C
82.71, H 9.25, N 8.04; found C 82.70, H 9.31, N 7.90.
Preparation of Di-2,9-(5-acetyloxypentyl)-1,10-phenanthroline
(4). A solution of pyridine (0.25 mL, 3.12 mmol) and DMAP (87 mg,
0.70 mmol) in CH2Cl2 (50 mL) was added dropwise at room
temperature to an argon-flushed, stirred solution of 3 (500 mg, 1.41
mmol) and acetyl chloride (0.222 mL, 3.12 mmol). After 4 h, the
resulting solution was washed with water, dried (MgSO4), filtered, and
evaporated to dryness. Column chromatography on SiO2 (CH2Cl2/5%
MeOH) yielded 4 as a colorless glassy product (330 mg, 0.76 mmol,
53%). 1H NMR (CDCl3, 200 MHz): δ ) 1.56 (m, 8H), 1.92 (m, 4H),
2.05 (s, 6H), 3.22 (t, J ) 6 Hz, 4H), 4.10 (t, J ) 6 Hz, 4H), 7.51 (d,
J ) 8 Hz, 2H), 7.71 (s, 2H), 8.15 (d, J ) 8 Hz, 2H); 13C NMR (CDCl3,
50 MHz): δ ) 21.01, 26.06, 28.52, 29.31, 39.30, 64.48, 122.32, 125.51,
127.11, 136.26, 145.01, 162.33, 171.83; IR (CH2Cl2): 1731 cm-1 (Cd
O); C26H32O4N2‚H2O (454.57): calcd C 68.70, H 7.54; found C 68.39,
H 7.24.
(4)2Cu‚BF4: 1H NMR (CD2Cl2, 200 MHz): δ ) 0.67 (m, 8H), 0.93
(m, 8H), 1.37 (m, 8H), 1.94 (s, 12H), 2.73 (t, J ) 6 Hz, 8H), 3.57 (t,
J ) 6 Hz, 8H), 7.85 (d, J ) 8 Hz, 4H), 8.11 (s, 4H), 8.62 (d, J ) 8 Hz,
4H); 13C NMR (CD2Cl2, 50 MHz): δ ) 20.69, 25.35, 27.53, 29.17,
39.98, 63.43, 124.92, 126.21, 127.75, 137.82, 142.81, 161.10, 170.67;
IR (CH2Cl2): 1732 cm-1 (CdO); C52H64N4O8CuBF4 (1023.45): calcd
C 61.03, H 6.30; found C 61.06, H 6.57.
(42) Barigelletti, F.; Belser, P.; von Zelewsky, A.; Juris, A.; Balzani, V.
J. Phys. Chem. 1985, 89, 3680.
(43) Barigelletti, F.; Juris, A.; Balzani, V.; Belser, P.; von Zelewsky, A.
J. Phys. Chem. 1987, 91, 1095.
(44) Everly, R. M.; Ziessel, R.; Suffert, J.; McMillin, D. R. Inorg. Chem.
1991, 30, 559.
(45) Nierengarten, J. F.; Dietrich-Buchecker, C. O.; Sauvage, J. P. New
J. Chem. 1996, 20, 685.

6294 J. Am. Chem. Soc., Vol. 123, No. 26, 2001
Felder et al.
Table 1. Photophysical Properties at 298 K in CH2Cl2 Solution
τc
10-3krd 10-6knr 10-6kq[O2]f
a
e
λmax(em)
(nm) 104Φem (ns)
b
(s-1
5.6
)
(s-1
)
(s-1
4.0
)
[Cu(1)2]+
[Cu(2)2]+
[Cu(3)2]+
[Cu(4)2]+
[Cu(5)2]+
[Cu(6)2]+ g
674
(748)
660
(724)
660
3
(4)
7
(10)
6
56
(72)
98
(132)
84
13.8
7.6
9.3
6.6
4.4
7.7
7.6
7.5
7.2
5.4
2.8
2.6
2.6
2.8
2.7
1.7
(724)
658
(718)
692
(8)
8
(11)
8
(107)
107
(152)
139
(718)
686
(745)
(12)
3
(4)
(224)
107
(130)
a Emission maxima from uncorrected and, in brackets, corrected
spectra. b Emission quantum yields in air-equilibrated and, in brackets,
air-free solutions. c Excited-state lifetimes in air-equilibrated and, in
brackets, air-free solutions. d Radiative decay rate constants in air-free
solution. e Nonradiative decay rate constants in air-free solution.
f Stern-Volmer rate constants for the quenching of the luminescent
excited state by O2, calculated from lifetimes in air-free and air-
equilibrated solutions. g From ref 34.
Figure 1. Absorption spectra of [Cu(1)2]+ (dash line), [Cu(2)2]+ (full
line), [Cu(5)2]+ (dot line), and [Cu(6)2]+ (dash-dot line) in CH2Cl2
solution at 298 K. Over the 380-700 nm region a multiplying factor
of 10 is applied.
(5)2Cu‚BF4: 1H NMR (CD2Cl2, 200 MHz): δ ) 0.89 (t, J ) 6 Hz,
12H), 1.17 (m, 16H), 2.18 (t, J ) 6 Hz, 8H), 6.34 (d, J ) 8 Hz, 8H),
7.36 (d, J ) 8 Hz, 8H), 7.83 (d, J ) 8 Hz, 8.09 (s, 4H), 8.50 (d, J )
8 Hz, 4H); 13C NMR (CD2Cl2, 50 MHz): δ ) 13.85, 22.07, 33.24,
34.77, 124.35, 126.27, 126.95, 127.62, 127.93, 135.99, 137.04, 143.16,
144.06, 156.41.
since they are very similar, if not identical, to those of [Cu-
(1)2]+ and [Cu(2)2]+.
The spectral shapes of complexes of 2,9-dialkylphenanthro-
lines 1-4 are substantially the same both in the UV and in the
vis spectral region, regardless the length of the alkyl chain. The
spectrum of [Cu(5)2]+, instead, is quite peculiar. In the UV
region, where the ligand-centered (π-π*) transitions of the
phenanthroline ligands are dominant, the difference recorded
for [Cu(5)2]+ is attributable to the presence of the phenyl
fragment conjugated to the phenanthroline unit. As far as the
vis region is concerned, both electronic and structural factors
(phenyl/phenanthroline π-stacking interactions, leading to flat-
tening distortions)22,48 can explain the pattern of the MLCT
bands of the phenyphenanthroline-type complex [Cu(5)2]+, as
already discussed.1 The pronounced shoulder around 550 nm
is commonly taken as a fingerprint for 2,9-aryl substitution of
the phenanthroline ring, when the other positions remain
free.24,28
Spectroscopic Investigations. The solvents used for spectroscopic
investigations were CH2Cl2 and CH3OH (Carlo Erba, spectrofluorimetric
grade). The samples were placed in fluorimetric 1-cm path cuvettes
and, when necessary, purged from oxygen by bubbling argon or by
using freeze-thaw-pump cycles. Absorption spectra were recorded
with a Perkin-Elmer λ5 spectrophotometer. Uncorrected emission
spectra were obtained with a Spex Fluorolog II spectrofluorimeter
(continuous Xe lamp, 150 W), equipped with a Hamamatsu R-928
photomultiplier tube. The corrected spectra were obtained via a
calibration curve determined by means of a 45-W quartz-halogen
tungsten filament lamp (Optronic Laboratories) calibrated in the range
400-1800 nm. The calibration curve is flat (no correction required)
between 400 and 550 nm and then increases steeply giving a correction
factor of 7, 25, and 210 at 700, 800, and 850 nm, respectively.
Fluorescence quantum yields in neat CH2Cl2 solvent were obtained from
spectra on an energy scale (cm-1) according to the approach described
by Demas and Crosby46 and using air-equilibrated [Os(phen)3]2+ in
acetonitrile (Φem ) 0.005)47 as standard. For steady-state and time-
resolved luminescence experiments at 77 K we employed a 1:1 (v/v)
mixture of CH2Cl2 and MeOH, that was found to yield sufficiently
transparent glasses. In this case, the samples were placed in Pyrex
capillary tubes (2 mm diameter) immersed in liquid nitrogen contained
in a homemade quartz dewar. Temperature-dependent measurements
in the range 300-96 K were also performed on CH2Cl2/MeOH samples
deaerated by means of several freeze-thaw-pump cycles. Homemade
1-cm path quartz cells were employed which were placed in the
modified holder of a liquid nitrogen Thor C600 cryostat. Emission
lifetimes in the ns-µs time scale were determined with an IBH single-
photon counting spectrometer equipped with a thyratron gated nitrogen
lamp working in the range 4-40 kHz (λexc ) 337 nm, 0.5 ns time
resolution); the detector was a red-sensitive (185-850 nm) Hamamatsu
R-3237-01 photomultiplier. The uncertainties are (2 nm, (20%, and
(5%, for emission maxima, emission intensities and quantum yields,
and excited-state lifetimes, respectively.
As is typical for complexes of dialkylphenanthrolines,1 the
complexes of ligands 1-4 exhibit a low-intensity absorption
feature above 500 nm that is accompanied by a much sharper
and intense absorption band around 460 nm (ꢀmax ca. 7000 M-1
cm-1), to be compared to the weaker band of [Cu(5)2]+ (ꢀmax
) 3200 M-1 cm-1, 438 nm).49
It is important to notice that, in the visible spectral region
(Figure 1), the absorption spectrum of the complex of the
asymmetric ligand 6 exhibits intermediate intensities of the
features around 450 and 550 nm, if compared to the spectra of
the symmetric alkyl- or phenyl-phenanthroline complexes. The
isosbestic point at 488 nm (ꢀ ) 2300 M-1 cm-1) for the spectra
of Figure 1 might be accidental; however, it could also indicate
some subtle regularity in intensity band variations on passing
from alkyl to phenyl substituents.
Emission Spectra at 298 K. The luminescence data at 298
K in CH2Cl2, along with some related quantities, are collected
in Table 1. All of the complexes show a luminescence band
that can be attributed to the deactivation from MLCT excited
states1,2 (Figure 2).
The recorded spectra require severe instrumental correction
in the region of interest (550-850 nm, see Experimental
Results and Discussion
Electronic Absorption Spectra. The absorption spectra of
[Cu(1)2]+, [Cu(2)2]+, [Cu(5)2]+, and [Cu(6)2]+ are displayed in
Figure 1; the spectra of [Cu(3)2]+ and [Cu(4)2]+ are not shown
(48) Cesario, M.; Dietrich-Buchecker, C. O.; Guilhem, J.; Pascard, C.;
Sauvage, J. P. J. Chem. Soc., Chem. Commun. 1985, 244.
(49) Dietrich-Buchecker, C. O.; Marnot, P. A.; Sauvage, J. P.; Kirchhoff,
J. R.; McMillin, D. R. J. Chem. Soc., Chem. Commun. 1983, 513.
(46) Demas, J. N.; Crosby, G. A. J. Phys. Chem. 1971, 75, 991.
(47) Kober, E. M.; Caspar, J. V.; Lumpkin, R. S.; Meyer, T. J. J. Phys.
Chem. 1986, 90, 3722.

Luminescent Cu(I)-Phenanthroline Complexes
J. Am. Chem. Soc., Vol. 123, No. 26, 2001 6295
Table 2. Photophysical Properties in CH2Cl2/MeOH (1:1, v/v)
Glass at 77 K, Unless Otherwise Noted
a
b
λmax(unc.)
τ
∆E298K-77K
(cm-1
(nm)
intensity
weak
(µs)
)
[Cu(1)2]+
[Cu(2)2]+
[Cu(3)2]+
[Cu(4)2]+
[Cu(5)2]+
[Cu(6)2]+ d
706
(738)
646
(676)
644
(678)
640
(672)
736
0.82
-180
-980
-940
-950
+980
-260
strong
strong
strong
faint
2.40
2.36
3.13
c
(772)
730
(760)
weak
∼ 0.5
a Emission maxima from uncorrected and, in brackets, corrected
spectra. b Calculated from emission bands maxima of the corrected
spectra. Negative sign means blue-shift, positive sign means red-shift.
c Not measured due to signal weakness. d In CH2Cl2 matrix, from ref
34.
Figure 2. Uncorrected luminescence spectra of [Cu(1)2]+ (dash line),
[Cu(2)2]+ (full line), and [Cu(5)2]+ (dot line) in CH2Cl2 solution at 298
K upon excitation at 440 nm. The optical density of all samples is
identical (0.240); thus, the spectra are comparable in terms of relative
emission intensity. The spectra corrected for the photomultiplier
response (normalized) are reported in the inset.
intramolecular quenching of [Cu(NN)2]+ MLCT excited states
was proposed for 2,9-disubstituted phenanthrolines with ap-
pended toluidine residues.44
Emission Spectra at 77 K. Luminescence data in CH2Cl2/
MeOH glass at 77 K, are collected in Table 2. Under these
conditions all of the complexes are luminescent, but the trend
of the intensity is quite variable along the series.
Section), so that the emission maxima are remarkably shifted
to longer wavelength upon correction (Table 1, Figure 2). The
uncorrected luminescence spectra of [Cu(1)2]+, [Cu(2)2]+, and
[Cu(5)2]+ (Figure 2) look somewhat different. However, upon
correction, the spectral shapes tend to uniform, and the only
remarkable difference is the 20 nm red-shift of the [Cu(1)2]+
maximum, compared to that of the other two compounds. The
shapes of the luminescence spectra of [Cu(2)2]+, [Cu(3)2]+, and
[Cu(4)2]+ are quite similar to each other, in line with the
absorption behavior.
It has been reported several times that the luminescence
intensity of [Cu(NN)2]+ systems decreases when the temperature
is decreased. To explain this unusual behavior, a model
involving two MLCT excited states in thermal equilibrium, that
is, a singlet (1MLCT) and a related triplet (3MLCT) have been
proposed (“two-state model”) by McMillin.16,19 The energy gap
between these states has been sometimes reported to be 1000-
2000 cm-1 19,52 and, at room temperature, population of the
upper-lying 1MLCT level is achieved to some extent. By
temperature decrease, such thermal activation is depressed until
The better luminescence performance of the complexes of
2-5 relative to [Cu(1)2]+ can be rationalized on the basis of
the protection exerted toward the copper ion by the cumbersome
substituents in the 2 and 9 positions of phenanthroline. These
residues contrast the excited-state distortion of the complex and
tend to prevent the formation of pentacordinated exciplexes
(with solvent molecules or counterions), that funnel the deac-
tivation of the potentially luminescent [Cu(NN)2]+ MLCT levels
via non radiative paths.17,50 Among the investigated complexes,
both in air-equilibrated and in oxygen-free solution, [Cu(5)2]+
exhibits the highest emission quantum yield and the longest
lifetime at room temperature, the former being comparable to
those of [Cu(2)2]+ and [Cu(4)2]+. The radiative deactivation rate
constant (kr ) Φem/τ) of [Cu(5)2]+ (Table 1) is the smallest of
the series, but this is largely compensated by the value of the
nonradiative rate constant (knr ) 1/τ), that turns out to be the
smallest as well.51 As a result, [Cu(5)2]+ is the best room-
temperature luminophore of the series. A comparison between
[Cu(2)2]+and [Cu(3)2]+ is interesting since the substituents on
the phenanthroline ligands are alkyl chains of similar length,
but with different terminal groups (Chart 1). The lower emission
quantum yield and shorter lifetime displayed by [Cu(3)2]+ can
be due to some internal exciplex dynamic quenching promoted
by the terminal basic OH groups, possibly interacting with the
copper ion in the excited state. A similar, more effective,
3
only the lowest MLCT level is significantly populated. The
radiative rate constant of the 3MLCT level is much smaller than
1
that of the MLCT one; thus, in the former case nonradiative
processes are comparably more important. As a result, deactiva-
3
tion of the MLCT state gives a much weaker luminescence
than 1MLCT, thus leading to the unconventional low-emission
intensity at low temperature.19,34,35,52 Furthermore, under the
latter conditions, the emission band is usually found to be red
3
shifted (emission from MLCT) in comparison to that in fluid
room-temperature solution (emission from higher-lying
1MLCT).19,34,35,52
By inspecting the emission data at only two temperature
values (298 and 77 K, Tables 1 and 2 respectively) one could
conclude that the above two-level model is unable to explain
all of our experimental results since, for instance, [Cu(2)2]+
(Figure 3), [Cu(3)2]+, and [Cu(4)2]+ (Table 2) are very strong
emitters at 77 K.
Noticeably their bright orange luminescence is observable
also by naked eye and, though a quantitative measure in a rigid
matrix is difficult, we can roughly estimate that the emission
quantum yields of the complexes of 2-4 are at least 100 times
higher than those of [Cu(1)2]+and [Cu(5)2]+. Another unex-
pected result is that, in passing from 298 to 77 K, complexes
of alkylphenanthrolines 1-4 display emission blue-shifting (very
small in the case of 1), while for the complex of arylphenan-
throline 5 the spectrum is red-shifted (Table 2).
(50) McMillin, D. R.; Kirchhoff, J. R.; Goodwin, K. V. Coord. Chem.
ReV. 1985, 64, 83.
(51) Since the luminescence of Cu(I)-phenanthrolines is attributable to
two excited states in thermal equilibrium (a higher-lying singlet MLCT
level and a lower-lying related triplet, see text) the calculated values of the
radiative and nonradiative rate constants are not strictly referred to one
specific excited state; therefore, they can be better defined as observed rates.
However, at least at room temperature, the luminescence is mainly
attributable to the singlet excited level and, as a first approximation, the
observed rate constants can be related to it.
(52) Parker, W. L.; Crosby, G. A. J. Phys. Chem. 1989, 93, 5692.

6296 J. Am. Chem. Soc., Vol. 123, No. 26, 2001
Felder et al.
Figure 3. Uncorrected luminescence spectra of [Cu(1)2]+ (dash line),
[Cu(2)2]+ (full line), and [Cu(5)2]+ (dot line) in a CH2Cl2/MeOH 1:1
(v/v) glass at 77 K upon excitation at 440 nm. The spectra corrected
for the photomultiplier response are reported in the inset. The
experimental emission intensity of [Cu(2)2]+ is by far the strongest (see
text), and the intensity on the ordinate scale is arbitrary, to allow a
comparison. The 77 K spectrum of [Cu(5)2]+ exhibits two maxima
separated by about 2000 cm-1. This value is likely to correspond to a
vibronic progression since the difference between the two emitting
excited states in this complex (McMillin’s model) is found to be 660
cm-1 (see text).
Figure 4. Uncorrected luminescence spectra of [Cu(2)2] in a CH2Cl2/
MeOH 1:1 (v/v) at different temperatures down to 190 K. In the inset
the trend at temperatures below 190 K is shown; the very weak spectra
at 190, 170, and 118 K are multiplied by a factor of 80 to allow a
comparison with the much more intense signals at lower temperatures
(108 and 96 K). λexc is always kept at 456 nm.
From the above temperature dependent trends in luminescence
intensity and emission spectral shift (when only two temperature
values are examined, i.e., 298 and 77 K), one could conclude
that the two-level model can be fully applied in the case of
[Cu(5)2]+ and [Cu(6)2]+ (emission intensity decrease and
spectral red-shift upon cooling), can reasonably hold in the case
of [Cu(1)2]+ (emission intensity decrease and slight blue-shift),
but can hardly explain the behavior of [Cu(2)2]+, [Cu(3)2]+,
and [Cu(4)2]+ (emission intensity increase and marked blue
shift). In an attempt to rationalize these conflicting experimental
findings we decided to make detailed investigations on the
temperature dependence of the luminescence properties of
[Cu(2)2]+ (taken as a representative of the series [Cu(2)2]+,
[Cu(3)2]+, and [Cu(4)2]+) and [Cu(5)2]+.
Figure 5. Uncorrected luminescence spectra of [Cu(5)2] in a CH2Cl2/
MeOH 1:1 (v/v) at different temperatures down to 96 K; λexc ) 440
nm. In the inset the trend between 190 and 96 K is emphasized.
Temperature Dependence of the Luminescence Spectra.
For [Cu(2)2]+ and [Cu(5)2]+ in deaerated CH2Cl2/MeOH 1:1
(v/v), we have recorded the luminescence spectra (Figures 4
and 5) and the lifetimes in the temperature interval 300-96 K;
in the range 130-110 K the solvent passes from fluid to frozen
state.
Figure 6 shows the trend of the luminescence band maxima
(top panel), and the relative emission intensities (bottom panel)
as a function of temperature.
96 K (frozen solvent) causes a steady destabilization of the
luminescent level by ca. 2000 cm-1 (ca. 1000 cm-1 by
comparing data at 298 and 77 K only, Table 2). A similar
behavior (in the frozen domain) has been frequently observed
for complexes of the Ru(II)-polypyridine family in polar
solvent and has been associated with the charge-transfer nature
(3MLCT) of the emitting levels;8 on the contrary, for [Cu(NN)2]+
complexes, no extended temperature-dependent investigations
have been reported to date. For [Cu(2)2]+, it is thus observed
an “oscillating” trend in the luminescence pattern as a function
of temperature (Figures 4 and 6): an intensity decrease and
spectral red-shift is found in the range 300-120 K then a
dramatic intensity increase and blue-shift is recorded by further
cooling. This latter finding is particularly interesting because
to date Cu(I)-bisphenanthrolines have been commonly con-
sidered poor luminophores in low-temperature rigid matrixes.
Below, we discuss the nature of the excited levels involved
in the emission processes and suggest a rationale to explain the
different excited-state dynamics of [Cu(2)2]+ and [Cu(5)2]+.
Temperature Dependence of the Excited-State Lifetimes.
For both [Cu(2)2]+ and [Cu(5)2]+ these investigations were
carried out with a single photon counting spectrometer and by
using cutoff filters in order to include the full emission region.
In all cases the emission decays occurred as single exponentials.
Close inspection of the results summarized in Figures 4-6
allows the following points to be listed.
(i) For the aryl-phenanthroline complex [Cu(5)2]+, on passing
from 300 to 96 K, the emission maximum is bathochromically
shifted (Figure 6a), thus pointing to an increased stabilization
of the luminescent level. This is accompanied by a steady
decrease in luminescence intensity (Figure 6b), in line with the
behavior of many [Cu(NN)2]+ complexes reported throughout
the years.35 As mentioned above, this is interpreted in terms of
the two-state model with the lowest-lying level being poorly
luminescent.19,34,35,52
(ii) For the alkyl-phenanthroline compounds [Cu(2)2]+, a more
complex behavior is found on lowering the temperature. Here,
stabilization of the emitting level is maximized in the interval
120-150 K (Figure 6a) but further lowering of temperature till

Luminescent Cu(I)-Phenanthroline Complexes
J. Am. Chem. Soc., Vol. 123, No. 26, 2001 6297
Table 3. Kinetic Parameters for Excited State Decaya
A
∆E
B
k0
(s-1
)
(cm-1
)
(s-1
)
(s-1
)
[Cu(2)2]+
[Cu(5)2]+
2.1 × 108
800
660
1.0 × 107
5.9 × 105
6.8 × 107
2.1 × 106
1.3 × 106
a In CH2Cl2/MeOH 1:1 (v/v); from fitting of eqs 1 of the text to the
experimental points of Figure 7; Tg was evaluated 117 ( 5 K.
B represents the decrease of nonradiative rate constant observed
on passing from fluid to frozen solvent. In the examined cases,
Tg occurred in the interval 110-130 K, with the analysis
providing Tg ) 117 ( 5 K. Table 3 collects parameters obtained
from the least-squares nonlinear fitting analysis of the data points
displayed in Figure 7.
According to the results of this analysis, some issues can be
highlighted.
(i) For both [Cu(2)2]+ and [Cu(5)2]+, the luminescence
lifetime increases monotonically on passing from room tem-
perature to 96 K (frozen solvent), Figure 7. According to the
two-state equilibrium model for the lowest MLCT excited states
of [Cu(NN)2]+, use of eq 1 provided ∆E ) 800 and 660 cm-1
for, respectively, [Cu(2)2]+ and [Cu(5)2]+ (Table 3); this
corresponds to the energy gap separating the two luminescent
levels.
Figure 6. Temperature dependence in the 300-96 K range of (a)
luminescence band maxima (cm-1, from uncorrected spectra) and (b)
relative emission intensities of [Cu(2)2] (λexc ) 456 nm, open points)
and [Cu(5)2] (λexc ) 440 nm, solid points) in CH2Cl2/MeOH 1:1 (v/v).
The emission intensity data of [Cu(5)2] (solid triangles) are multiplied
by a factor of 10 for the sake of clarity.
As for the nature of the two states (see above) it has been
suggested that the higher-lying one is a singlet (1MLCT) and
the lower-lying a triplet (3MLCT);19 this can explain the increase
of the excited state lifetime accompanied by the bathochromic
emission shift, upon temperature decrease. In our case only the
trend in temperature dependence of the lifetimes is “conven-
tional”; instead, for [Cu(2)2]+, the frozen state of the solvent
causes a highly efficient population of the higher-lying emitting
level, since the (very intense) emission is hypsochromically
shifted. Interestingly, at 96 K the measured lifetime of [Cu(2)2]+
is 1510 ns, that is, twice the value of [Cu(5)2]+ (780 ns).
(ii) For both [Cu(2)2]+ and [Cu(5)2]+, the lifetime undergoes
a stepwise change in the glass-to-fluid transition region, 100-
130 K, see plot of Figure 7. This behavior is typical of CT
levels of molecules embedded in polar solvents and is related
to the fact that in frozen media these levels are destabilized
because of the lack of solvent relaxation.42,43 In our case the
solvent dielectric constants are 8.9 and 32.7 for CH2Cl2 and
MeOH respectively, so the 1:1 mixture can be regarded as a
solvent of intermediate polarity. The influence of the solvent is
clearly larger for [Cu(2)2]+ than for [Cu(5)2]+ since the B
parameter, that describes the jump observed in the plots of
Figure 7, is 1.0 ×107 and 2.1 ×106 s-1, respectively (Table 3).
This suggests that for [Cu(2)2]+ the luminescence might be
associated with excited states bearing more CT character than
for the latter case. In other words, according to a simplified
picture, the (metal) promoted electron spreads over the LUMO
of the ligand, so that the average distance of the electron from
the metal center is larger in 2 than in 5.20 Therefore, the dipole
vector associated with the CT state is expected to be larger in
the case of the [(2)CuII(2-)]+ excited complex than for
[(5)CuII(5-)]+. Another explanation for the different glass-to-
fluid transition in the two compounds has to do with the blocking
of structural distortions in frozen solvent, as is pointed out
below.
Figure 7. Temperature dependence of the luminescence lifetimes in
the 300-96 K range for [Cu(2)2] (open squares) and [Cu(5)2] (solid
squares) in CH2Cl2/MeOH 1:1 (v/v); λexc ) 337 nm.
According to the two-state model, we have analyzed the data
with an Arrhenius-type approach;42,43 therefore, the reciprocal
of the lifetime (1/τ) was plotted against the reciprocal of
temperature (1000/T), Figure 7.
As one can see from the Figure, the 1/τ versus 1/T trend is
qualitatively similar for both complexes and the experimental
data points can be fitted with the following equations in both
cases.
∆E
RT
1/τ ) A exp -
+ k′
(1a)
(1b)
0
(
)
B
k′ ) k0 +
0
1 + exp[C(1/T - 1/Tg)]
(1a) is an Arrhenius-type equation, where a term taking care of
the glass-to-fluid transition region is added (k′, eq 1b).42,43 In
0
eq 1a, A and ∆E are, respectively, the preexponential factor
and the energy barrier for thermal redistribution of the excited
species between the levels of the two-state model. k′ (eq 1b)
0
includes a low-temperature limiting rate constant, k0 (that
comprises the radiative and nonradiative contributions at 96 K),
whereas B accounts for effects occurring around the glass-to-
fluid transition temperature (Tg). Within this empirical descrip-
tion and given that the luminescence lifetime of the examined
complexes is governed by nonradiative processes (i.e. τ ≈ knr-1),
Influence of the Ground- and Excited-States Geometries
on the Photophysical Properties. Absorption spectra. The
ideal ground-state geometry for [Cu(NN)2]+ complexes corre-
sponds to a quasi-tetrahedral D2d symmetry.1-3 Upon light
absorption, the population of the lowest-lying MLCT excited

6298 J. Am. Chem. Soc., Vol. 123, No. 26, 2001
Felder et al.
Figure 8. Relative orientation of the two ligand planes (A and B) and
of the reference θz, θx, and θy angles. yz is the plane of the sheet; the
black circles represent the N phenanthroline atoms; the Cu ion coincides
with the origin of the reference axes; z is the direction bisecting both
N-Cu-N triangles; the vector ê lies on the B plane, the vector η is
perpendicular to it (see ref 15).
Figure 9. Orbital interaction schemes under local C2ν symmetry
between the Cu ion (at the center of the reference system) and ligand
A.
transitions is provided, according to a very simplified description
of the orbital interactions under a local C2ν symmetry.16
For phen, the LUMO/LUMO+1 energy gap is evaluated 0.8
eV by using standard extended Huckel molecular orbital
(EHMO) calculations, to be compared with an experimental
energy difference of the two vis absorption features in [Cu(5)2]+
and [Cu(6)2]+ of ca. 0.6 eV.
states leads to geometrical distortions, whose extent depends
on specific structural factors related to the substitution patterns
of the phenanthroline ligand. Figure 8 provides an illustration
of the parameters that describe the distortions from an ideal
tetrahedral geometry.15
Emission Spectra and Lifetimes. The correlation between
geometric and luminescence properties in [Cu(NN)2]+ com-
plexes is more difficult to discuss, since several factors should
be taken into account.16 They include (i) the effect of specific
distortions due to the d9 excited-state electronic configuration
at the metal center, (ii) the fact that two very close emitting
MLCT levels can be involved, and (iii) the possible role of
spin-orbit coupling in determining the symmetry (i.e., the
forbidden character) of the lowest-lying cluster of MLCT states.
At any rate, in practice, the luminescent MLCT excited states
of [Cu(NN)2]+ complexes are usually reported to be stabilized
by decreasing temperatures; this may result in very low (if any)
luminescence intensity, as it happens for our [Cu(5)2]+ complex.
Two effects could account for this uncommon spectroscopic
behavior: (i) the typical decrease of emission yield when the
energy of an emitting level is decreased (energy-gap law),39-41
(ii) the intermolecolar quenching of [Cu(NN)2]+ excited states
via formation of exciplexes with donating solvents in fluid
media.17 The stepwise 1/τ vs 1/T behavior around 117 K, Figure
7, indicates that on passing from fluid to rigid medium the
nonradiative deactivation steps are depressed to a different extent
Of great importance are the θz and θx angles that define the
“flattening” and “rocking” distortions of the phenanthroline
ligands with respect to each other. In practice, combinations of
various types of distortions are likely to occur.
The absorption spectra of [Cu(2)2]+, [Cu(5)2]+ and [Cu(6)2]+
(Figure 1) provide a nice illustration of the control exerted by
the molecular geometry over the optical properties. We notice
first that for [Cu(2)2]+, whose geometry is assumed to be close
to D2d, only one intense vis absorption band is present (λmax
)
458 nm, ꢀ ) 7300 M-1 cm-1). According to McMillin, such
absorption band can be associated to a MLCT z-polarized
transition along the C2 axis bisecting the Cu-NN angle (Figure
8).16 This involves one orbital of the ed(xz,yz) metal set and one
(π* LUMO) orbital from the ligand system.
The presence of one or two phenyl residues on the chelating
phenanthroline, causes a marked change in the spectral shape
(Figure 1). This is a consequence of interligand (phenyl/
phenanthroline) π-stacking interactions that favor distortions
from the tetrahedral geometry.22,53 Interestingly, the integrated
areas Y of the absorption profiles of [Cu(2)2]+, [Cu(5)2]+, and
[Cu(6)2]+ in the 25000-14300 cm-1 range (400-700 nm,
Figure 1) are 3200, 3000, and 2500 M-1 cm-2 respectively.
The fact that such values are reasonably comparable may suggest
that the strong transition associated with the absorption band
of [Cu(2)2]+ actually undergoes a splitting into two main
contributions (two bands around 450 and 550 nm, respectively)
in the case of [Cu(5)2]+ and of [Cu(6)2]+, as a consequence of
symmetry lowering. At odds with what was hypothesized by
Shinozaki and Kaizu,54 small deviations of the θz and θx angles
(5-15°) are sufficient for yielding a neat dual band behavior,
as recently reported by Karpishin.28 Thus, if deviations from
the D2d arrangement are not so large to induce scrambling of
the metal orbital energy layout, it might well be that the two
observed low-energy bands of Figure 1, with maxima around
450 and 550 nm, correspond to transitions from metal orbitals
to the LUMO+1 and LUMO levels of the ligands. In Figure 9,
an illustration for the dyzfψLUMO+1 and dxyføLUMO z-allowed
for [Cu(2)2]+ and [Cu(5)2]+, B ) 1.0 × 107 and 2.1 × 106 s-1
,
respectively. In the case of the diphenyl-phenanthroline ligand
([Cu(5)2]+), comparison of the plots of Figures 4 and 7 suggests
a minor role for “energy gap law” effect, since the change of
state of the solvent negligibly affects the emission energy
position. On this basis, the observed change in nonradiative rate
for [Cu(5)2]+ (B ) 2.1 × 106 s-1) is likely due to inhibition of
exciplex quenching owing to a reduced mobility of the partners
in the frozen solvent.
On the contrary, for [Cu(2)2]+, the luminescent level is
destabilized by about 2000 cm-1 on freezing the solvent, and a
very intense luminescence is observed. We interpret this
behavior as a consequence of the blocking of the complex
geometry to the ground state arrangement in the rigid matrix.
In this case, substantial “energy-gap law” effects are expected.
We found B ) 1.0 × 107 s-1 for [Cu(2)2]+, and a comparison
with the B value obtained for [Cu(5)2]+ (see above) may suggest
that in the former only 20% of B is related to quenching by the
solvent, the remaining effect being ascribable to changes in
excited-state energy level (energy gap law). The luminescence
(53) Meyer, M.; Albrecht-Gary, A. M.; Dietrich-Buchecker, C. O.;
Sauvage, J. P. Inorg. Chem. 1999, 38, 2279.
(54) Shinozaki, K.; Kaizu, Y. Bull. Chem. Soc. Jpn. 1994, 67, 2435.

Luminescent Cu(I)-Phenanthroline Complexes
J. Am. Chem. Soc., Vol. 123, No. 26, 2001 6299
behavior of [Cu(3)2]+ and [Cu(4)2]+, that is identical to that of
[Cu(2)2]+, can be rationalized with the same arguments.
The prominence of structural over electronic factors, to
explain the observed “anomalous” luminescence behavior at low
temperatures for [Cu(2)2]+, [Cu(3)2]+, and [Cu(4)2]+, is con-
firmed by the luminescence properties of [Cu(1)2]+. Such
complex, though of alkylphenanthroline-type, has been known
to be a poor emitter at low temperatures since a long time.19
The small methyl residues are likely not able to prevent some
minimal excited-state distortion even in rigid matrix, just enough
to yield weak luminescence.
π-stacking interactions, see absorption spectrum) and [Cu(1)2]+
(where distortions can still occur in the excited state even below
120 K, see above). On the contrary, long alkyl chains on the
phenanthroline ligands prevent both type of distortions. As a
result, the lower-lying poorly emitting MLCT level is not
apparently involved, and the observed strong luminescence
arises from the (sole) upper lying MLCT state.
We believe that this work provides a relevant breakthrough
toward the optimization of the luminescence performances of
Cu(I)-phenanthrolines,1-3 since it is suggested that both
bulkiness44 and alkyl-nature of the ligands are necessary features
to observe luminescence in frozen media. Our observations,
coupled with those recently reported by the groups of McMillin24
and Karpishin29 can give clear guidelines for the synthesis of
highly luminescent and long-lived complexes both in fluid
solution and in rigid matrix. In the latter case the Cu(I) complex
is not necessarily embedded in low-temperature glasses but can
be a room-temperature solid sample;55 indeed, it has been
recently demonstrated that even pristine [Cu(phen)2]+ is weakly
luminescent in the solid state.25 It is thus conceivable that, for
instance, complexes bearing long alkyl chains in the 2,9 and
methyl residues in the 3,8 positions of the phenanthroline ring
can be ideal [Cu(NN)2]+ luminophores, displaying strong
luminescence both in fluid and in rigid media. Further, the
availability of simple Cu(I)-phenanthrolines displaying long-
lived and strongly luminescent excited states in rigid (solid)
matrix would open new perspectives in terms of practical
applications (e.g. photovoltaic technology and chemical sensing),
thus increasing the competition toward [Ru(bpy)3]2+-type
complexes.1
Conclusions
A relevant issue for to the design of Cu(I)-phenanthroline
complexes exhibiting improved luminescence properties (i.e.
high emission intensity and long excited-state lifetimes) is
concerned with a full understanding of the structural and
electronic properties of the luminescent MLCT excited states.
The long-term holding of the so-called two-level model has
provided guidelines in this respect.16 According to this model,
the low-temperature emission is foreseen very weak. However,
we show here that at low temperature, when the solvent is rigid,
some complexes may exhibit enhanced luminescence. Our
findings are not in contrast with the two-state model but, on
the contrary, bring further evidence to support its validity. The
picture that stems from this investigation shows that two emitting
MLCT excited states are always present as long as excited-
state distortions can occur. At room temperature, despite strong
differences in the absorption spectra, the luminescence properties
are similar for all complexes, the variability of emission
performances being predominantly ascribable to “external”
factors such as bimolecular exciplex quenching.17 When the
temperature is decreased, yet remaining in a fluid domain, the
emission band is decreased in intensity and shifted to lower
energy for all complexes, in agreement with the two-state model.
Then, when the solvent is transformed into a rigid matrix (below
120 K in our cases) the trends differentiate. The “regular”
behavior is still followed in the case of [Cu(5)2]+ (where
distortions are intrinsically present due to intramolecular
Acknowledgment. This work was supported by the Italian
CNR and French CNRS. We thank Mr. M. Minghetti for
technical assistance. Financial support from the European Union
(TMR Network ERBFMRX CT 980226) is gratefully
acknowledged.
JA0043439
(55) Eggleston, M. K.; Fanwick, P. E.; Pallenberg, A. J.; McMillin, D.
R. Inorg. Chem. 1997, 36, 4007.