A white-light-emitting molecule: Frustrated energy transfer between constituent emitting centers

Article abstract of DOI:10.1021/ja902533f

White-light-emitting single molecules are promising materials for use in a new generation of displays and light sources because they offer the possibility of simple fabrication with perfect color reproducibility and stability. To realize white-light emiss

Full text of DOI:10.1021/ja902533f

Published on Web 05/29/2009
A White-Light-Emitting Molecule: Frustrated Energy Transfer
between Constituent Emitting Centers
Sanghyuk Park,† Ji Eon Kwon,† Se Hun Kim,† Jangwon Seo,† Kyeongwoon Chung,†
Sun-Young Park,‡ Du-Jeon Jang,‡ Begon˜a Milia´n Medina,§ Johannes Gierschner,§,|
and Soo Young Park*,†
Department of Materials Science and Engineering, Seoul National UniVersity,
ENG 445, Seoul 151-744, Korea, Department of Chemistry, Seoul National UniVersity,
NS60, Seoul 151-742, Korea, Institut de Cie`ncia Molecular (ICMOL), UniVersity of Valencia,
P.O. Box 22085, 46071 Valencia, Spain, and Madrid Institute for AdVanced Studies, IMDEA
Nanoscience, UAM, Modulo C-IX, AV. Toma´s y Valiente 7, Campus de Cantoblanco,
28049 Madrid, Spain
Received March 31, 2009; E-mail: parksy@snu.ac.kr
Abstract: White-light-emitting single molecules are promising materials for use in a new generation of
displays and light sources because they offer the possibility of simple fabrication with perfect color
reproducibility and stability. To realize white-light emission at the molecular scale, thereby eliminating the
detrimental concentration- or environment-dependent energy transfer problem in conventional fluorescent
or phosphorescent systems, energy transfer between a larger band-gap donor and a smaller band-gap
acceptor must be fundamentally blocked. Here, we present the first example of a concentration-independent
ultimate white-light-emitting molecule based on excited-state intramolecular proton transfer materials. Our
molecule is composed of covalently linked blue- and orange-light-emitting moieties between which energy
transfer is entirely frustrated, leading to the production of reproducible, stable white photo- and
electroluminescence.
Introduction
white light covering the visible range from 400 to 700 nm. For
instance, Coppo and colleagues reported white-light emission
Synthesizing white-light-emitting single molecules is currently
an exciting challenge in synthetic chemistry because of the
possibility of innovative applications to new classes of displays1-3
and light sources.4,5 Compared to the multicomponent molecular
emitters of white-light-emitting organic light-emitting devices
(WOLEDs), use of a single molecule would enable easy
fabrication with perfect color reproducibility and stability.6-8
All of the white-light-emitting molecules reported to date,
without exception, are composed of covalently linked emitter
components and utilize partial energy transfer from a wide band-
gap donor component to a smaller band-gap acceptor to generate
from a single molecule in solution; the molecule was composed
of luminescent iridium and europium complexes, and partial
energy transfer was involved.9 Yang and co-workers synthesized
a pH-dependent organic white-light-emitting fluorophore com-
posed of benzo[a]- and [b]xanthene derivatives.10 Because
energy transfer characteristics in conventional fluorescent or
phosphorescent systems depend strongly on donor-acceptor
distance and mutual orientation, and thus on the concentrations
and/or morphologies of these materials in the final device, the
reported systems showed white-light emission only under very
specific conditions. Therefore, the synthesis of efficient and
color-stable white-light-emitting single molecules, not affected
by the environment, is exceedingly difficult using the strategies
reported to date.
To achieve white-light emission on a molecular scale, without
concentration-dependent energy transfer problems, energy trans-
fer between donor and acceptor must be fundamentally blocked.
To date, various concepts, based on an unpopulated ground state
of the energy acceptor, have been proposed; these include use
of an excimer,8 the exciplex approach,11 interligand energy
† Department of Materials Science and Engineering, Seoul National
University.
‡ Department of Chemistry, Seoul National University.
§ University of Valencia.
| Madrid Institute for Advanced Studies.
(1) Andersson, P.; Nilsson, D.; Svensson, P. O.; Chen, M. X.; Malmstrom,
A.; Remonen, T.; Kugler, T.; Berggren, M. AdV. Mater. 2002, 14,
1460.
(2) Kido, J.; Kimura, M.; Nagai, K. Science 1995, 267, 1332.
(3) Chen, Y.; Au, J.; Kazlas, P.; Ritenour, A.; Gates, H.; McCreary, M.
Nature 2003, 423, 136.
(4) Hayes, R. A.; Feenstra, B. J. Nature 2003, 425, 383.
(5) Sun, Y. R.; Giebink, N. C.; Kanno, H.; Ma, B. W.; Thompson, M. E.;
Forrest, S. R. Nature 2006, 440, 908.
(9) Coppo, P.; Duati, M.; Kozhevnikov, V. N.; Hofstraat, J. W.; De Cola,
L. Angew. Chem., Int. Ed. 2005, 44, 1806.
(6) Segal, M.; Singh, M.; Rivoire, K.; Difley, S.; Van Voorhis, T.; Baldo,
M. A. Nat. Mater. 2007, 6, 374.
(10) Yang, Y. J.; Lowry, M.; Schowalter, C. M.; Fakayode, S. O.; Escobedo,
J. O.; Xu, X. Y.; Zhang, H. T.; Jensen, T. J.; Fronczek, F. R.; Warner,
I. M.; Strongin, R. M. J. Am. Chem. Soc. 2006, 128, 14081.
(11) Mazzeo, M.; Pisignano, D.; Della Sala, F.; Thompson, J.; Blyth,
R. I. R.; Gigli, G.; Cingolani, R.; Sotgiu, G.; Barbarella, G. Appl.
Phys. Lett. 2003, 82, 334.
(7) Mazzeo, M.; Vitale, V.; Della Sala, F.; Anni, M.; Barbarella, G.;
Favaretto, L.; Sotgiu, G.; Cingolani, R.; Gigli, G. AdV. Mater. 2005,
17, 34.
(8) Liu, Y.; Nishiura, M.; Wang, Y.; Hou, Z. M. J. Am. Chem. Soc. 2006,
128, 5592.
9
10.1021/ja902533f CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 14043–14049 14043

A R T I C L E S
Park et al.
transfer (ILET),12 and excited-state intramolecular proton
transfer (ESIPT) systems.13 However, these “frustrated
energy transfer” strategies for constructing a white-emitting
molecular dyad have yet to be demonstrated, although few cases
of white-emitting molecule based on “partial energy transfer”
have been reported.14 As a promising step toward our goal of
designing white-light-emitting molecules devoid of the energy
transfer problem, we exploited a pair of hydroxy-substituted
tetraphenyl imidazole derivatives as complementary and inde-
pendent emissive components based on frustration of energy
transfer between these materials. This arises because of ESIPT,
as previously demonstrated with a molecular mixture,13 leading
to an energy state mismatch arising from the lack of a ground-
state population in the acceptor.15
of fluorescence kinetic profiles with excitation at the front face by
347 nm laser pulses. Emission wavelengths were selected by
combining band-pass and cutoff filters. Fluorescence decay times
were extracted by biexponential fitting procedures through decon-
volution employing instrumental response functions. Differential
scanning calorimetry (DSC) was carried out under a nitrogen
atmosphere at a heating rate of 20 °C/min on a Perkin-Elmer DSC7.
Cyclic voltammetric experiments were performed using a model
273A machine (Princeton Applied Research) with a one-compart-
ment electrolysis cell consisting of a platinum working electrode,
a platinum wire counter electrode, and a quasi Ag+/Ag electrode
as reference. Measurements were performed in 0.5 mM CH2Cl2
solution, with tetrabutylammonium tetrafluoroborate as supporting
electrolyte, at a scan rate of 50 mV/s. Each oxidation potential was
calibrated using ferrocene as a reference.
Calculations. All (time-dependent) density functional theory
(TD-DFT) calculations were carried out in the gas phase using the
TURBOMOLE 5.10 quantum-chemical package, employing the
B3LYP function and the 6-311G** basis set.32 TD-DFT was shown
to reproduce the E* f K* ESIPT process until the biradical
character began to dominate, which, in the gas phase, led to fast
internal conversion via a conical intersection governed by torsional
relaxation within the hydroxyphenyl imidazole unit.33 Thus, TD-
DFT sufficiently described the radiative processes of HPI and HPNI.
To this end, vertical emission was calculated using a coplanar
geometry of the K* form of the hydroxyphenyl imidazole backbone.
EL Device Fabrication. EL devices were fabricated with a
configuration of (A) ITO/NPD (40 nm)/EML (W1, 30 nm)/BPhen
(50 nm)/LiF (1 nm)/Al (100 nm) and (B) ITO/2-TNATA (60 nm)/
NPD (20 nm)/EML (W1, 40 nm)/BPhen (20 nm)/LiF (1 nm)/Al
(100 nm). All organic layers were deposited onto UV-O3-treated
indium-tin-oxide (ITO)-coated glass. All organic layers were
deposited by thermal evaporation under a base pressure of <10-7
Torr. Deposition of a hole-transport layer of 600 Å (device B only;
see Figure S10 in Supporting Information for details), a thick layer
of 4,4′,4′′-tris(2-naphthylphenylamino)triphenylamine (2-TNATA),
and a 400 or 200 Å thick layer of N,N′-diphenyl-N,N′-bis(1-
naphthalenyl)-(1,1′-biphenyl)-4,4′-diamine (NPD) was followed by
deposition of the emitting layer (W1). Over this layer, hole-blocking
and electron-transport layers (500 or 200 Å of 4,7-diphenyl-1,10-
phenanthroline [BPhen]) were sequentially deposited. Finally, 10
Å thick LiF and 1000 Å thick aluminum were successively
deposited to form a cathode. The active area was 2 mm × 2 mm.
Current-voltage-luminescence characteristics were obtained using
a Keithley 237 source measurement unit and a Photoresearch PR-
650 spectrometer.
Here, we report the first example of a concentration-
independent ultimate white-light-emitting molecular dyad based
on two ESIPT keto-emitting units. These molecules are co-
valently linked blue- and orange-light-emitting moieties between
which energy transfer is frustrated, and the combination thus
shows broad white photoluminescence covering the entire visible
range after single-wavelength excitation, as well as stable white
electroluminescence, upon simple OLED device fabrication.
Experimental Section
Measurements. Chemical structures were fully identified by 1H
NMR (JEOL, JNM-LA300), 13C NMR (Bruker, Avance DPX-300),
MALDI-TOF mass spectrometry (Applied Biosystems Inc.; Voyager-
DE STR Biospectrometry Workstation), GC-mass spectrometry
(JEOL, JMS-AX505WA), and elemental analysis (CE Instruments,
EA1110). Absorption spectra were obtained using either of two
UV-vis spectrophotometers (Scinco S-2040 and Shimadzu UV-
1650PC). Photoluminescence emission and excitation spectra were
obtained using a fluorescence spectrophotometer (Varian, Cary
Eclipse); excitation spectra were corrected for characteristics of the
light source. Emission spectra were corrected for wavelength-
dependent sensitivity of the detector. Photoluminescence quantum
efficiencies (ΦPL values) for solutions were obtained using 9,10-
diphenylanthracene as reference.16 The ΦPL values of thin films
on quartz plates were measured using a 6 in. integrating sphere
(Labsphere, 3P-GPS-060-SF) equipped with a 325 nm CW He-Cd
laser (Omnichrome, Series 56) and a PMT detector (Hamamatsu,
PD471) attached to a monochromator (Acton Research, Spectrapro-
300i). The detailed procedure followed to obtain solid-state ΦPL
data has been described elsewhere.17 Pulse-excited emission spectra
were measured using an actively/passively mode-locked 25 ps Nd:
YAG laser (Quantel, YG701) and an intensified CCD (Princeton
Instruments, ICCD576G) attached to a spectrometer (Acton Re-
search, Spectrapro-500) as the excitation source and detector,
respectively. Samples were excited at the front face with pulses of
347 nm, generated through a Raman shifter filled with methane at
15 atm, and pumped by the fourth-harmonic pulses (266 nm) of
the laser. A 10 ps streak camera (Hamamatsu, C2830) with a CCD
detector (Princeton Instruments, RTE128H) was used for detection
Results and Discussion
The standard method for white-light generation typically
involves mixing of the three primary colors red, green, and
blue2,18-20 or combining two complementary colors.8,21 How-
ever, the simple mixing of luminescent materials can be
problematic because color balancing is difficult to control as a
result of near-inevitable energy transfer, leading to dominant
emission from lower band-gap components, even when they
are present in very small amounts.22 This energy transfer
(12) You, Y.; Kim, K. S.; Ahn, T. K.; Kim, D.; Park, S. Y. J. Phys. Chem.
C 2007, 111, 4052.
(18) Gong, X.; Wang, S.; Moses, D.; Bazan, G. C.; Heeger, A. J. AdV.
Mater. 2005, 17, 2053.
(13) Kim, S.; Seo, J.; Jung, H. K.; Kim, J. J.; Park, S. Y. AdV. Mater.
2005, 17, 2077.
(19) (a) Aharon, E.; Kalina, M.; Frey, G. L. J. Am. Chem. Soc. 2006, 128,
15968. (b) Wu, C. C.; Sturm, J. C.; Register, R. A.; Thompson, M. E.
Appl. Phys. Lett. 1996, 69, 3117.
(14) (a) Klymchenko, A. S.; Pivovarenko, V. G.; Ozturk, T.; Demchenko,
A. P. New J. Chem. 2003, 27, 1336. (b) Klymchenko, A. S.;
Yushchenko, D. A.; Mely, Y. J. Photochem. Photobiol. A 2007, 192,
93.
(20) Kido, J.; Hongawa, K.; Okuyama, K.; Nagai, K. Appl. Phys. Lett. 1994,
64, 815.
(15) (a) Goodman, J.; Brus, L. E. J. Am. Chem. Soc. 1978, 100, 7472. (b)
Kasha, M. J. Chem. Soc., Faraday Trans. 2 1986, 82, 2379.
(16) deMello, J. C.; Wittmann, H. F.; Friend, R. H. AdV. Mater. 1997, 9,
230.
(21) (a) Liu, J.; Guo, X.; Bu, L. J.; Xie, Z. Y.; Cheng, Y. X.; Geng, Y. H.;
Wang, L. X.; Jing, X. B.; Wang, F. S. AdV. Funct. Mater. 2007, 17,
1917. (b) Qin, D. S.; Tao, Y. Appl. Phys. Lett. 2005, 86, 113507.
(22) (a) Van Der Meer, B. W.; Coker, G.; Chen, S. Y. S. Resonance Energy
Transfer: Theory and Data; Wiley: New York, 1994. (b) Berggren,
M.; Dodabalapur, A.; Slusher, R. E.; Bao, Z. Nature 1997, 389, 466.
(17) Berlman, I. B. Handbook of Fluorescence Spectra of Aromatic
Molecules; Academic Press: New York, 1971.
9
14044 J. AM. CHEM. SOC. VOL. 131, NO. 39, 2009

A White-Light-Emitting Molecule
A R T I C L E S
Figure 1. Schematic representation of the ESIPT photocycle and the structures of blue-, orange-, and white-light-emitting ESIPT molecules. (a) Characteristic
four-level photocycle scheme of the ESIPT process. (b) Molecular structure of HPI, HPNI, and HPNIC. (c) Structures of white-light-emitting molecules
based on ESIPT dyads. W1 and W2 were used to generate white light. The reference molecule HPI-DCM, which contains a conventional fluorophore, was
also synthesized for comparative purposes.
originates from the interaction between an excited state donor
with an acceptor in its electronic ground state. The extent of
energy transfer depends on several parameters, including the
radiative decay rate of the donor and the spectral overlap of
donor emission and acceptor absorption, but especially on
donor-acceptor separation and orientation, and thus critically
on component concentration and the morphology of the final
device. Therefore, complete frustration of energy transfer
between donor and acceptor components is the only way to
obtain stable color mixing in a single molecule; this has not
been achieved to date.
To realize this aim, we designed a single molecular dyad
composed of two ESIPT moieties in place of ordinary fluores-
cent materials. Each ESIPT moiety is characterized by a fast
enol (E) to keto (K) cyclic intramolecular phototautomerization
and demonstrates a high-energy UV enol absorption coupled
with tunable keto emission.13,23-26 As summarized in Figure
1a, ESIPT-exhibiting molecules in the ground state exist
exclusively as enol (E) forms, but upon photoexcitation, they
undergo tautomerization into keto forms (E* f K*) via an
extremely fast and irreversible ESIPT process occurring in the
subpicosecond time domain.27 After the excited keto form (K*)
decays radiatively to the intermediate ground state (K), the initial
E form is instantaneously recovered. Therefore, predominant
absorption from E and emission from K* results in an
anomalously large Stokes shift with no self-absorption, providing
an ideal system for proton-transfer lasers27,28 and UV photo-
stabilizers.29 Moreover, it has been shown that proper molecular
design and synthesis allowed the ESIPT keto emission wave-
lengths to be tuned over the whole range of the visible spectrum
but with retention of enol form absorption in the UV region.30
Therefore, the synthesis of a dyad composed of appropriate
ESIPT fluorophores with complementary emission colors was
expected to provide a novel and uncomplicated strategy for
synthesis of a white-light-emitting single molecule by complete
suppression of donor-acceptor energy transfer.
In the first step, because blue and orange are almost exactly
complementary,21 we considered a combination of the blue-
emitting ESIPT fluorophore 2-(1,4,5-triphenyl-1H-imidazol-2-
yl)phenol (HPI, λmax ) 470 nm, Figure 1b) and the orange-
emitting fluorophore 2,5-bis(5-(4-tert-butylphenyl)-1,3,4-oxadiazol-
2-yl)benzene-1,4-diol (DOX, used previously for demonstration
(27) Khan, A. U.; Kasha, M. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 1767.
(28) Chou, P.; Mcmorrow, D.; Aartsma, T. J.; Kasha, M. J. Phys. Chem.
1984, 88, 4596.
(23) Tarkka, R. M.; Zhang, X. J.; Jenekhe, S. A. J. Am. Chem. Soc. 1996,
118, 9438.
(24) Fischer, M.; Wan, P. J. Am. Chem. Soc. 1999, 121, 4555.
(25) (a) Park, S.; Kim, S.; Seo, J.; Park, S. Y. Macromolecules 2005, 38,
4557. (b) Park, S.; Seo, J.; Kim, S. H.; Park, S. Y. AdV. Funct. Mater.
2008, 18, 726. (c) Park, S.; Kim, S.; Seo, J.; Park, S. Y. Macromol.
Res. 2008, 16, 385. (d) Park, S.; Kwon, O. H.; Lee, Y. S.; Jang, D. J.;
Park, S. Y. J. Phys. Chem. A 2007, 111, 9649.
(29) Goeller, G.; Rieker, J.; Maier, A.; Stezowski, J. J.; Daltrozzo, E.;
Neureiter, M.; Port, H.; Wiechmann, M.; Kramer, H. E. A. J. Phys.
Chem. 1988, 92, 1452.
(30) (a) Chen, K. Y.; Hsieh, C. C.; Cheng, Y. M.; Lai, C. H.; Chou, P. T.
Chem. Commun. 2006, 4395. (b) Chou, P. T.; Huang, C. H.; Pu, S. C.;
Cheng, Y. M.; Liu, Y. H.; Wang, Y.; Chen, C. T. J. Phys. Chem. A
2004, 108, 6452. (c) Seo, J.; Kim, S.; Park, S.; Park, S. Y. Bull. Korean
Chem. Soc. 2005, 26, 1706.
(26) Klymchenko, A. S.; Demchenko, A. P. J. Am. Chem. Soc. 2002, 124,
12372.
9
J. AM. CHEM. SOC. VOL. 131, NO. 39, 2009 14045

A R T I C L E S
Park et al.
Figure 2. Diagram of the ESIPT process in the white-light-emitting molecular dyad. Calculated energy level in W1, with the HPI moiety on the left and
HPNI on the right, according to TD-DFT calculations.
of frustrated energy transfer; λmax ) 580 nm; see Scheme S4 in
Supporting Information (SI)) as constituent emitting centers for
white-light emission.13,31 However, it appeared that the rigid
molecular structure of DOX made it insoluble in common
organic solvents and thus made the synthesis and handling of
DOX fragments difficult. Fortunately, a simpler alternative
orange keto-emitting ESIPT unit was successfully synthesized
by replacing the phenyl group of HPI by the naphthalene group
[3-(1,4,5-triphenyl-1H-imidazol-2-yl)naphthalen-2-ol (HPNI) of
Figure 1b]. In the newly synthesized HPNI molecule, the
naphthalene group significantly stabilized the energy of the
excited-state keto form by enhancement of conjugation via
delocalization of π electrons along the naphthalene backbone
(the dashed line on the HPNI structure).32 Consequently, HPNI
showed a huge bathochromic shift of photoluminescence (PL)
compared to HPI (λmax ) 570 vs 470 nm). The emission was
tuned even more to the red (λmax ) 590 nm) by linking the
4,5-phenyl rings of HPNI to synthesize 3-(1-phenyl-1H-phenan-
thro[9,10-d]imidazol-2-yl)naphthalen-2-ol (HPNIC, Figure 1b).
Finally, by covalently linking molecular pairs, the dyad mol-
ecules W1 and W2 were synthesized (Figure 1c).
In the design of W1 and W2 molecular dyads, an ether linkage
between the fluorophores was chosen because diphenylether is
a simple but chemically very stable cross-linking group.33
Quantum-chemical calculations on W1 and W2 using density
functional theory (DFT B3LYP/6-311G**, see SI for details)
indicated an almost perpendicular molecular conformation of
the phenyl rings within the diphenylether unit (Figures S1-S3
in SI). This further reduces the electronic communication
between the ESIPT moieties and limits intermolecular stacking
in the solid state.34,35
Although the synthetic route to the W1 and W2 dyads (see
Schemes S1 and S2 in SI) comprised only two initial steps, the
reaction scheme was later adjusted to include four steps for
convenience of purification of the crude product (compare
Schemes S1 and S3 in SI). Briefly, the HPI derivative with an
acetamide side group was synthesized by refluxing salicylal-
dehyde, benzil, 4,4′-oxidianiline, and ammonium acetate in
acetic acid. Next, the acetamide group was deblocked by
hydrochloric acid to give the free amine. To synthesize the HPNI
and HPNIC materials, 2-hydroxy-3-naphthaldehyde was pre-
pared by direct formylation of 2-naphthol using tert-butyllithium
at room temperature. Finally, W1 and W2 were obtained with
high yields of 65 and 60%, respectively. Note that easy synthesis
of the W1 and W2 dyads are essential for real applications in
optoelectronic and lighting devices.
Figure 3a shows the UV/vis absorption and PL spectra of
the individual chromophores HPI, HPNI, and HPNIC in solution
(CHCl3, c ) 1 × 10-5 M). Common electronic absorption by
HPI and HPNI was found in the region 320-350 nm, where
both species can be simultaneously photoexcited. Low-energy
absorption features correspond to singlet π-π* excitation of
the enol ground state (E) to the first excited state, E f E* (S0
f S1), consistent with the time-dependent (TD)-DFT calcula-
tions, as shown in Figure 2; for details see SI. After photoex-
citation (E f E*), intramolecular proton transfer (ESIPT)
generates the keto excited state (K*), calculated to be lower in
energy than E* by 0.5 and 0.84 eV for HPI and HPNI,
respectively. Next, intense keto PL with maxima at 470 nm
(HPI), 570 nm (HPNI), and 590 nm (HPNIC) were observed
(see Table 1). The large Stokes’ shifts of about 8600, 9500,
and 10 100 cm-1, respectively, confirmed that emission arose
(31) Park, S.; Kwon, O. H.; Kim, S.; Park, S.; Choi, M. G.; Cha, M.; Park,
S. Y.; Jang, D. J. J. Am. Chem. Soc. 2005, 127, 10070.
(32) Nagaoka, S.; Nakamura, A.; Nagashima, U. J. Photochem. Photobiol.
A 2002, 154, 23.
(34) Ahlrichs, R.; Bar, M.; Haser, M.; Horn, H.; Kolmel, C. Chem. Phys.
Lett. 1989, 162, 165.
(35) (a) Wu, Y. G.; Lawson, P. V.; Henary, M. M.; Schmidt, K.; Bredas,
J. L.; Fahrni, C. J. J. Phys. Chem. A 2007, 111, 4584. (b) Barbatti,
M.; Aquino, A. J. A.; Lischka, H.; Schriever, C.; Lochbrunner, S.;
Riedle, E. Phys. Chem. Chem. Phys. 2009, 11, 1406.
(33) Ma, H.; Chen, B. Q.; Sassa, T.; Dalton, L. R.; Jen, A. K. Y. J. Am.
Chem. Soc. 2001, 123, 986.
9
14046 J. AM. CHEM. SOC. VOL. 131, NO. 39, 2009

A White-Light-Emitting Molecule
A R T I C L E S
Table 1. Optical and Photophysical Properties of the Compounds
under Study: Absorption and PL Maxima (λmax,abs, λmax,em), Molar
Extinction Coefficients (ε), PL Quantum Yields, and Decay Times
broad emission because of an increase in HPI absorption (ε )
1.77 × 104 M-1 cm-1), and emission was thus comparable to
that of HPNI (ε ) 1.71 × 104 M-1 cm-1). Similarly, at 350 nm
excitation wavelength, the emission became more orange
because of the lower absorption of HPI compared to that of
HPNI (details are given in Figure S6 in SI).
The actual emission spectrum of W1 or W2 can be simply
approximated by a sum of emissions from the individual
components from the ratio of their extinction coefficients and
emission intensities obtained under these conditions, as seen
by comparison of the green solid and dashed lines in Figure
3b,c. Note that the summation procedure is applicable at any
other excitation wavelength, as shown in Figure S6 in SI,
yielding unambiguous evidence for frustrated energy transfer
between the emitting components of the dyad structure. On the
basis of the same reasoning, the quantum yield ΦPL of the dyad
was estimated from those of its components (c1,c2) as follows:
(ΦPL, τPL
)
b
λ
max,abs, nm
εa L mol-1 cm-1
λ
max,em, nm
ΦPL τPL (ns) at λ (nm)
HPI
HPNI 331
HPNIC 337
320
0.42 × 104 470
1.14 × 104 570
1.14 × 104 590
1.59 × 104 HPI: 472
HPNI: 564
0.35
0.22
0.16
0.27
2.0 (450)
5.9 (580)
3.8 (580)
2.5 (450)
6.5 (680)
3.4 (450)
2.7 (580)
W1
HPI: 320
HPNI: 332
HPI: 320
W2
1.59 × 104 HPI: 473
HPNIC: 586
0.22
HPNIC: 333
a HPI, HPNI, W1 (346 nm); HPNIC, W2 (345 nm). b The PL
quantum yields (ΦPL) of HPI, HPNI, HPNIC, W1, and W2 were
measured using an integrating sphere and an emission-tunable CW
He-Cd laser.16,17
from K* f K transitions, whereas E* f E emission was seen
as only very weak shoulders.13,25 After reaching the keto ground
state (K), the enol form (E) was instantaneously recovered, with
the latter being about 1.0 eV lower than K with no transient
energy barrier; see Figure 2. The excitation spectra, monitored
at the respective emission bands, were spectrally identical to
the lowest absorption band, indicating that both PL bands
originated from the same initially excited species. We wish to
stress that absorptions of the orange-emitting HPNI and HPNIC
molecules are located in the UV region and show no significant
spectral overlaps with HPI emission. The other noteworthy
feature of HPNI is that the ground state of actual emitter (K2)
is only transient and thus statistically unpopulated. Therefore,
in a 1:1 mixture of donor (D ) HPI) and acceptor (A ) HPNI,
HPNIC), energy transfer (which is proportional to the D-A
electronic interaction including spectral overlap)36 is virtually
blocked, unlike in common D-A systems containing ordinary
blue- and orange-emitting molecules.
On the basis of this theoretical background, we expected to
see broad white-light emission after 1:1 blending or covalent
tethering of the blue- and orange-light-emitting components.
Indeed, in the synthesized molecular dyads W1 and W2, the
full spectrum of visible light was covered by the two simulta-
neously occurring emissions (Figure 3b,c). Because energy
transfer is completely blocked between the emitting centers, the
dyad emissions are simply the superposition of those from the
individual centers, as shown in Figure 3b,c for W1 and W2,
respectively.
The actual emission color of the dyad is quite sensitive to
and precisely controllable by the excitation wavelength because
the extinction coefficients of HPI and HPNI (HPNIC) are not
identical within their entire absorption ranges. This is an
interesting phenomenon since precise color-tuned emission from
a single molecule using varying excitation wavelengths is very
uncommon. In particular, as depicted in Figure 3, W1 and W2
could be tuned to exhibit pure white emission by 346 and 345
nm excitation, respectively (see photographs in Figure 4a),
where blue- and orange-light emission intensities were well
equilibrated. For W1, the molar absorption coefficients (ε values)
of HPI and HPNI at 346 nm were 4.2 × 103 and 1.14 × 104
M-1 cm-1, respectively, but the emission intensity ratio between
HPI and HPNI in Figure 3b was less pronounced because of
the higher PL quantum yield of HPI compared to HPNI (Table
1). At 320 nm excitation, however, W1 showed a blue-dominant
ΦPL(c1 + c2) ) {ΦPL(c1)ε(c1) + ΦPL(c2)ε(c2)}/ε(c1 + c2)
(1)
For W1, with c1 ) HPI and c2 ) HPNI, the quantum yield
was estimated to be ΦPL(HPI + HPNI) ) 0.28, quite close to
the measured value of ΦPL(W1) ) 0.27 (see Table 1). In our
study, the fluorescence quantum yields of the W1 molecule as
well as its emission spectrum were found to be virtually invariant
with the environmental changes including the solvent polarity
and the added protic solvent (see Figure S13 in SI), unlike the
conventional ESIPT materials. This is uniquely attributed to the
strong and planar intramolecular H bond in HPI enclosed with
bulkier steric groups like 1- and 5-phenyl substituents.
For comparison, we investigated HPI-DCM (Figure 1c),
containing a classical (non-ESIPT) fluorophore. This compound
showed almost completely red-shifted orange emission with only
a small shoulder in the blue HPI region, and simple summation
of components was obviously inapplicable (see Figure 3d). This
might be attributed to efficient energy transfer from HPI to the
DCM fragment because of the great spectral overlap between
the HPI emission and the DCM absorption at around 470 nm
(Figure 3d and Figure S8 in SI). This observation emphasizes
the beneficial features of frustrated energy transfer in the W1
and W2 dyads.
Further evidence for frustrated energy transfer was derived
from fluorescence lifetime measurements.35,37 Unlike conven-
tional observations on resonant energy transfer, where the
presence of acceptor molecules leads to shorter fluorescence
donor decay times (τPL D values), the donor unit (HPI) lifetimes
in W1 with τPL,D ) 2.5 ns and in W2 with τPL,D ) 3.4 ns were
comparable to or even slightly greater than that of HPI
(2.0 ns); see Table 1 and Figure S10 in SI for details. Because
of the total frustration of energy transfer, the white-light
emissions from W1 and W2 were also completely independent
of concentration or state. When we measured PL spectra at
different concentrations, the peak-normalized spectra of all
samples were identical (Figure 4c). Furthermore, white-light
emission was observed not only in solution but also in the solid
state (thin films of dyes dispersed in poly(methylmethacrylate),
PMMA; see Figure 4a). In addition to the frustrated energy
transfer between D-A units, it is considered that both twisted
ether linkages and bulky aromatic rings in the HPI and HPNI
units assist in preventing direct stacking of active chromophores,
thereby maintaining appropriate intermolecular distances and
suppressing concentration-dependent PL quenching, and thus
permitting strong emission.31
(36) Turro, N. J. Modern Molecular Photochemistry; University Science
Books: Sausalito, CA, 1991; Chapter 9.
(37) Jarvis, G. B.; Mathew, S.; Kenny, J. E. Appl. Opt. 1994, 33, 4938.
9
J. AM. CHEM. SOC. VOL. 131, NO. 39, 2009 14047

A R T I C L E S
Park et al.
Figure 3. Optical properties of synthesized molecules in solution (CHCl3, c ) 1 × 10-5 M) at room temperature. (a) Absorption (Abs) and normalized
photoluminescence (PL) of HPI (blue), HPNI (green), and HPNIC (red). (b) Abs and PL (excitation at 346 nm) spectra of HPI (blue), HPNI (red), and W1
(green solid line). The green dashed line indicates the expected spectrum of W1 as a superposition of HPI and HPNI (see text). (c) Abs and PL (excitation
at 345 nm) spectra of HPI (blue), HPNIC (red), and W2 (green solid line). The green dashed line indicates the expected spectrum of W2 as a superposition
of HPI and HPNIC. (d) Abs and PL (excitation at 345 nm) spectra of HPI (blue), DCM-EtOH (red), HPI-DCM (green solid line), and the expectation
(green dashed line).
Finally, the proposed PL additivity from complementary
emitting centers was verified by examination of CIE coordinates
(x,y) of compound emission colors. Initial CIE coordinates of
HPI (0.16, 0.23), HPNI (0.49, 0.50), and HPNIC (0.52, 0.47)
clearly lay on a straight line (see dashed line in Figure 4b),
passing through the ideal white values (0.33, 0.33), thus
constituting complementary colors. It is noteworthy that the CIE
coordinates of both solutions and films were similar. The slight
differences may be attributed to the higher polarizability of the
surrounding medium in the solid state.38,39
To synthesize single-molecule-based white-light-emitting
organic light-emitting devices, using W1 as the emission layer,
two simple electroluminescence (EL) devices were fabricated
with a configuration of (A) ITO/2-TNATA(60 nm)/NPD (20
nm)/EML(W1, 40 nm)/BPhen(20 nm)/LiF(1 nm)/Al(100 nm)
and (B) ITO/NPD(40 nm)/EML(W1, 30 nm)/BPhen(50 nm)/
LiF(1 nm)/Al(100 nm), with and without insertion of a hole
injection layer, respectively (see Figures S11 and S12 and Table
S1 in SI for details). The W1 electroluminescence (EL) spectra
from either A or B appeared very similar to the PL spectra,
with blue and orange emission bands, generated from the HPI
and HPNI moieties, respectively (Figure 4d). The development
of EL is attributable to the presence of an electrically initiated
ESIPT in each constituent emitting center (HPI/HPNI); the
exciton is formed by recombination of an electrically generated
hole E+ and an electron E- within the emitting layer, yielding
E*, which then decays via K*, as in PL.13,23 However, a
significant dependency of HPI and HPNI emission intensity
ratios on the device structure was observed (Figure 4d), which
can be rationalized by DFT calculations. Of the potential
electron-transporting species (HPI-, HPNI-) it was calculated
that HPNI- is preferentially formed because of the much higher
electron affinity (EA) compared to HPI (Table S3 in SI).
However, the ionization potentials of both HPI and HPNI are
of the same order, suggesting that formation of both HPI+ and
HPNI+ might sensitively depend on the local environment.
Considering that these cationic species are responsible for
primary exciton formation in geminate recombination, a sensi-
tive dependency of emission characteristics on device morphol-
ogy, particularly the presence or absence of a hole injection
layer, was both expected and observed. Much like the excitation
wavelength control on the emission color (vide supra), the
injection and transport control by the device structure in EL is
rationalized to work similarly.
In contrast to the behavior of the more blue-side emitting
device A, device B, without the 2-TNATA layer, yielded time-
and bias-stable white-light EL with CIE (x,y) coordinates of
(0.34, 0.29); see Figure 4b. The stable EL spectra were probably
attributed to the high Tg of W1 (∼143 OC) and W2 (∼153 OC)
molecules, which prevents any morphological changes of the
emission layer under device operation. The turn-on voltage was
6.7 V, and the maximum luminous efficiency was 0.98 cd/A,
corresponding to an external quantum efficiency of 0.76%. The
maximum brightness of the device was 1092 cd/m2 (see Figures
S11 and S12 and Table S2 in SI for details). Although EL device
performance has not yet been optimized, it is encouraging that
a single-molecule-based white-light-emitting OLED has been
(38) Gierschner, J.; Cornil, J.; Egelhaaf, H. J. AdV. Mater. 2007, 19, 173.
(39) Cornil, J.; Beljonne, D.; Calbert, J. P.; Bredas, J. L. AdV. Mater. 2001,
13, 1053.
9
14048 J. AM. CHEM. SOC. VOL. 131, NO. 39, 2009

A White-Light-Emitting Molecule
A R T I C L E S
Figure 4. Photo- and electroluminescence from the white-light-emitting molecular dyads. (a) Photographs of HPI, W1, and HPNI molecules in solution (1
× 10-5 M in CHCl3) and thin films of the compounds on PMMA (c ) 1 wt %) under UV light (365 nm hand-held UV lamp, 1.2 mW ·cm-2). (b) Emission
colors in a CIE 1931 chromaticity diagram: Ideal white (4, 0.33, 0.33); HPI (3, 0.16, 0.23); HPNI (>, 0.49, 0.50); HPNIC (<, 0.52, 0.47); PL of W1 (],
0.33, 0.37); EL of W1 (O, 0.34, 0.29); EL of W2 (0, 0.33, 0.35). (c) PL spectra of W1 at different concentrations (1 × 10-3 to 1 × 10-5 M in CHCl3). The
excitation wavelength was 346 nm. The inset shows the normalized spectra of each sample. (d) EL spectra of device A (blue) and device B (green) at 10
mA/cm2. The inset shows the operation image of device B at 10 mA/cm2.
Acknowledgment. This work was supported by the Korea
Science and Engineering Foundation (KOSEF) through the National
Research Laboratory Program (2006-03246) and Creative Research
Initiatives Program (0417-20090011) funded by the Ministry of
Education, Science of the Korea Government. We are grateful for
the instrumental and partial financial support of Dongwoo FineChem
Co., Ltd. D.-J.J. and S.-Y.P. thank the SRC program (R11-2007-
012-01002-0) and the BK 21 Program, respectively. B.M.M. thanks
the Ministerio deCiencia e Innovacio´n of Spain (MCI) for a “Juan
de la Cierva” contract. J.G. is a “Ramo´n y Cajal” Research Fellow,
financed by the MCI; at present he is a visiting researcher at the
ICMol, Valencia, under the auspices of the Consolider Ingenio 2010
Nanoscience program. J.G. and B.M.M. thank P. Viruela for helpful
discussions.
fabricated. Finally, it should be mentioned that our HPI, HPNI,
and HPNIC moieties showed excellent photostability upon UV
irradiation. Even when molecules were repeatedly exposed to
355 nm picosecond Nd:YAG laser pulses (∼5 × 105 shots), no
discernible photobleaching or decomposition was observed using
an excitation power of 2 mJ cm-1 pulse-1
.
In conclusion, an ideal material for a white-light source
should be cost-effective, stable, robust, emit over the whole
visible spectrum, not suffer from self-absorption, and its pure
color should be easily reproducible.40 With this goal in mind,
we have successfully synthesized and characterized, for the
first time, a white-light-emitting single molecule dyad,
consisting of two noninteracting chromophores showing
excited-state intramolecular proton transfer (ESIPT). This
unprecedented example of a concentration-independent ul-
timate white-light-emitting molecule not only blocks energy
transfer between a wide band-gap donor and a smaller band-
gap acceptor but also shows reproducible and stable white
photo- and electroluminescence. The molecule is expected
to have a significant impact on the development of a new
class of displays and light sources.
Note Added after ASAP Publication. The version published
on May 29, 2009 contained an error in equation 1. The corrected
version was published July 13, 2009.
Supporting Information Available: Synthesis, measurements,
calculation, and EL device fabrication details are given. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(40) Bowers, M. J.; McBride, J. R.; Rosenthal, S. J. J. Am. Chem. Soc.
2005, 127, 15378.
JA902533F
9
J. AM. CHEM. SOC. VOL. 131, NO. 39, 2009 14049