Nearly 100% internal phosphorescence

efficiency in an organic light-emitting device

Cite as: Journal of Applied Physics 90, 5048 (2001); https://doi.org/10.1063/1.1409582
Submitted: 08 May 2001 • Accepted: 13 August 2001 • Published Online: 31 October 2001

Chihaya Adachi, Marc A. Baldo, Mark E. Thompson, et al.

ARTICLES YOU MAY BE INTERESTED IN

Electroluminescence of doped organic thin films

Journal of Applied Physics 65, 3610 (1989); https://doi.org/10.1063/1.343409

100% phosphorescence quantum efficiency of complexes in organic semiconductor films

Applied Physics Letters 86, 071104 (2005); https://doi.org/10.1063/1.1862777

Blue organic electrophosphorescence using exothermic host–guest energy transfer

Applied Physics Letters 82, 2422 (2003); https://doi.org/10.1063/1.1568146

Journal of Applied Physics 90, 5048 (2001); https://doi.org/10.1063/1.1409582 90, 5048

© 2001 American Institute of Physics.

JOURNAL OF APPLIED PHYSICS VOLUME 90, NUMBER 10 15 NOVEMBER 2001

Nearly 100% internal phosphorescence efficiency in an organic

light emitting device
Chihaya Adachi, Marc A. Baldo, Mark E. Thompson,a) and Stephen R. Forrest
Center for Photonics and Optoelectronic Materials (POEM), Department of Electrical Engineering,
Princeton University, Princeton, New Jersey 08544
共Received 8 May 2001; accepted for publication 13 August 2001兲
We demonstrate very high efficiency electrophosphorescence in organic light-emitting devices
employing a phosphorescent molecule doped into a wide energy gap host. Using
bis共2-phenylpyridine兲iridium共III兲 acetylacetonate 关 共ppy兲2Ir共acac兲兴 doped into 3-phenyl-4-
共1⬘-naphthyl兲-5-phenyl-1,2,4-triazole, a maximum external quantum efficiency of 共19.0⫾1.0兲% and
luminous power efficiency of 共60⫾5兲 lm/W are achieved. The calculated internal quantum
efficiency of 共87⫾7兲% is supported by the observed absence of thermally activated nonradiative loss
in the photoluminescent efficiency of 共ppy兲2Ir共acac兲. Thus, very high external quantum efficiencies
are due to the nearly 100% internal phosphorescence efficiency of (ppy兲2Ir共acac兲 coupled with
balanced hole and electron injection, and triplet exciton confinement within the light-emitting layer.
© 2001 American Institute of Physics. 关DOI: 10.1063/1.1409582兴

I. INTRODUCTION phosphorescent materials which harvest both singlet and trip-

let excitons, ␩ int can approach 100%, in which case we can
Inorganic, direct band gap semiconductor heterostructure anticipate ␩ ext⬃20%.
light-emitting devices and lasers comprised of III–V com- The recent fabrication of extremely efficient electrophos-
pounds based on GaAs and InP, are known to have nearly phorescent OLEDs employing Pt and Ir complexes suggests
100% internal quantum efficiency 共␩ int).1 Since holes and that devices with internal quantum efficiencies of 100% are
electrons in the valence and conduction bands, respectively, achievable through radiative recombination of both singlet
are considered as free particles, radiative and nonradiative and triplet excitons.4 –13 Attempts to observe electrophospho-
transitions in these direct band gap materials do not involve rescence 共EP兲 were reported using keto-coumarin14 and
the intermediate formation of excitons. In contrast, many op- benzophenone15 derivatives. The efficiency of these materi-
tical processes in organic thin films are mediated by excitons als was low, ␩ ext⬍1%, even at 77 K, because the rate of
which can ultimately lead to a reduction in electrolumines- phosphorescent light emission was comparable to the excited
cence efficiency in polymer and molecular organic light state nonradiative decay rate. However, organometallic com-
emitting devices 共OLEDs兲. Besides nonradiative pathways pounds which introduce spin–orbit coupling due to the cen-
due to strong exiton–phonon coupling, the fraction of singlet tral heavy metal atom show a relatively high ligand based
excitons 共␹兲 under electrical excitation is 0.25 in a molecular phosphorescence efficiency 共␾ p ⬎20%) even at room tem-
solid such as aluminum tris共8-hydroxyquinoline兲 (Alq3 ),2 perature because of the strong radiative transition moment of
limits ␩ int to only 25% when fluorescence-based light emit- n – ␲ * , ␲ – ␲ * and the metal-to-ligand charge transfer
ting molecules are employed. The OLED external quantum states.12 Using this approach, demonstrations of high-
efficiency ( ␩ ext) follows: efficiency 共␩ ext⬎5%) OLEDs were made possible using
2,3,7,8,12,13,17,18-octaethyl-21H,23H-porphine platinum
␩ ext⫽ ␩ int␩ ph⫽ ␥ ␩ ex␾ p ␩ ph , 共1兲 共II兲,4,5 and fac tris共2-phenylpyridine兲iridium 关Ir共ppy兲3] and
its derivatives.6 –13
where ␩ ph is the light out-coupling efficiency, ␩ ex is the frac- Here, we demonstrate very high efficiency EP-OLEDs
tion of total excitons formed which result in radiative transi- employing the green electrophosphorescent molecule, bis共2-
tions 共␩ ex⬃1/4 for fluorescent molecular dyes, and 1 for phenylpyridine兲 iridium共III兲acetylacetonate 关 (ppy兲2Ir共acac) 兴 ,
phosphorescent materials兲, ␥ is the ratio of electrons to holes with a maximum ␩ ext⫽(19.0⫾0.5)% and luminous power
共or vice versa, to maintain ␥ ⭐1兲 injected from opposite con- efficiency of ␩ p ⫽(60⫾5) lm/W.16 We show that these val-
tacts 共the electron–hole charge-balance factor兲, and ␾ p is the ues correspond to ␩ int⫽(87⫾7)%. We also demonstrate that
intrinsic quantum efficiency for radiative decay 共including the very high internal quantum efficiency is due to direct
both fluorescence and phosphorescence兲. If only singlets are exciton formation at the guest phosphor, along with subse-
radiative in fluorescent materials, ␩ ext is limited to ⬃5%, quent exciton confinement within the emissive layer.
assuming3 ␩ph⬃1/2n 2 ⬃20% for a glass substrate with index
of refraction n⫽1.5. In contrast, by using high efficiency
II. EXPERIMENTAL METHOD
a兲
Present address: Department of Chemistry, University of Southern Califor- Organic layers were deposited by high-vacuum 共10⫺6
nia, Los Angeles, CA 90089. Torr兲 thermal evaporation onto a clean glass substrate pre-

0021-8979/2001/90(10)/5048/4/$18.00 5048 © 2001 American Institute of Physics

J. Appl. Phys., Vol. 90, No. 10, 15 November 2001 Adachi et al. 5049

FIG. 2. Concentration dependence of EL spectra on 共ppy兲2Ir共acac兲 of an

ITO/HMTPD 共60 nm兲/1%, 2%, and 6%-共ppy兲2Ir共acac兲:TAZ 共25 nm兲/Alq3
FIG. 1. The external quantum and power efficiencies of an ITO/HMTPD
共50 nm兲/Mg:Ag/Ag OLED. Inset: Concentration dependence of the external
共60 nm兲/12%-共ppy兲2Ir共acac兲:TAZ 共25 nm兲/Alq3 共50 nm兲/Mg:Ag/Ag OLED.
quantum efficiency 共␩ ext) and driving voltage of a 共ppy兲2Ir共acac兲 device
A maximum external quantum efficiency of ␩ ext⫽(19.0⫾0.5)% and power
with a structure: ITO/HMTPD 共60 nm兲/X%-共ppy兲2Ir共acac兲:TAZ 共25 nm兲/
efficiency of ␩ p ⫽(60⫾5) lm/W were obtained. Inset: Molecular structure
Alq3 共50 nm兲 Mg:Ag/Ag 共X⫽1% to 20%兲.
of 共ppy兲2Ir共acac兲.

coated with an indium tin oxide 共ITO兲 layer 共160 mm thick兲

with a sheet resistance of ⬃20 ⍀/sq. Prior to use, the sub- Figure 2 shows the dependence of the 共ppy兲2Ir共acac兲 EP
strate was degreased with solvents and cleaned in a UV- spectrum on doping concentration in TAZ. At concentrations
ozone chamber before it was loaded into the evaporation ⬎6%, we only observe the electroluminescence 共EL兲 com-
system. First, a 60 nm thick hole transport layer ponent centered at a wavelength of ␭⫽520 nm due to the
of 4,4⬘-bis关N,N⬘-共3-tolyl兲amino兴-3,3⬘-diemthyl biphenyl (ppy兲2Ir共acac) phosphorescence. At 共ppy兲2Ir共acac兲 concen-
共HMTPD兲8 was deposited, followed by a 25 nm thick trations less than 2%, an additional deep blue emission 共␭
light-emitting layer 共EML兲 consisting of17 (ppy) 2 Ir共acac兲 ⫽395 nm兲 due to HMTPD fluorescence is also observed. The
codeposited with8 a 3-phenyl-4-共1⬘-naphthyl兲-5-phenyl- inset of Fig. 2 shows 共ppy兲2Ir共acac兲 concentration depen-
1,2,4-triazole 共TAZ兲 electron-transport host layer. A 50 nm dences of ␩ ext and drive voltage at 1 mA/cm2. A maximum
thick layer of Alq3 was then deposited onto the EML surface ␩ ext was observed at 共ppy兲2Ir共acac兲 concentrations from 5%
to transport and inject electrons into the EML. A shadow to 12%, while a significant decrease in ␩ ext was observed at
mask with 1 mm diameter openings was used to define the both higher and lower concentrations. In addition, a gradual
cathode consisting of a 150 nm thick Mg:Ag 共10:1兲 layer, decrease of the drive voltage was observed with an increase
with a 20 nm thick Ag cap. Similar results to those reported in 共ppy兲2Ir共acac兲 concentration.
here were also obtained employing a LiF/Al cathode. We also measured the temperature dependence of the EP
Current density 共J兲–voltage 共V兲–luminance 共L兲 charac- and photoluminescence 共PL兲 quantum efficiencies of the
teristics were measured using an HP4145B semiconductor EML from T⫽15 to 300 K. Figure 3 shows the temperature
parameter analyzer, with the quantum efficiency directly ob- dependence of the PL spectra in a 共ppy兲2Ir共acac兲 doped TAZ
tained by placing the OLED approximately 3 mm above the film. The temperature dependencies of the integrated PL and
center of a large diameter 共1.13 cm兲 calibrated Si photodi- EP intensities under a constant current density of J
ode. Since almost all of the emitted light from the OLED ⫽0.1 mA/cm2 are shown in the insets of Fig. 3. While the
substrate surface is detected by the photodiode, this method spectral width decreased slightly with temperature, both the
avoids systematic errors introduced by corrections needed to PL and EP intensities are temperature independent.
account for non-Lambertian spatial emission patterns.7 Fur-
thermore, this direct measurement of the quantum efficiency
also eliminates errors often introduced by first measuring
device luminance and then converting the data to efficiency
after determination of the OLED emission spectrum.18

III. RESULTS

Figure 1 shows the EP-OLED efficiency with 12%-

共ppy兲2Ir共acac兲 doped into TAZ. A maximum ␩ ext⫽(19.0
⫾0.5)% and ␩ p ⫽(60⫾5) lm/W are obtained at a current of
J⫽1.5 ␮A/cm2 and luminance of 1.3 cd/m2. The device ex-
hibits a gradual decrease in ␩ ext with increasing current at
FIG. 3. Temperature dependence of the PL spectrum of a 12%-
J⬎10 mA/cm2 due to triplet–triplet annihilation.19,20 Even 共ppy兲2Ir共acac兲:TAZ film from T⫽275 to 15 K. Inset: Temperature depen-
at a luminance of ⬃1000 cd/m2 corresponding to J dence of the PL intensity of a 12%-共ppy兲2Ir共acac兲:TAZ film and EP intensity
⫽2.1 mA/cm2, ␩ ext⫽(13.7⫾0.5)% was observed. of the OLED at J⫽0.1 mA/cm2.
5050 J. Appl. Phys., Vol. 90, No. 10, 15 November 2001 Adachi et al.

FIG. 4. Calculated light out-coupling factor for glass (n⫽1.5兲/ITO 共160

nm兲/organic layer 共125 nm兲/Mg:Ag/Ag. The abscissa shows the distance of FIG. 5. Absorption and emission spectra of 共ppy兲2Ir共acac兲. Inset: Energy
the light emitting layer from the Mg:Ag cathode. level diagram of 共ppy兲2Ir共acac兲. The ligand singlet 共1 Ligand) and triplet
( 3 Ligand) states, 共1 MLCT兲 and ( 3 MCLT) were determined from the absorp-
tion and emission spectra. ⌽ NF , ⌽ ISC , ⌽ PI , and ⌽ NP are quantum yields
for nonemissive transitions from 1 MLCT intersystem crossing, intrinisc
IV. DISCUSSION phosphorescent transitions, and nonemissive transitions from 3 MLCT, re-
spectively.
A. Internal electrophosphorescent quantum efficiency
To determine the internal quantum efficiency of the EP-
OLED, we begin by calculating the radiative modes in the ␩ int⫽ ␥␾ p ␩ ex⫽ ␥␬ p / 关 ␬ p ⫹ ␬ NP兴关共 1⫺ ␹ 兲 ⫹ ␹ ␾ ISC兴 . 共3兲
OLED optical microcavity21,22 employing the theory of From the temperature independence of the PL and EP effi-
Chance et al.23 Dyadic Green’s functions are used to com- ciencies 共see inset of Fig. 3兲, we infer that the temperature-
pute the radiative decay rates, allowing us to consider arbi- dependent nonradiative pathways are almost negligible in a
trarily complex structures. This treatment also accounts for 共ppy兲2Ir共acac兲:TAZ solid-state film even at room tempera-
nonradiative losses due to dipole coupling with surface plas- ture. Note, however, that some nonradiative processes 共e.g.,
mon modes at the metalorganic cathode interface. We model triplet–triplet and triplet–polaron annihilation, and field-
the OLED in Fig. 1 as a four-layer structure: the Mg:Ag dependent exciton dissociation, etc.兲 do not depend strongly
cathode is considered semi-infinite with a refractive index of on temperature. From our measurements of ␩ int⫽(87
n Mg⫽0.25⫹i4.36;24 the organic layers are represented by a ⫾7)%, Eq. 共3兲 implies that the 共13⫾7兲% loss in total effi-
single layer of thickness 125 nm with refractive index n ciency must arise either from residual nonradiative processes
⫽1.7;21 the ITO has a thickness of 160 nm and a refractive in 共ppy兲2Ir共acac兲 or because ␾ ISC or ␥ is less than 1. How-
index of n⫽1.9⫹i0.01,25 where the absorption was calcu- ever, it is unlikely that ␾ ISC⬍1 since, in that case, we would
lated from transmission measurements; and finally the semi- anticipate fluorescent emission from the 1 MCLT to ground
infinite glass layer has a refractive index of n⫽1.5. The glass state on a time scale of ⬃10 ns, but this emission is not
substrate is approximately 1 mm thick, allowing for the use observed at any temperature or pumping intensity used in
of ray optics to calculate the angular emission pattern within these studies. Hence, we conclude that both the photolumi-
the glass, and consequently the coupling into air. For lumi- nescent efficiency of 共ppy兲2Ir共acac兲 and ␥ are at least ⬃0.9.
nescence at the HMTPD 共ppy兲2Ir共acac兲:TAZ interface 75 nm
from the cathode, we calculate an output coupling efficiency
B. Exciton formation process
of ␩ p ⫽(22⫾2)% 共Fig. 4兲. Hence, for ␩ ext⫽(19.0⫾0.5)%,
we obtain ␩ int⫽(87⫾7)%. The EP spectral characteristics in Fig. 2 suggest that at
With our estimate of ␩ int , we can infer the efficiencies 共ppy兲2Ir共acac兲 concentrations less than 2%, hole injection
of the molecular transitions leading to EP using Fig. 5, which from the HMTPD highest occupied molecular orbital
shows the absorption and emission spectra of 共ppy兲2Ir共acac兲 共HOMO兲 into the TAZ HOMO is energetically unfavorable,
identifying transitions from several excited state manifolds. and carrier recombination partly occurs within HMTPD 关c.f.
The energy level scheme inferred from these spectra is energy level diagram in Fig. 6共a兲兴. The large energy differ-
shown in the inset. Phosphorescence proceeds via either di- ence between the HOMO level of HMTPD and TAZ of ⬃1.0
rect injection into the triplet metal ligand charge transfer eV prevents direct hole injection from HMTPD into TAZ.
state ( 3 MLCT兲, or via intersystem crossing 共ISC兲 from the The accumulated holes at the interface then recombine with
singlet charge transfer state ( 1 MLCT兲. Now, the phosphores- electrons injected from TAZ layer, leading to blue HMTPD
cence quantum yield 共␾ p 兲 follows: emission in addition to exciton formation at 共ppy兲2Ir共acac兲.
This analysis is consistent with the dominant electron trans-
␾ p ⫽ ␬ p / 关 ␬ p ⫹ ␬ NP兴 , 共2兲
port characteristics of TAZ. At 共ppy兲2Ir共acac兲 concentrations
where ␬ p is the phosphorescence emission rate and ␬ NP that higher than 6%, on the other hand, there is no HMTPD blue
of nonemissive triplet decay. Under electrical excitation, on emission 共Fig. 2兲. Thus, we conclude that 共ppy兲2Ir共acac兲 ex-
the other hand, both singlet and triplet excitons are directly citon formation occurs directly on 共ppy兲2Ir共acac兲 from holes
created on either the guest or host molecules with a statistical injected from HMTPD and electrons primarily transported at
splitting of ␹⫽25% singlets and (1⫺ ␹ )⫽75% triplets.2 the TAZ lowest unoccupied molecular orbital 共LUMO兲 en-
Thus, ␩ int follows 关c.f. Eqs. 共1兲 and 共2兲兴: ergy 关Fig. 6共b兲兴. A decrease of the driving voltage with in-
J. Appl. Phys., Vol. 90, No. 10, 15 November 2001 Adachi et al. 5051

ACKNOWLEDGMENTS

The authors are grateful to Universal Display Corpora-

tion, the National Science Foundation MRSEC program, and
the Defense Advanced Research Projects Agency for their
support of this research.

1
L. A. Coldren and S. W. Corzine, Diode Lasers and Photonic Integrated
Circuits 共Wiley, New York, 1995兲, p. 54.
2
M. A. Baldo, D. F. O’Brien, M. E. Thompson, and S. R. Forrest, Phys.
Rev. B 60, 14422 共1999兲.
3
N. C. Greenham, R. H. Friend, and D. D. C. Bradley, Adv. Mater. 6, 491
共1994兲.
4
M. A. Baldo, D. F. O’Brien, Y. You, A. Shoustikov, S. Sibley, M. E.
Thompson, and S. R. Forrest, Nature 共London兲 395, 151 共1998兲.
5
D. F. O’Brien, M. A. Baldo, M. E. Thompson, and S. R. Forrest, Appl.
Phys. Lett. 74, 442 共1999兲.
6
M. A. Baldo, S. Lamansky, P. E. Burrows, M. E. Thompson, and S. R.
Forrest, Appl. Phys. Lett. 75, 4 共1999兲.
7
T. Tsutsui, M. J. Yang, M. Yahiro, K. Nakamura, T. Watanabe, T. Tsuji, Y.
Fukuda, T. Wakimoto, and S. Miyaguchi, Jpn. J. Appl. Phys., Part 2 38,
FIG. 6. Energy diagrams of the ITO/HMTPD/ L1502 共1999兲.
共共ppy兲2Ir共acac兲:TAZ/Alq3 /MgAg/Ag EP-OLED showing the relative posi- 8
C. Adachi, M. A. Baldo, and S. R. Forrest, Appl. Phys. Lett. 77, 904
tions of the HOMO and LUMO levels of the various organic layers, corre- 共2000兲.
sponding to dopant concentrations of 共a兲 ⬍2% 共ppy兲2Ir共acac兲 and 共b兲 ⬎6% 9
C.-L. Lee, K. B. Lee, and J.-J. Kim, Appl. Phys. Lett. 77, 2280 共2000兲.
共ppy兲2Ir共acac兲 in a TAZ host. 10
C. Adachi, M. A. Baldo, S. R. Forrest, S. Lamansky, M. E. Thompson, and
R. C. Kwong, Appl. Phys. Lett. 78, 1622 共2001兲.
11
C. Adachi, M. A. Baldo, M. E. Thompson, and S. R. Forrest, Bull. Am.
creasing phosphor concentration is evident for the direct hole Phys. Soc. 46, 863 共2001兲.
12
injection process. Since the HOMO levels of HMTPD and S. Lamansky, P. Djurovich, D. Murphy, F. Abdel-Razzaq, C. Adachi, P. E.
Burrows, S. R. Forrest, and M. E. Thompson, J. Am. Chem. Soc. 123,
共ppy兲2Ir共acac兲 are aligned at 5.6 eV, direct hole injection at 4304 共2001兲.
the HMTPD/EML interface should reduce the drive voltage, 13
C. Adachi, R. C. Kwong, and S. R. Forrest, Organic Electronics 2, 37
as observed. 共2001兲.
14
Recall that the electron–hole charge-balance factor, M. Morikawa, C. Adachi, T. Tsutsui, and S. Saito, 51st Fall Meeting, Jpn.
Soc. Appl. Phys., Paper 28a-PB-8 共1990兲.
␥ⱗ0.9. Since most holes injected into the 共ppy兲2Ir共acac兲 15
S. Hoshino and H. Suzuki, Appl. Phys. Lett. 69, 224 共1996兲.
HOMO level recombine with electrons at this interface, and 16
These results are similar to those recently reported for OLEDs employing
since hole transport in the doped TAZ layer is probably more the analogous phosphor, Ir共ppy兲3 doped into a hole-transporting host by
likely than electron transport through the HMTPD HTL, we M. Ikai, S. Tokito, Y. Sakamoto, T. Suzuki, and Y. Taga, Appl. Phys. Lett.
speculate that the hole density is slightly higher than that of 79, 156 共2001兲.
17
electrons, possibly leading to deviation in ␥ from its ideal S. Lamansky, P. Djurovich, D. Murphy, F. Abdel-Razaq, R. Kwong, I.
Tsyba, M. Bortz, B. Mui, R. Bau, and M. E. Thompson, Inorg. Chem. 40,
value of 1. After direct exciton formation on 共ppy兲2Ir共acac兲, 1704 共2001兲.
the exciton radiatively decays due to the wide energy gap of 18
J. Kido and Y. Iizumi, Appl. Phys. Lett. 73, 2721 共1998兲.
a TAZ host which confines triplet excitons on the guest mol- 19
C. Adachi, M. A. Baldo, and S. R. Forrest, J. Appl. Phys. 87, 8049 共2000兲.
ecule. No host fluorescence and phosphorescence even at a
20
M. A. Baldo, C. Adachi, and S. R. Forrest, Phys. Rev. B 62, 10967 共2000兲.
21
V. Bulovic, V. B. Khalfin, G. Gu, P. E. Burrows, D. Z. Garbuzov, and S. R.
low temperature thus ensures good exciton confinement on Forrest, Phys. Rev. B 58, 3730 共1998兲.
the phosphor guest. 22
J-S Kim, P. K. H. Ho, N. C. Greenham, and R. H. Friend, J. Appl. Phys.
88, 1073 共2000兲.
V. SUMMARY
23
R. R. Chance, A. Prock, and R. Sibley, Adv. Chem. Phys. 37, 1 共1978兲.
24
Handbook of Optical Constants of Solids 共Academic, Orlando, FL, 1985兲.
In conclusion, we demonstrated a very high-efficiency 25
J. S. Kim, P. K. H. Ho, N. C. Greenham, and R. H. Friend, J. Appl. Phys.
EP-OLED approaching 100% internal quantum efficiency. 88, 1073 共2000兲.
26
R. H. Jordan, L. J. Rothberg, A. Dodabalapur, and R. E. Slusher, Appl.
The high internal phosphorescence efficiency and charge bal-
Phys. Lett. 69, 1997 共1996兲.
ance in the structure are responsible for the high efficiency. 27
G. Gu, D. Z. Garbuzov, P. E. Burrows, S. Venkatesh, and S. R. Forrest,
From these results, we find that further increases in OLED Opt. Lett. 22, 396 共1997兲.
28
efficiency will only be obtained by developing schemes for C. F. Madigan, M. H. Lu, and J. C. Sturm, Appl. Phys. Lett. 76, 1650
increasing light out-coupling by incorporating 共2000兲.
29
M. Borodilsky, T. F. Kraus, R. Coccioli, R. Vrijen, R. Bhat, and E.
microcavities,21,26 shaped substrates,27,28 or an index match- Yablonovitch, Appl. Phys. Lett. 75, 1036 共1999兲.
ing medium29,30 in combination with the use of phosphores- 30
T. Yamazaki, K. Sumioka, and T. Tsutsui, Appl. Phys. Lett. 76, 1243
cent molecular dyes. 共2000兲.