Content uploaded by Gary James Richards
Author content
All content in this area was uploaded by Gary James Richards on Mar 19, 2015
Content may be subject to copyright.
April 15, 2004 / Vol. 29, No. 8 / OPTICS LETTERS 869
Optical amplification in a first-generation
dendritic organic semiconductor
Justin R. Lawrence, Graham A. Turnbull, and Ifor D. W. Samuel
Organic Semiconductor Centre and Ultrafast Photonics Collaboration, School of Physics and Astronomy,
University of St. Andrews, St. Andrews, Fife KY16 9SS, United Kingdom
Gary J. Richards and Paul L. Burn
Chemistry Research Laboratory, Department of Chemistry, University of Oxford, Mansfield Road, Oxford, OX1 3TA, United Kingdom
Received December 23, 2003
We report a study of a new class of organic semiconductor as an optical gain medium. We demonstrate am-
plification of violet light by use of stimulated emission in a solution of a first-generation bis-f luorene-cored
semiconducting dendrimer. Amplification is also observed in the solid state by means of amplified sponta-
neous emission in an optically pumped dendrimer planar waveguide. Gains of
36 dB cm
21
at 420 nm and
26 dB cm
21
at 390 nm in solution and 350 dB cm
21
in the solid state are obtained. These results show that
semiconducting dendrimers have potential as visible laser and amplifier materials. © 2004 Optical Society
of America
OCIS codes: 060.2280, 060.2300, 060.2310, 230.4000.
Organic semiconductors are proving to be successful
alternatives to inorganic materials for light-emitting
diodes, visible lasers, and optical amplifiers.
1–3
There
are three classes of organic semiconductors: thermal-
ly evaporated small molecules, solution-processable
conjugated polymers, and dendrimers. Of these,
dendrimers are the youngest technology and are be-
ginning to attract significant attention. They exhibit
high photoluminescence (PL) quantum yields and
are easily processed from solution. Semiconducting
dendrimers have a modular architecture and consist of
a core and dendrons, each of which contain conjugated
units, and solubility-conferring surface groups. This
allows for the independent tuning of the electrical, op-
tical, and processing properties.
4,5
Dendrimers have
significant advantages over semiconducting polymers,
for which attempting to tune the solubility can lead to
undesirable changes in the emission properties.
6
The
dendritic structure also allows for a wide range of
emissive chromophores that in many cases would be
unsuited for use in an unmodified form or as a repeat
unit of a polymer. Charge transport in dendrimer
films can be controlled by varying the dendrimer
generation, which changes the spacing of the cores.
Lupton
et al.
5
found in a tris(distyrylbenezenyl) amino
cored dendrimer a reduction in hole mobility with
increasing generation, indicating that hole transport
occurs by means of hopping between the cores of the
molecules. To date, most semiconducting dendrimer
work has centered on light-emitting diodes, and there
have been some excellent results.
7
Nonf luorescent and electrically insulating den-
drimers have recently been studied for lasers.
8,9
The
nonfluorescent dendrimers were used to separate
molecules of a conventional laser dye to prevent
aggregation.
8
Also, laser emission from a polymer
waveguide doped with a rhodamine-cored electrically
insulating dendrimer
9
has been reported. In this
Letter we demonstrate strong amplification of light in
a first-generation (G1) dendrimer, which was previ-
ously shown to be semiconducting by the time-of-f light
technique.
10
The molecular structure of the G1 den-
drimer is shown in the inset in Fig. 1(b). The core
consists of a bis-f luorene unit capped with two G1
triphenyl-based dendrons containing 2-ethyl-hexyloxy
surface groups. It was made by coupling the G1
boronic-acid-focused dendron with
7, 7
0
dibromobisflu-
orene using palladium catalysis. The absorption and
PL spectra of the G1 dendrimer solution (10
24
Min
a 1-cm cell) are shown in Fig. 1(a). The absorption
maximum is centered at 350 nm, whereas the PL has
two peaks (390 and 425 nm) and two shoulders (450
and 475 nm).
To establish the presence of gain, a solution of the
G1 dendrimer was used to amplify a weak probe pulse
to an intense pulse.
3
A nitrogen laser (500-ps pulses,
337 nm) was used as a common excitation source for
both the amplifier and probe dye laser. Its output
was split into two beams of
⬃600-mJ pulses that were
then focused to a stripe of 200 mm 3 10 mm to trans-
versely pump the cells containing the dendrimer and
stilbene 3 laser dye solutions. The output beam from
the dye laser was attenuated by use of neutral-density
filters and focused through the excited region of the
10-mm-long amplifier cell. The probe beam was then
detected with either an optical energymeter or a CCD
spectrograph.
To measure the magnitude of the gain, a solution
containing 0.31 g兾l(1.83 3 10
24
M) of bis-f luorene G1
dendrimer in toluene was used as an optical amplifier.
The intensity of the probe beam was measured before
and after passing through the 1-cm path-length ampli-
fier cuvette. The gain achieved as a function of probe
intensity is shown in Fig. 1(b). The maximum gain
achieved is 36 dB cm
21
at 420 nm and 26 dB cm
21
at
390 nm. Solutions of different concentrations yielded
lower gains, and we believe that the reduction in gain
at higher concentrations is due to a poor overlap of
0146-9592/04/080869-03$15.00/0 © 2004 Optical Society of America
870 OPTICS LETTERS / Vol. 29, No. 8 / April 15, 2004
Fig. 1. (a) Absorption and emission spectra of the den-
drimer solution. (b) Gain (at 420 nm) versus probe en-
ergy of the dendrimer optical amplifier. Inset, dendrimer
structure.
the probe beam with the small excited region of the
solution. At lower concentrations, meanwhile, the
pump beam was too weakly absorbed to yield the ex-
citation density necessary for substantial amplifi-
cation.
3
As the probe energy was increased, the
amplifier gain was found to drop as a result of gain
saturation. The curve in Fig. 1(b) is a theoretical fit
to the 420-nm data using the well-known expression
for a homogeneously saturated pulsed amplifier
11
:
E
OUT
苷 E
S
ln兵1 1 关exp共E
IN
兾E
S
兲 2 1兴exp共G兲其 . (1)
Here G is the small signal gain coefficient, E
S
is the
saturation energy, E
OUT
is the energy extracted from
the amplif ier, and E
IN
is the unamplified probe en-
ergy. We found that the data can be convincingly de-
scribed by Eq. (1), and from the fit we obtained a small
signal gain of 36 dB cm
21
.
Amplification in organic semiconductors arises from
stimulated emission in a four-level system similar
to organic laser dyes.
1
The stimulated emission
cross section s for the bis-f luorene dendrimer can
therefore be determined from G 苷 sN
ex
l, where N
ex
is the excitation density and l is the amplifier length.
We estimate the gain cross section to be s 艐 3.4 3
10
218
cm
2
. The accuracy of this value is limited by
incomplete knowledge of the excitation density as the
probe pulse travels through the amplifier. These
values compare reasonably well with the gain of
44 dB cm
21
and the cross section of 5.3 3 10
217
cm
2
at 600 nm obtained with poly[2-methoxy-5-(3
0
,
7
0
-dimethyloctyloxy)-paraphenylenevinylene] (OC
1
C
10
-
PPV)
3
and 5.0 3 10
217
cm
2
at 420 nm in t, t
0
-didecyloxy-
II-distyrlbenzene laser dye.
12
However, dendrimers
are less-developed materials than dyes and conjugated
polymers, leaving considerable scope for improvement.
This illustrates the potential of semiconducting den-
drimers for substantial gain in compact amplifiers.
Although high gain can be obtained in dendrimer
solutions, liquids are inconvenient for use in practical
devices. We therefore investigated the potential for
amplification in the solid state. Dendrimers have
important advantages over dyes for solid-state ampli-
fication. They can be spin coated to give good quality
films, and the dendritic structure greatly reduces
quenching caused by aggregation. These dendrimers
are semiconductors with the potential for electrically
pumped amplification in the future.
Slab waveguides were formed by spin coating the G1
dendrimer solution onto quartz substrates to produce
200-nm-thick films, which yield high absorption of the
pump beam. The waveguides were optically pumped
at 337 nm by the nitrogen laser described previously,
which was focused to a 280-mm-wide stripe near the
edge of the film. The energy of the pump pulse was
controlled by neutral-density filters, and light emitted
from the waveguide edge was collected with an optical
fiber and analyzed by a CCD spectrometer.
Figure 2(a) shows the normal PL of a dendrimer
film and the line-narrowed spectrum when the film
is pumped at an energy of 0.12 mJmm
21
. Comparing
the PL spectra of the solution and the solid state shows
that the shoulders and peaks of the two spectra are the
same and the small differences could be due to self-
absorption of emission in the film or differences in the
dielectric environment between the solution and the
solid state. The film PL quantum yield is 40%, sug-
gesting that aggregation does not dominate the photo-
physics. The emission spectrum was found to narrow
significantly with increasing pump energy and was
accompanied by a superlinear increase in intensity of
the output radiation. The linewidth dropped from 53
to 3.3 nm. The spontaneously emitted photons were
waveguided and amplified by stimulated emission, a
process known as amplified spontaneous emission
1,13
(ASE). ASE preferentially occurs where the gain ex-
ceeds the absorption and scattering losses by the great-
est amount. The ASE peak is at 425 nm, which is at
a longer wavelength than the PL maximum (418 nm)
because the ground-state absorption is reduced.
To quantify the level of gain from the ASE measure-
ments described above, an adjustable slit was placed
in the pump beam
13
to allow the length of the pumped
stripe to be varied. The emission was then measured
as the pump stripe length was increased from 0.3 to
1.8 mm and the pump energy from 20 to 120 nJ mm
21
.
Figure 2(b) shows the emission intensity at 425 nm as
a function of stripe length. The data were fitted using
I共l兲 苷 关共AP
0
兲兾g共l兲兴exp关g共l兲l兴 2 1 (Ref. 14) to extract a
value for the net gain coefficient g共l兲. The emission
intensity is I共l兲,AP
0
is the spontaneous emission
April 15, 2004 / Vol. 29, No. 8 / OPTICS LETTERS 871
Fig. 2. (a) Emission spectrum of a bis-fluorene dendrimer
film below and above ASE threshold. (b) ASE intensity as
a function of pump energy and stripe length.
proportional to the pump energy, and l is the pump
stripe length. At low pump energy (20 nJ mm
21
) the
net gain was found to be 10 cm
21
. At higher energy
(120 nJ mm
21
), net gains of 46 cm
21
were observed. A
maximum gain of 79 cm
21
(350 dB cm
21
) was achieved
while pumping with 30 mJmm
21
. These results
compare well with values reported for other materials,
e.g., 74 cm
21
for polyf luorene with a pump energy
of 9 mJ.
15
The loss coefficient of the dendrimer
waveguide at 425 nm was measured to be 1.4 cm
21
(6 dB cm
21
).
In summary, we have demonstrated strong am-
plification in both solutions and thin films of a new
class of organic semiconductor. Gains of 36 dB cm
21
in solution and 350 dB cm
21
in the solid state have
been obtained. A loss coefficient of 1.4 cm
21
has been
measured. These values are comparable with other
organic semiconductors, such as conjugated polymers.
Dendrimers offer considerable scope for molecular
design, e.g., changing the core chromophore should
make it possible to tune the gain across the visible
spectrum in a similar way to dendrimer light-emitting
diodes.
4
We therefore conclude that light-emitting
dendrimers show promise as a gain medium for optical
amplifiers.
The authors are grateful to the Engineering and
Physical Sciences Research Council, CDT Oxford,
Ltd., and the Royal Society for financial support.
I. Samuel’s e-mail address is idws@st-and.ac.uk;
P. Burn’s is paul.burn@chem.ox.ac.uk.
References
1. M. D. McGehee and A. J. Heeger, Adv. Mater. 12, 1655
(2000).
2. R. H. Friend, R. W. Gymer, A. B. Holmes, J. H.
Burroughes, R. N. Marks, C. Taliani, D. D. C. Bradley,
D. A. Dos Santos, J. L. Brédas, M. Logdlund, and
W. R. Salaneck, Nature 397, 121 (1999).
3. J. R. Lawrence, G. A. Turnbull, and I. D. W. Samuel,
Appl. Phys. Lett. 80, 3036 (2002).
4. M. Halim, I. D. W. Samuel, J. N. G. Pillow, and P. L.
Burn, Synth. Met. 102, 1113 (1999).
5. J. M. Lupton, I. D. W. Samuel, R. Beavington, P. L.
Burn, and H. Bässler, Adv. Mater. 13, 258 (2001).
6. M. Pope and C. E. Swenberg,
Electrical Processes in Or-
ganic Crystals and Polymers (Oxford Science, Oxford,
UK, 1999), Chap. 9.
7. J. P. J. Markham, S. C. Lo, S. W. Magennis, P. L. Burn,
and I. D. W. Samuel, Appl. Phys. Lett. 80, 2645 (2002).
8. S. Yokoyama, A. Otomo, and S. Mashiko, Appl. Phys.
Lett. 80, 7 (2002).
9. A. Otomo, S. Yokoyama, T. Nakahama, and S. Mashiko,
Appl. Phys. Lett. 77, 3881 (2000).
10. J. P. J. Markham, T. D. Anthopoulos, I. D. W. Samuel,
G. J. Richards, P. L. Burn, C. Im, and H. Bässler, Appl.
Phys. Lett. 81, 3266 (2002).
11. T. L. Koch, L. C. Chiu, and A. Yariv, Opt. Commun.
40, 364 (1982).
12. W. Holzer, A. Penzkofer, S. H. Gong, D. D. C. Bradley,
X. Long, and A. Bleyer, Chem. Phys. 224, 315 (1997).
13. M. D. McGehee, R. Gupta, S. Veenstra, E. K. Miller,
M. A. Diaz-Garcia, and A. J. Heeger, Phys. Rev. B 58,
7035 (1998).
14. K. L. Shaklee and R. F. Leheny, Appl. Phys. Lett. 18,
475 (1971).
15. G. Heliotis, D. D. C. Bradley, G. A. Turnbull, and
I. D. W. Samuel, Appl. Phys. Lett. 81, 415 (2002).