Download Amplification of optical pulse sequences at a high repetition rate in a

APPLIED PHYSICS LETTERS 91, 011105 共2007兲
Amplification of optical pulse sequences at a high repetition rate
in a polymer slab waveguide
D. Amarasinghe, A. Ruseckas, A. E. Vasdekis, G. A. Turnbull, and I. D. W. Samuela兲
Ultrafast Photonics Collaboration and Organic Semiconductor Centre,
SUPA, School of Physics and Astronomy, University of St. Andrews, North Haugh, St. Andrews,
Fife KY16 9SS, United Kingdom
共Received 8 May 2007; accepted 8 June 2007; published online 2 July 2007兲
Amplification of three short light pulses in a 140 ps time window at 5 kHz repetition rate has
been demonstrated using a compact amplifier based on the conjugated polymer
poly共9,9⬘-dioctylfluorene-co-benzothiadiazole兲. The amplifier was optically pumped and gratings
were used to couple the signal into and out of the film. A gain of 22 dB was observed for a signal
pulse temporally aligned with the pump pulse in a 1 mm waveguide. For a signal pulse delayed by
140 ps, the maximum gain achieved was 14 dB. The results are a step towards the use of polymer
amplifiers in data communications. © 2007 American Institute of Physics.
关DOI: 10.1063/1.2753542兴
There is an increasing demand for faster internet speeds
and digital services both at home and in the workplace, creating a need for short-haul data communications with high
bandwidth at low cost.1 Polymer optical fibers 共POFs兲 are
promising for this application because they are flexible, ductile under strain, have low manufacturing costs, and can be
installed on-site.1,2 Compatible optoelectronic devices, such
as mixers, amplifiers, and repeaters, are needed for the lowloss transmission window of 500– 660 nm of POF. Conjugated polymers are attractive media for lasers and optical
amplifiers because of their high optical gain, simple fabrication, ability to be pumped by diode lasers,3–5 and future potential to be electrically pumped.6 Polymer optical amplifiers
have previously been demonstrated, which amplify a single
light pulse with repetition rates in the range of hertz to a few
kilohertz in solution7,8 and in the solid state.9–11 Polymer
lasers, meanwhile, have recently been demonstrated to be
capable of bursts of pulses at repetition frequencies of a few
megahertz.12 However, optical data communications involves streams of light pulses at high repetition rates and a
practical amplifier needs to have identical gain for each
pulse. In this letter, we demonstrate and characterize the amplification of a short packet of pulses with 70 ps pulse spacing using a polymer optical amplifier.
To achieve uniform amplification of a train of signal
pulses, one requires a gain medium with a sufficiently long
gain lifetime. However, gain lifetime is shorter at high pump
energy densities due to amplified spontaneous emission
共ASE兲,9 so we have used a material with a low exciton diffusion coefficient in order to reduce the exciton-exciton annihilation rate. The conjugated polymer poly共9,
9⬘-dioctylfluorene-co-benzothiadiazole兲 共F8BT兲 was also
chosen because it is commercially available 共from American
Dye Source兲 and has a long fluorescence lifetime. Compared
with another commercially available and high gain polymer,
poly关2-methoxy-5-共2⬘-ethylhexyloxy兲-p-phenylene
vinylene兴, F8BT has a gain lifetime approximately eight times
longer. This allows us to achieve approximately even
a兲
Electronic mail: [email protected]
amplification
of
three
pulses
in
a
140 ps
window.
The amplifier was configured as a polymer slab waveguide, shown in Fig. 1共b兲. The active polymer layer was
deposited by spin coating to form a 250 nm thick film on a
silica substrate with gratings made by reactive ion etching.
The gratings were used to couple the signal pulses into the
polymer waveguide and out. The region in between the gratings was optically pumped to provide gain. To ensure that
there was no optical feedback, the output coupler was angled
at 20° to the optical axis.
The pump pulse 共at 497 nm兲 and signal pulses 共at
580 nm兲 were generated from an optical parametric amplifier
pumped by an amplified Ti:sapphire laser system at 5 kHz.
The 100 fs pulses were stretched to 10 ps in a TF-10 glass
block to match the resolution of the streak camera. The pulse
sequence was created using an optical interlever which consisted of a partial reflector and a mirror, as shown in Fig.
FIG. 1. 共Color online兲 共a兲 Schematic of the setup. 共b兲 Schematic of polymer
amplifier. 共c兲 Absorbance 共solid symbols兲, photoluminescence 共open symbols兲, amplified spontaneous emission 共dot-dash line兲 spectra of F8BT, and
signal spectra 共line兲. Arrow shows the position of the pump pulse.
0003-6951/2007/91共1兲/011105/3/$23.00
91, 011105-1
© 2007 American Institute of Physics
Downloaded 09 Jul 2007 to 138.251.31.216. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
011105-2
Appl. Phys. Lett. 91, 011105 共2007兲
Amarasinghe et al.
FIG. 2. Streak camera traces of amplified 共dash line兲 and nonamplified
pulses 共solid line兲. The pump energy density is 1.6 ␮J / cm2 and the signal
energy was 0.04 nJ in a waveguide length of 622 ␮m.
1共a兲. A quarter waveplate was used to rotate the polarization
so that the pulses were deflected from the input path by a
Glan polarizer. Amplification is achieved through stimulated
emission and the output of the pulse sequence was recorded
using a streak camera. The gain was calculated from the
intensity ratio of the amplified to the unamplified signal
energy, after subtraction of the photoluminescence
background,7,9
冋
G共dB兲 = 10 log
册
Pon − Pb
,
Poff
共1兲
where Pon is the intensity of the amplified signal, Pb is background 共generally Pb ⬍ 0.01 of Pon兲, and Poff is the signal
intensity without pump.
Figure 2 shows the streak camera traces of the amplifier
output with and without pump for a 622 ␮m waveguide. In
this case, all three pulses in the train were amplified by ⬃16
times. The gain of the amplifier was measured for a range of
waveguide lengths. Examples are shown in Fig. 3. The signal
pulse energy coupled into the amplifier was 0.04 nJ. The
gains for the first signal pulse were 19, 17, 15, and 8 dB for
waveguide lengths of 1022, 822, 622, and 322 ␮m, respectively. The gain for each peak increased with pump energy
FIG. 4. Gain dependence on signal energy for signal pulses delayed by 10,
70, and 140 ps with respect to the pump pulse in a 1022 ␮m long waveguide 共symbols兲 and the theoretical fit 共dash lines兲. The pump energy density
was kept constant at 2 ␮J / cm2. Signal energy was calculated from input
pulse energy taking into account a 20% light coupling efficiency into waveguide which corresponds to the value just after the input grating.
until saturation effects took place. The gain cross section was
calculated from the linear region of the plots 共at low pump
density兲 using the equation,
␴Nl = ln
Pon − Pb
,
Poff
共2兲
where N is the exciton density and l is the length of the
polymer guide. This gives ␴ = 3共±2兲 ⫻ 10−16 cm2, which is
similar to previously reported values.8,13,14
The gain dependence on signal energy was also measured with the three signal systems. Figure 4 shows the
variation for each pulse 共symbols兲, the theoretical curve
共dash lines兲 was fitted to the experimental results using the
well known equation for net gain in an optical
amplifier:8,15,16
G=
冋 冋 冉 冊 册册
C3
A
ln 1 + G0 exp
−1
C3
A
,
共3兲
where
C3 =
Ein共␴abs + ␴se兲
,
hvin
共4兲
where Ein is the signal energy input, G0 is the small signal
gain, ␴abs,se is the absorption 共abs兲 and stimulated emission
共se兲 cross section, h is Planck’s constant, vin is the signal
frequency, and A is the cross sectional area of the signal.
Fitting Eq. 共3兲 to the data gives G0 = 20 dB for the first two
pulses and G0 = 14 dB for the third pulse. Assuming that absorption at this wavelength is negligible 共␴abs = 0兲, the stimulated emission cross section was calculated to be 4共±1兲
⫻ 10−16 cm2 for all three curves and is in good agreement
with the value calculated using Eq. 共2兲. Finally, we compared
the time dynamics of a single signal pulse system with that
from the pulse sequence, shown in Fig. 5. The single signal
pulse was detected with a photodiode coupled to a lock-in
amplifier. In this case the pump and signal pulses were
100 fs long and the signal pulse energy was kept at 1 nJ just
after the input coupler. The gain decreases slightly with an
increase of time delay between pump and signal pulses. This
time dependence is confirmed using a variable time delay
FIG. 3. Gain dependence on pump energy density for signal pulses delayed
by 10, 70, and 140 ps with respect to the pump pulse on two waveguide
lengths of 1022 and 322 ␮m.
Downloaded 09 Jul 2007 to 138.251.31.216. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
011105-3
Appl. Phys. Lett. 91, 011105 共2007兲
Amarasinghe et al.
obtained in a 1 mm waveguide. Our results show promise
that polymer technology is growing towards a higher performance for use with low cost, short haul polymer optical
fibers.
The authors are grateful to the Engineering and Physical
Sciences Research Council and the Scottish Funding Council
for financial support, and to L. O’Faolain and T.F. Krauss for
assistance in making the grating couplers.
1
FIG. 5. Gain obtained for signal pulses at different time delays for a waveguide length of 1022 ␮m. The symbols show time dependence for three
signal pulses. Pump energy densities are 2.6 共open symbols兲 and 1.4 ␮J / cm2
共solid symbols兲 using a signal energy of 0.04 nJ. The solid line shows the
time dependence for a single signal pulse with a pump energy density of
2.8 ␮J / cm2, measured using a lock-in detection for which a higher signal
energy had to be used 共1 nJ兲, which resulted in a lower gain. The dotted
lines are a guide for the eye.
and lock-in detection 共solid line兲, for which a higher signal
pulse energy had to be used, resulting in lower gain values.
Optical amplifiers are needed to be able to amplify a
long sequence of optical pulses at high repetition rates. The
work in this letter is a step toward this goal, showing that
multiple pulses can be amplified. An improvement would be
to adjust the length of pump pulse to match the signal pulse
sequence. This would enable a longer pulse train to be amplified with each pulse experiencing the same amplification
factor. Stability is important for practical applications. We
were able to stably amplify 107 pulse sequences at pump
energy densities ⬎1 ␮J / cm2, which gave gains of ⬎12 dB.
The experiment was conducted in vacuum. For practical applications, the device would have to be encapsulated to prevent photodegradation due to contact with oxygen and water
vapor, such as in organic light emitting diodes.17
We have shown that F8BT gives sufficiently long lived
gain to demonstrate amplification of three signal pulses in a
compact solid state structure. A high gain of up to 20 dB was
I. T. Monroy, H. P. A. Vd Boom, A. M. J. Koonen, G. D. Khoe, Y.
Watanabe, Y. Koike, and T. Ishigure, Opt. Fiber Technol. 9, 159 共2003兲.
2
Y. Koike, T. Ishigure, M. Satoh, and E. Nihei, Pure Appl. Opt. 7, 201
共1998兲.
3
I. D. W. Samuel and G. A. Turnbull, Chem. Rev. 共Washington, D.C.兲 107,
1272 共2007兲.
4
A. E. Vasdekis, G. Tsiminis, J.-C. Ribierre, Liam O’Faolain, T. F. Krauss,
G. A. Turnbull, and I. D. W. Samuel, Opt. Express 14, 9211 共2006兲.
5
T. Riedl, T. Rabe, H.-H. Johannes, W. Kowalsky, J. Wang, T. Weimann, P.
Hinze, B. Nehls, T. Farrell, and U. Scherf, Appl. Phys. Lett. 88, 241116
共2006兲.
6
C. Pflumm, C. Karnutsch, M. Gerken, and U. Lemmer, IEEE J. Quantum
Electron. 41, 316 共2005兲.
7
J. R. Lawrence, G. A. Turnbull, and I. D. W. Samuel, Appl. Phys. Lett. 80,
3036 共2002兲.
8
G. Heliotis, D. D. C. Bradley, M. Goossens, S. Richardson, G. A.
Turnbull, and I. D. W. Samuel, Appl. Phys. Lett. 85, 6122 共2004兲.
9
D. Amarasinghe, A. Ruseckas, A. E. Vasdekis, M. Goossens, G. A.
Turnbull, and I. D. W. Samuel, Appl. Phys. Lett. 89, 201119 共2006兲.
10
M. Goossens, G. Heliotis, G. A. Turnbull, A. Ruseckas, J. R. Lawrence, R.
Xia, D. D. C. Bradley, and I. D. W. Samuel, Proc. SPIE 28–36, 5937
共2005兲.
11
M. A. Reilly, C. Marinelli, C. N. Morgan, R. V. Penty, I. H. White, M.
Ramon, M. Ariu, R. Xia, and D. D. C. Bradley, Appl. Phys. Lett. 85, 5137
共2004兲.
12
T. Rabe, K. Gerlach, T. Riedl, H.-H. Johannes, W. Kowalsky, J.
Niederhofer, W. Gries, J. Wang, T. Weimann, P. Hinze, F. Galbrecht, and
U. Scherf, Appl. Phys. Lett. 89, 081115 共2006兲.
13
R. Xia, G. Heliotis, Y. Hou, and D. D. Bradley, Org. Electron. 4, 165
共2003兲.
14
M. Zavelani-Rossi, S. Perissinotto, G. Lanzani, M. Salerno, and G. Gigli,
Appl. Phys. Lett. 89, 181105 共2006兲.
15
T. K. Koch, L. C. Chiu, and A. Yariv, J. Appl. Phys. 53, 6047 共1982兲.
16
M. A. Reilly, B. Coleman, E. Y. B. Pun, R. V. Penty, I. H. White, M.
Ramon, R. Xia, and D. D. C. Bradley, Appl. Phys. Lett. 87, 231116
共2005兲.
17
S. K. Park, J. Oh, C. Hwang, J. Lee, Y. S. Yang, H. Y. Chu, and K. Kang,
ETRI J. 27, 545 共2005兲.
Downloaded 09 Jul 2007 to 138.251.31.216. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp