APPLIED PHYSICS LETTERS 89, 201119 共2006兲
Broadband solid state optical amplifier based on a semiconducting polymer D. Amarasinghe, A. Ruseckas, A. E. Vasdekis, M. Goossens, G. A. Turnbull, and I. D. W. Samuela兲 Ultrafast Photonics Collaboration and Organic Semiconductor Centre, School of Physics and Astronomy, SUPA, University of St Andrews, North Haugh, St Andrews, Fife KY16 9SS, United Kingdom
共Received 25 July 2006; accepted 5 October 2006; published online 17 November 2006兲 A compact, solid state optical amplifier based on the conjugated polymer poly关2-methoxy-5共2⬘-ethylhexyloxy兲-p-phenylene vinylene兴 has been demonstrated. The amplifier was optically pumped. Gratings were used to couple the signal into and out of the film. The transmitted signal was amplified over 100 times in a 1 mm long waveguide giving 21 dB gain at 630 nm. A gain of ⬎13 dB was observed at 615 and 650 nm giving a gain bandwidth of ⬎26 THz. The gain dynamics at pump densities below 5 J / cm2 are described by an exciton-exciton annihilation model. At higher pump intensities, amplified spontaneous emission and photoinduced losses become significant. © 2006 American Institute of Physics. 关DOI: 10.1063/1.2390649兴 Polymers are attractive materials for optical components because they are ductile, have low manufacturing costs, and are versatile in their processing techniques. Polymer integrated circuits and polymer optical fibers are promising key technologies for short-haul data transmission. Such passive components are favored to replace twisted pair cables at remote stations and hence reduce the bottleneck in local area networks.1 Simple integration of active components such as optical amplifiers or switches could dramatically widen application areas. Semiconducting polymers combine these manufacturing advantages with attractive optical and electrical properties and offer potential as broadband optical gain media. Polymer lasers have been widely studied,2–7 but very little work has been done on optical amplifiers.8–10 The potential of high optical gain in semiconducting polymers has previously been shown in solution.9 For practical applications, a solid state amplifier would be more suitable. Compared with laser dyes,10 conjugated polymers exhibit less concentration quenching of the luminescence and can even be highly emissive in undiluted films. This enables efficient electroluminescence11 and offers the prospect of electrically pumped stimulated emission.6,12 Transient absorption experiments have shown that optical gain of 50 cm−1 can be achieved through very thin 共⬃100 nm兲 conjugated polymer films.13–15 In this letter, we show that the prototypical conjugated polymer poly关2-methoxy-5-共2⬘-ethylhexyloxy兲p-phenylene vinylene兴 共MEH-PPV兲 can be used to make a high gain solid state optical amplifier, and we explore the dynamics of such a device. A significant technical challenge for developing solid state polymer amplifiers comes in achieving efficient optical coupling to the polymer waveguide. With a view to eventual electrical excitation, the polymer film should optimally be ⬃100 nm thick, which makes it difficult to fabricate good quality end facets. We address this here by using Bragg gratings to surface-couple light in and out of the active waveguide. Figure 1 shows a schematic of one of the polymer optical amplifiers studied in this work. The device consisted a兲
Electronic mail:
[email protected]
of a 100 nm thick layer of MEH-PPV 共American Dye Source 100RE兲 spin coated onto a patterned fused silica substrate. A pair of substrate gratings, separated by a polymer channel of known length, was used to couple signal light pulses into and out of the thin film waveguide. The gratings, of period ⌳ = 440 nm, were defined using electron-beam lithography on poly共methyl methacrylate兲 and then transferred into the substrate using reactive ion etching. To avoid optical feedback and lasing between input and output gratings, the axis of the output grating was rotated by 20°. This also ensured that the amplified output was spatially separated from any surface reflections. The polymer channel was optically pumped with 100 fs pulses at 500 nm 共5 kHz repetition rate兲 to generate population inversion. Weak signal pulses were coupled through the channel and amplified, and the change in their transmission was measured. The time interval between the pump and signal pulses, which were also 100 fs long, was varied with an optical delay line. The gain was calculated from the intensity ratio of the amplified to the unamplified signal energy, after subtraction of the photoluminescence background 共measured with the signal beam blocked兲 using G 共dB兲 = 10 log 关共Pon − PB兲 / Poff兴, where Pon is the amplified signal, Poff is the unamplified signal energy, and PB is the background luminescence with PB ⬍ 0.01 of Pon. With the pump laser off, we initially measured the dependence of signal transmission on waveguide length. We thus determined a coupling efficiency of 17% at each grating and waveguiding losses of 50 cm−1 at 630 nm. The pump energy density was then varied with the signal pulse energy fixed at ⬃2 nJ inside the waveguide, just
FIG. 1. Schematic of the polymer amplifier.
0003-6951/2006/89共20兲/201119/3/$23.00 89, 201119-1 © 2006 American Institute of Physics Downloaded 30 Jan 2007 to 138.251.31.216. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
201119-2
Appl. Phys. Lett. 89, 201119 共2006兲
Amarasinghe et al.
FIG. 2. Gain dependence on pump energy density for a 630 nm signal wavelength for two waveguide lengths 共622 and 1022 m兲. The inset shows variation of gain with signal energy coupled into a waveguide length of 822 m with the pump energy density fixed at 2 J / cm2. The solid line is a fit to gain saturation in a pulsed amplifier.
after the input coupler. Figure 2 shows the optical gain measured at a signal wavelength of 630 nm in two channels of lengths 622 and 1022 m. The gain increases with the increase in the pump energy and saturates at 21 dB in the 1022 m channel, which corresponds to over 100 times amplification of the signal pulse. For the same wavelength, the 622 m waveguide length showed a gain of ⬃18 dB. Amplification in conjugated polymers arises from stimulated emission in a four level system as in organic laser dyes.7 In this case the optical gain is9 G = Nl,
共1兲
where is the gain cross section, N is the excitation density, and l is the waveguide length. The value was calculated using the linear part of the excitation density dependence. The average value for channel lengths of 622, 822, and 1022 m with a signal wavelength of 630 nm is 共4 ± 1兲 ⫻ 10−17 cm2. It agrees with the value reported for a similar polyparaphenylenevinylene polymer.9 The dependence of the gain on input signal energy was measured at a pump density of 2 J / cm2. The inset of Fig. 2 shows that the gain follows the characteristic saturation dependence of pulsed amplifiers9 and that even higher gain may be obtained for signal energy ⬍0.2 nJ. The solid line, in Fig. 2, is the theoretical fit from the expression for a saturated pulsed amplifier,9,16 Eout = Es ln兵1 + exp关共Ein/Es兲 − 1兴exp共G兲其,
FIG. 3. Maximum gain per millimeter 共square solid symbols兲, gain cross section 共circular open symbols兲, absorbance 共dotted line兲, and stimulated emission spectra 共solid line兲. The intensity of the latter was calculated by multiplying the spontaneous emission spectrum by 3.
from the fluorescence spectrum except that the experimental gain value at 615 nm is 9 dB higher. At this wavelength there are finite ground state absorption losses, which decrease with excitation and lead to an overestimation of gain. The gain due to stimulated emission is expected to be ⬃27 dB/ mm at 615 nm consistent with values of gain cross section 共open symbols兲. To explore the gain saturation mechanism we studied the time dynamics. The gain decays faster with increasing pump intensity and the corresponding increase in exciton density 共Fig. 4兲. This can be due to exciton-exciton annihilation, which occurs when the density of excitons reaches a sufficient level where the two excitons within the annihilation radius combine to form a single exciton in a higher state. This leads to a decrease of exciton density and an associated decrease of the optical gain. The gain lifetime at low excitation density was approximately 140 ps. Exciton-exciton annihilation can be modeled using the rate equation15,17,18 dN = − kN − ␥N2 . dt
共3兲
This equation can be solved for the constant ␥ as N共t兲 =
N共0兲exp共− kt兲 , 1 + 共␥/k兲N共0兲关1 − exp共− kt兲兴
共4兲
where N共t兲 is the exciton density at time t, k = 1 / 140 ps is the decay rate with no annihilation, and ␥ is the annihilation rate. A plot of 1 / N共t兲 vs exp共kt兲 gives ␥ ⬇ 共3 ± 1兲 ⫻ 10−9 cm3 / s.
共2兲
where G is the small signal gain coefficient and Es is the saturation signal energy. To explore the scope for broadband amplification we took gain measurements at signal wavelengths of 650 and 615 nm 共Fig. 3兲. The maximum gain at 650 nm was 15 dB in the 1022 m long waveguide. At 615 nm losses were higher due to finite ground state absorption, which precluded the accurate measurement of the nonamplified signal in long waveguides. Instead a shorter waveguide length of 422 m was used. This gave a maximum gain of 16 dB. Gain cross sections, which were calculated from the linear region of the pump dependence, at 615 and 650 nm were very similar to that at 630 nm, all measured with a signal energy of 2 nJ, which is high enough to significantly deplete the available gain. The spectral variation of gain values 共dB/mm, solid symbols兲 follow the stimulated emission spectrum calculated
FIG. 4. Gain vs time delay between pump and signal pulses 共symbols兲 in a 622 m long waveguide. The solid lines show a fit for exciton-exciton annihilation. Downloaded 30 Jan 2007 to 138.251.31.216. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
201119-3
Appl. Phys. Lett. 89, 201119 共2006兲
Amarasinghe et al.
This is similar to reported values in other conjugated polymers.14,15,18 There is a good agreement between calculated and experimental gain dynamics at pump energy densities up to 5 J / cm2 共Fig. 4兲. At higher energy densities, the experimental gain is lower than the calculated one by a constant fraction. We observed a line narrowing of the photoluminescence which is a signature of amplified spontaneous emission 共ASE兲 at ⬎10 J / cm2. The time resolution in the amplifier experiment is limited to ⬃2 ps by beam geometry. In a higher resolution experiment we found that ASE depletes the excited state population in ⬍1 ps 共not shown兲, so this very fast decay would not be resolved in Fig. 4. The generation of charge carriers as a result of exciton-exciton annihilation at high excitation density cannot be excluded too. They absorb in the gain region and reduce the gain. To summarize, we have demonstrated a compact, solid state optical amplifier based on the conjugated polymer MEH-PPV. A maximum gain of 21 dB at 630 nm was observed in a waveguide length of 1022 m. A substantial gain of ⬎13 dB was achieved at 615 and 650 nm giving a gain bandwidth ⬎26 THz. The gain dynamics showed faster decay with increasing excitation density due to exciton-exciton annihilation, while the maximum gain is limited by ASE. The present results are encouraging for further development of MEH-PPV for broadband organic amplifiers for use with polymer optical fibers for low cost, short-haul data transmission. We are grateful to L. O’Faolain and T.F. Krauss for assistance with making the grating couplers, and to the Engi-
neering and Physical Sciences Research Council for financial support. 1
Y. Koike, T. Ishigure, M. Satoh, and E. Nihei, Pure Appl. Opt. 7, 201 共1998兲. 2 N. Tessler, G. J. Denton, and R. H. Friend, Nature 共London兲 382, 695 共1996兲. 3 D. Moses, Appl. Phys. Lett. 60, 3215 共1992兲. 4 I. D. W. Samuel and G. A. Turnbull, Mater. Today 7, 28 共2004兲. 5 U. Scherf, S. Riechel, U. Lemmer, and R. F. Mahrt, Curr. Opin. Solid State Mater. Sci. 5, 143 共2001兲. 6 N. Tessler, Adv. Mater. 共Weinheim, Ger.兲 11, 363 共1999兲. 7 M. D. McGehee and A. J. Heeger, Adv. Mater. 共Weinheim, Ger.兲 12, 1655 共2000兲. 8 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 5937, 28 共2005兲. 9 J. R. Lawrence, G. A. Turnbull, and I. D. W. Samuel, Appl. Phys. Lett. 80, 3036 共2002兲. 10 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. 11 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. Lögdlund, and W. R. Salaneck, Nature 共London兲 397, 121 共1999兲. 12 C. Pflumm, C. Karnutsch, M. Gerken, and U. Lemmer, IEEE J. Quantum Electron. 41, 316 共2005兲. 13 B. J. Schwartz, F. Hide, M. R. Andersson, and A. Heeger, Chem. Phys. Lett. 265, 327 共1997兲. 14 G. J. Denton, N. Tessler, N. T. Harrison, and R. H. Friend, Phys. Rev. Lett. 78, 733 共1997兲. 15 A. Ruseckas, M. Theander, L. Valkunas, M. R. Andersson, O. Inganäs, and V. Sundström, J. Lumin. 76–77, 474 共1998兲. 16 T. L. Koch, L. C. Chiu, and A. Yariv, Opt. Commun. 40, 364 共1982兲. 17 T. Nguyen, I. B. Martini, J. Liu, and B. J. Schwartz, J. Phys. Chem. B 104, 237 共2000兲. 18 R. G. Kepler, V. S. Valencia, S. J. Jacobs, and J. J. McNamara, Synth. Met. 78, 227 共1996兲.
Downloaded 30 Jan 2007 to 138.251.31.216. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
