An organic laser based on a monodisperse star-shaped oligofluorene gain medium has been embodied in mechanically flexible format with distributed feedback templated from a holographic master grating. Laser emission was obtained from 425 to 442.5 nm with lowest soft pump threshold at 14.4 μJ/cm2 (2.7 kW/cm2). We compare the performance of such lasers with and without encapsulation. Encapsulation enables stable operation in ambient atmosphere at a 1/e degradation energy dosage of 53 J/cm2.
© 2010 Optical Society of America
In recent years, much effort has been put into the development of solid-state organic light emitters. In particular, the success of organic light emitting diodes combined with advances in polymeric waveguides has been accompanied by increased interest in organic semiconductor lasers. Such devices have a number of benefits compared to their inorganic counterparts, including simplified production methods and the avoidance of highly toxic materials . Another attractive property of organics is their compatibility with a large range of different materials thus allowing for example the usage of mechanically flexible substrates [2–6].
One of the key challenges in the development of these lasers is their susceptibility to photo-induced oxidation which limits their useful lifetime. To date the majority of organic film laser demonstrations required the laser structure to be pumped in vacuum or in a protective inert gas atmosphere . For example the first flexible organic semiconductor laser reported  was characterized in vacuum. In general, only limited information on quantitative analysis of organic laser degradation, both in vacuum and aerobic operating conditions, has been published. One possible approach to address the issue of photo-degradation is to create a multilayer structure where the active medium is enclosed in between layers of a transparent polymer that serves as an oxygen and water barrier. Encouraging results by Richardson et al. [7, 8] on red-emitting MEH-PPV are based on such an encapsulation technique.
A commonly used resonator layout for organic thin-film lasers is based on distributed feedback (DFB), where the optical feedback and out-coupling are provided by Bragg reflections of different order from a grating within the film . DFB gratings are typically custom made by electron beam lithography, two-beam interference or deep-ultraviolet photo-lithography. These methods can be quite complicated and expensive thus counteracting the organic semiconductors’ important potential for of low-cost production. In order to simplify matters soft-lithographic imprint methods to replicate a master grating have been developed [9–11]. Our approach is based on soft lithography using commercially available holographic reflective gratings as a master. Commercially available gratings have been used before by Song, Wenger and Friend  for the fabrication of unencapsulated devices on glass substrates. DFB resonators are particularly suitable for building flexible lasers and a number of reports on such devices have been published. However, all reported flexible devices, with the exception of the LPPP-based work by Scherf et al. [2, 4], are based on dye-doped polymers with thresholds ranging from 16 μ J/cm2  to 800 μ J/cm2 . The lasers presented here are based on an oligomeric organic semiconductor.
Monodisperse star-shaped oligomers, are a very recent addition to the family of organic laser materials. In particular, tris(trifluorene)truxenes (T3), which are basically nanoscale emitters, were shown to have a high photoluminescence quantum yield and to form low-loss film onto a hard substrate when deposited from solution [13, 14]. Consequently, low threshold for optical gain and laser oscillation were reported for such non-flexible devices [14, 15]. However, no data on the operating lifetime was given. Also, to our knowledge, no encapsulation work on monodisperse oligomers has been published yet. In this manuscript, we report mechanically flexible lasers that use encapsulated T3 nanochromophores as the active region and operate at deep-blue wavelengths from 425 to 442 nm. Furthermore, we show that even without encapsulation the truxenes have better photo-stability properties than other star-shaped molecules.
In the next section, the device fabrication is described. We then present the emission characteristics under pulsed ultra-violet (UV) optical pumping. The power transfer function shows a soft threshold which is fitted with a microresonator model and an ASE based theory for the threshold transition. Angle resolved spectroscopy is used to analyze the DFB grating effect. Finally, device operational lifetimes are examined.
An important feature of the lasers presented here is their simple production. The grating structure is copied from the master grating into a UV-transparent polymer that is obtained by photo-curing 1,4-cyclohexyldimethanol divinyl ether (CHDV) with added photo-acid generator (PAG). As PAG we use 4-octyloxy diphenyliodonium hexafluoroantimonite which is dissolved in the CHDV at a weight ratio of 0.2 %. The solution is then mixed in an ultrasonic bath for at least 5 min and degassed. It is in principle possible to incorporate the T3 molecules directly into the CHDV matrix resulting in a nanocomposite that is capable of providing optical gain  though this paper will focus on stacked structures where the undoped CHDV grating is over-coated by a layer of pure T3. A schematic of the structure is shown in Figs. 1(a) and 1b.
After preparation, the PAG/CHDV solution is drop-coated onto the substrate. As a substrate we use commercially available actetate sheets of 0.1 mm thickness made for ink-jet printed transparencies.
The substrate with the film on it is pressed onto the master grating. We use a commercially available holographic reflection grating with 3600 lines/mm, i.e. a period of 278 nm. The grating surface is a polymer overcoated with aluminum. The photoresist is cured by flooding with UV light at an exposure dose of 30.5 J/cm2 through the substrate while in contact with the master grating. Finally, the grating stucture is peeled off from the master. An approximate area of 1 cm2 with high quality grating can be obtained by this method which is sufficient to produce several devices at a time. Analysis by AFM [a scan profile is shown in Figs. 1(c) and 1(d)] reveals that the resulting grating has an average modulation depth of 21 nm in a scanned area of 3.7 × 3.7 μm2. The thickness of the CHDV layer has been measured with a caliper and a value of 30 μm was found. The refractive index of this material is nCHDV = 1.472 (Sigma-Aldrich, Inc.).
The T3 molecules are spin-coated from 20 mg/ml toluene solution onto the grating at a spinning speed of 2700 to 3200 RPM, resulting in films of different thickness (100 nm – 150 nm) which impacts the laser emission wavelength as discussed below in Section 3.2. Further over-coating with optical adhesives provides protection of the chromophores against oxygen and moisture. We present devices that are encapsulated with Norland optical adhesive (NOA) 88 by drop-coating a 1 mm thick top layer of refractive index nNOA = 1.56 (Norland Products, Inc.). The transmission spectrum of acetate is shown in Fig. 2 together with PL and absorption spectra of the chromophores. Detailed analysis of the truxene PL and absorption has been done by Kanibolotsky et al.  and the refractive index of the active material is nT3 = 1.77 .
3. Optical characterization
The organic DFB lasers were optically pumped with pulsed ultraviolet laser light at 355-nm wavelength, 5-ns pulse duration and a 10-Hz repetition rate. The pump source is a frequency tripled Q-switched Nd:YAG laser delivering pump energies ranging from 6 nJ to 3 mJ per pulse. A knife-edge method has been employed to measure the pump beam dimensions. A full width half maximum (FWHM) of 2.9 mm across the beam has been measured at 6 cm and 40 cm distance from the attenuator wheel, i.e. the beam divergence is negligible within the length scales used in the setup. The samples were placed at a 45° angle with respect to the beam axis resulting in an elliptical pump spot of 2.9 mm × 4.1 mm FWHM. Energy densities given in the text and figures are peak values assuming a Gaussian shaped beam cross section. The vertical emission from the sample surface was monitored by a 50-μm-core optical fiber which is connected to a CCD-spectrometer with a maximum spectral resolution of 0.13 nm. The detection angle is controllably variable which enables mapping of the beam profile with an angular resolution of 0.45° (7.8 mrad). During measurements, the samples are exposed to normal ambient air at room temperature. No vacuum or inert gases are used to protect the active material. A schematic of the setup is shown in Fig. 3. Control of the pump energy is achieved by monitoring the pump power with a powermeter and attenuation by a λ/2-waveplate followed by a polarizer and an attenuator wheel. A dichroic mirror is used to filter out 532 nm light that is leaking from the laser aperture. It is possible to focus the pump spot using a 75 mm spherical lens. In this case, an approximate pump area of 0.0014 cm2 has been measured (area with intensity higher than half the peak value) by monitoring the pump spot luminescence with a CCD camera.
Upon appropriate pulsed UV excitation, laser action is observed with a soft threshold behavior that is discussed in the next section. Above threshold, a fan-shaped beam is visible which is the typical mode profile of one dimensional DFB grating lasers.
3.1. Threshold crossing
When pumping close to threshold for optical gain, we observe a gradual steepening of the laser emission intensity from the initial photo-luminescence (PL) slope towards the final laser slope, rather than a well defined sharp threshold. Theoretical work by Stéphan  and Boucher et al. [18, 19] suggests that such a soft threshold can be expected to be fairly typical for microresonator lasers in general and DFB lasers in particular. The origin of this behavior is the significant contribution of ASE and hence unsaturated gain when close to threshold.
It is possible to understand soft threshold behavior as a property of ASE using a variation of the variable stripe length equation that includes saturation and has been discussed before by Pert , Dal Negro et al.  and Costela et al. [22,23]:Equation (1) can be integrated numerically from x = 0, IASE (0) = 0 to x = 1. It is assumed that IASE (1) is directly proportional to the detected intensity. This model has four parameters, Fth, Isat, a and b, that have to be optimized when fitting experimental data. Boucher and Féron  suggested that a three-parameter model may be suitable for microresonators:
For fitting experimental data, the saturation intensity Isat is treated as a linear parameter and solved for by linear regression while the other parameters are optimized using the fminsearch function from Matlab (details on this function can be found in the Matlab documentation ). Both models fit experimental measurements quite well as shown in Fig. 4. In the example given in Fig. 4, the two models agree with each other extremely closely, both in curve shape and in the values for threshold and saturation intensity (Table 1). More generally we observe that both models rarely deviate from each other by more than 10 %. In conclusion the three-parameter model, Equation (2), should be preferred for fitting microresonator thresholds because it has one numerical variable less, needs significantly less computation time and yields similar results as the ASE based model.
The power transfer function of the lasers was extracted from the spectra at different pump fluences. In Fig. 5 we see that a typical laser spectrum consists of a dominant peak and a couple of smaller satellite peaks. The individual laser emission peaks have a spectral width of 0.25 nm or less. If we integrate the intensity over the spectral range of the dominant peak, we get the slope measurement shown in the inset of Fig. 5. For the best sample, a threshold pump fluence of 14.4 μJ/cm2 (2.7 kW/cm2) has been found by fitting model [Eq. (2)] with the spontaneous emission coupling coefficient being κ = 0.13. This value is just below the ASE threshold of a neat truxene film on a hard silica substrate (16 μ J/cm2, 4 kW/cm2) but higher than the lowest truxene DFB laser threshold on a corrugated silica substrate (2.7 μ J/cm2, 270 W/cm2) measured by Tsiminis et al. . Both of these earlier reported values were achieved using protection of the sample by vacuum. The higher threshold in our devices can be explained by slightly lower grating quality and poorer confinement of the transverse optical mode to the active layer because of reduced refractive index contrast in between the layers. An early laser demonstration with a different type of star-shaped oligomers by Xia et al.  revealed an even lower threshold of 0.16 – 0.28 μ J/cm2 (38 – 65 W/cm2) though we will show in Section 3.3 that the pyrene-cored molecules in that report are less stable than truxene-cored oligofluorenes.
A sample encapsulated by drop-coating had a higher threshold of 63 μ J/cm2 (12.2 kW/cm2). The main factors contributing to the increase of threshold are believed to be a lensing effect due to the curvature of the of the encapsulation surface, reduced overlap of the transverse optical mode with the active layer and pump absorption in the optical adhesive.
Precise measurements of the threshold of high-quality devices pumped with the focussed (0.0014 cm2) pump spot were not possible because limits of the setup were reached, especially in terms of detector sensitivity. It is clear however, that the lowest threshold pump energies in this configuration are slightly below 30 nJ (∼ 21 μ J/cm2, 4 kW/cm2). This indicates that for pump spot sizes ranging from 0.0014 cm2 to 0.093 cm2 (1:1.4 elliptical pump spot shape) the threshold pump fluence is independent of the pump spot size.
3.2. Properties of the laser beam
Above threshold, a fan shaped laser beam was observed on a white screen placed parallel to the surface of the sample. Photographic images of the operating sample and the laser beam are shown in Figs. 6(a) and 6(b), respectively. Figure 6(c) shows a photo of the sample under illumination from a UV lamp. The set of graphics in Fig. 7 shows polarisation resolved spectral maps across the beam waist. In these measurements the effect of the grating yields an X-shaped structure. The wavelength of the crossing point of this structure is the position of the stop band . Laser oscillation occurs at that wavelength if the material is able to provide optical gain. The position of the stop-band can be controlled by the spin-coating parameters thus allowing tuning of the laser wavelength between 425 nm and 442 nm, corresponding to an in-plane effective refractive index in between neff = 1.529 and neff = 1.590. In general, an increase of the spinning speed will blue-shift the stop-band because the film thickness and thus the effective refractive index decrease. The effective refractive index can also be estimated from the waveguide materials’ refractive indices given in Section 2 by solving the boundary conditions of the fundamental TE0 at the top and bottom surfaces of the active layer for a given layer thickness. The above values are consistent with the thickness of the T3 layer ranging from 100 nm to 150 nm. Encapsulation with NOA 88 raised the effective index to 1.592 (442.5 nm laser wavelength of the encapsulated device, corresponding to a 110 nm thick T3 layer) which is higher than the index of the cladding material. Optical confinement is therefore still provided in this configuration.
The spectral maps plotted in Fig. 7(a) show that the output has a low angular divergence of 1° (17.5 mrad) and is TE polarised, i.e. the electric field is polarised parallel to the grating. This is consistent with the laser operating in the TE0 mode. In some cases, an unpolarised isotropic ASE background is observed. It is likely that in these cases the ASE originates from regions of poor grating quality that lie within the large area of the unfocussed pump spot. When the pump beam is focussed by a spherical 75-mm lens, we observe that satellite peaks in the spectrum are reduced and the angular beam divergence is improved to (or below) the resolution of our setup of 0.45° (7.8 mrad) as can be seen in Fig. 7(b). This effect is probably due to reduced influence of inhomogenities in the sample structure.
A crucial aspect of organic solid-state light emitters is their susceptibility to photo-induced oxidation which is ultimately limiting the device lifetime. The ASE intensity of an unencapsulated T3 layer that is directly exposed to air follows an exponential decrease with the number of absorbed pump pulses as shown in Fig. 8 for a film of T3 on silica pumped at 0.5 mJ /cm2. During the 1/e lifetime, the sample was exposed to a total pump energy of Fdeg = 7 J/cm2. This is substantially better than the result by Xia et al.  on pyrene-cored oligomers which is Fdeg = 176 mJ/cm2 within the time during which the laser output decreased to half the initial value.
Possible figures of merit for the characterization and comparison of laser operational lifetimes are the already introduced degradation energy dosage Fdeg that the sample has been exposed to during its lifetime and the fraction of this degradation dosage and the threshold pump fluence γ: = Fdeg/Fth. In case of the pyrenes this value is γ= 6.4 × 105 – 1.1 × 106. To our knowledge, only two flexible laser reports include measurements of the lifetime: for Alq3:DCM based lasers Fdeg = 1.6 – 16 J/cm2 and γ= 105 – 106 where found , though it has not been specified how the lifetime was determined, and a recent experiment on Coumarin540-doped NOA88  revealed a degradation energy dosage of only Fdeg = 0.8 J/cm2 at γ= 1.14 × 103, where the lifetime was defined as the time during which the laser output decayed to 10% of its initial value. The best performance for non-flexible devices so far has been reported by Richardson et al.  where the MEH-PPV laser structure was sandwiched in between two glass plates that were glued together by NOA68. Within the lifetime during which the laser output dropped to half the initial intensity, a degradation energy dosage of 1.27 MW/cm2 was achieved with a high threshold of 8 mJ/cm2 and thus γ= 1.6 × 108.
3.3.1. Lifetime of truxene lasers
Unencapsulated lasers had a 1/e degradation dosage of 11.5 ±5.8 J/cm2 (γ= (8.0±4.0) ×105) which is slightly higher than that of T3 on silica. The large variance is possibly related to partial incorporation of the T3 molecules into the CHDV matrix. A sample encapsulated by drop-coated optical adhesive yielded Fdeg = 53 J/cm2 and γ= 8.4 × 105. This indicates that encapsulation by optical adhesives can indeed enhance device lifetime. The lifetime measurements that yielded these results are plotted in Fig. 8 for comparison. The pump fluence was kept constant at 0.5 mJ/cm2 per pulse for all samples, though investigations by Richardson et al. [7,8] on MEH-PPV suggest that the degradation dosage remains the same for various pump levels. Remarkably, there is an initial phase where the laser output is stable and at some point the intensity suddenly starts to drop exponentially. The reason for this behavior is currently under investigation. During the degradation process, a redshift of the laser emission wavelength by about 1.5 nm is observed. A degradation-induced wavelength shift of DFB laser emission has been reported before for MEH-PPV based devices  and was attributed to a change in refractive index of the active layer, though in that case a blue-shift was observed.
For the first time to our knowledge, the promising category of organic lasers based on monodisperse star-shaped oligofluorenes has been assessed in mechanically flexible and encapsulated format. As a further feature of the work, the distributed feedback structures in these lasers were obtained by soft lithographic templating from a commercially available master grating. Encapsulation increases the 1/e degradation energy dosage in ambient from 11.5 J/cm2 to Fdeg = 53 J/cm2. Measured relative to the threshold of the encapsulated device, this latter value correponds to Fdeg/Fth = 8.4 × 105.
The authors thank the Engineering and Physical Sciences Research Council for funding under the grant EP/F05999X/1, HYPIX, Hybrid organic semiconductor/gallium nitride/CMOS smart pixel arrays.
References and links
2. C. Kallinger, M. Hilmer, A. Haugeneder, M. Perner, W. Spirkl, U. Lemmer, J. Feldmann, U. Scherf, K. Müllen, A. Gombert, and V. Wittwer, “A Flexible Conjugated Polymer Laser,” Adv. Mater. 10, 920–923 (1998). [CrossRef]
3. M. Berggren, A. Dodabalapur, R. E. S. A. Timko, and O. Nalamasu, “Organic solid-state lasers with imprinted gratings on plastic substrates,” Appl. Phys. Lett. 72, 410–411 (1998). [CrossRef]
4. S. Riechel, C. Kallinger, U. Lemmer, J. Feldmann, A. Gombert, V. Wittwer, and U. Scherf, “A nearly diffraction limited surface emitting conjugated polymer laser utilizing a two-dimensional photonic band structure,” Appl. Phys. Lett. 77, 2310–2312 (2000). [CrossRef]
5. S. Riechel, U. Lemmer, J. Feldmann, S. Berleb, A. G. Mückl, W. Brütting, A. Gombert, and V. Wittwer, “Very compact tunable solid-state laser utilizing a thin-film organic semiconductor,” Opt. Lett. 26, 593–595 (2001). [CrossRef]
6. M. Lu, B. Cunningham, S.-J. Park, and J. Eden, “Vertically emitting, dye-doped polymer laser in the green (λ ∼ 536 nm) with a second order distributed feedback grating fabricated by replica molding,” Opt. Commun. 281, 3159–3162 (2008). [CrossRef]
7. S. Richardson, “The Fabrication and Lithography of Conjugated Polymer Distributed Feedback Lasers and Development of their Applications,” Ph.D. thesis, University of St. Andrews, School of Physics and Astronomy (2007).
8. S. Richardson, O. P. M. Gaudin, G. A. Turnbull, and I. D. W. Samuel, “Improved operational lifetime of semiconducting polymer lasers by encapsulation,” Appl. Phys. Lett. 91, 261104 (2007). [CrossRef]
9. D. Pisignano, L. Persano, P. Visconti, R. Cingolani, G. Gigli, G. Barbarella, and L. Favaretto, “Oligomer-based organic distributed feedback lasers by room-temperature nanoimprint lithography,” Appl. Phys. Lett. 83, 2545–2547 (2003). [CrossRef]
10. J. R. Lawrence, G. A. Turnbull, and I. D. W. Samuel, “Polymer laser fabricated by a simple micromolding process,” Appl. Phys. Lett. 82, 4023–4025 (2003). [CrossRef]
11. M. Ichikawa, Y. Tanaka, N. Suganuma, T. Koyama, and Y. Taniguchi, “Low-Threshold Photopumped Distributed Feedback Plastic Laser Made by Replica Molding,” Jpn. J. Appl. Phys. 42, 5590–5593 (2003). [CrossRef]
12. M. H. Song, B. Wenger, and R. H. Friend, “Tuning the wavelength of lasing emission in organic semiconducting laser by the orientation of liquid crystalline conjugated polymer,” J. Appl. Phys. 104, 033107 (2008). [CrossRef]
13. A. L. Kanibolotsky, R. Berridge, P. J. Skabara, I. F. Perepichka, D. D. C. Bradley, and M. Koeberg, “Synthesis and Properties of Monodisperse Oligofluorene-Functionalized Truxenes: Highly Fluorescent Star-Shaped Architectures,” J. Am. Chem. Soc. 126, 13695–13702 (2004). [CrossRef] [PubMed]
14. G. Tsiminis, Y. Wang, P. E. Shaw, A. L. Kanibolotsky, I. F. Perepichka, M. D. Dawson, P. J. Skabara, G. A. Turnbull, and I. D. Samuel, “Low-threshold organic laser based on an oligofluorene truxene with low optical losses,” Appl. Phys. Lett. 94, 243304 (2009). [CrossRef]
15. Y. Wang, G. Tsiminis, Y. Yang, A. Ruseckas, A. L. Kanibolotsky, I. F. Perepichka, P. J. Skabara, G. A. Turnbull, and I. D. Samuel, “Broadly tunable deep blue laser based on a star-shaped oligofluorene truxene,” Synth. Met. 160, 1397–1400 (2010). [CrossRef]
16. B. Guilhabert, N. Laurand, J. Herrnsdorf, Y. Chen, A. R. Mackintosh, A. L. Kanibolotsky, E. Gu, P. J. Skabara, R. A. Pethrick, and M. D. Dawson, “Amplified spontaneous emission in free-standing membranes incorporating star-shaped monodisperse π-conjugated truxene oligomers,” J. Opt. 12, 035503 (2010). [CrossRef]
17. G. Stéphan, “Semiclassical study of the laser transition,” Phys. Rev. A 55, 1371–1384 (1997). [CrossRef]
18. Y. Boucher, A. Deryagin, V. Kuchinskii, and G. Sokolovskii, “Near-threshold spectral and modal characteristics of a curved-grating quantum-well distributed-feedback laser,” Nanotechnology 14, 615–618 (2003). [CrossRef]
19. Y. Boucher and P. Féron, “Generalized transfer function: A simple model applied to active single-mode microring resonators,” Opt. Commun. 282, 3940–3947 (2009). [CrossRef]
20. G. Pert, “Output characteristics of amplified-stimulated-emission lasers,” J. Opt. Soc. Am. B 11, 1425–1435 (1994). [CrossRef]
21. L. D. Negro, P. Bettotti, M. Cazzanelli, D. Pacifici, and L. Pavesi, “Applicability conditions and experimental analysis of the variable stripe length method for gain measurements,” Opt. Commun. 229, 337–348 (2004). [CrossRef]
22. A. Costela, O. García, L. Cerdán, I. García-Moreno, and R. Sastre, “Amplified spontaneous emission and optical gain measurements from pyrromethene 567 - doped polymer waveguides and quasi-waveguides,” Opt. Express 16, 7023–7036 (2008). [CrossRef] [PubMed]
23. A. Costela, O. García, L. Cerdán, I. García-Moreno, and R. Sastre, “Amplified spontaneous emission and optical gain measurements from pyrromethene 567 - doped polymer waveguides and quasi-waveguides: erratum,” Opt. Express 16, 7587 (2008). [CrossRef]
24. The Mathworks, Matlab Documentation. Available online at http://www.mathworks.com/.
25. R. Xia, W.-Y. Lai, P. A. Levermore, W. Huang, and D. D. C. Bradley, “Low-Threshold Distributed-Feedback Lasers Based on Pyrene-Cored Starburst Molecules with 1,3,6,8-Attached Oligo(9,9-Dialkylfluorene) Arms,” Adv. Funct. Mater. 19, 2844–2850 (2009). [CrossRef]
26. G. A. Turnbull, P. Andrew, M. J. Jory, W. L. Barnes, and I. D. W. Samuel, “Relationship between photonic band structure and emission characteristics of a polymer distributed feedback laser,” Phys. Rev. B 64, 125122 (2001). [CrossRef]