1. Introduction

Two critical research issues in organic optoelectronics are to reduce the cost of organic LEDs and to improve their external efficiency. One approach to improve the extraction efficiency is to use two-dimensional (2D) photonic crystals (PhCs) [1-3]. A PhC structure enhances the light emitted from the active layer by slowing the propagation speed of the photons, thus increasing the coupling to the out-of-plane radiative modes. Moreover, the fabrication of these PhCs in an organic layer can be done by nanoimprint lithography (NIL), which is a simple and cost efficient process [4]. Another approach is to increase the spontaneous recombination rate of the emitters. This can be based on the energy transfer between light emitters and surface plasmons (SPs). Several results on enhanced light emission via SPs have been described [5, 6]. It has been shown that plasmon resonances of thin metal layers can be chosen to enhance the efficiency of semiconductor light sources [7-10]. K. Okamoto et al. demonstrated a significant enhancement of the internal quantum efficiency of semiconductor quantum wells by controlling the energy transfer between quantum wells and surface plasmons generated in a thin metal layer [11-12]. Recently, comparable studies have been performed with dye-doped organic films [13] and with conjugated polymer films [14, 15] in close proximity to metal surfaces. Similar coupling processes have been observed in both organic and inorganic structures with an important enhancement in the spontaneous emission intensity of emitters. In this paper, we report on the combination of the two approaches mentioned above to enhance the light-emission efficiency of organic thin films. An active polymer film deposited on a metal surface is patterned by NIL and the SP energy is matched to that of the emitter in the PhC, reaching up to a × 27 enhancement

2. Experiment

A stamp with different lattice constant PhCs was fabricated in a silicon wafer by using electron-beam lithography and dry etching. The electron-beam exposure was carried out on a Jeol 6000 equipment with a dose of 130 μC/cm² under a beam current of 100 pA on single layer of a ZEP 520 resist (positive tone resist from Zeon Corporation). Development is carried out during 30 sec in a solution of ZED N50 (Zeon Corporation). The silicon stamp is then etched to a depth of 350 nm by inductively coupled plasma etching and treated with a self-assembled anti-adhesive monolayer (tridecafluoro-1, 1, 2, 2-tetrahydrooctyl trichlorosilane deposited in vapour phase).

 

Fig. 1. (a). Scanning electron micrograph of a nanoimprinted two-dimensional PhC with a 380 nm lattice constant honeycomb array of holes (holes depth 350 nm), (b). cross-section schematic of the studied system.

Download Full Size | PDF

A thermal NIL process is used to replicate these 2D periodic patterns in a dye-doped polymer. The dye-doped polymer is composed of rhodamine 6G (from Sigma Aldrich) directly dissolved with a concentration of 5×10^-4 mol/L in a printable polymer (mr-NIL 6000 from micro resist technology), which is optically transparent in the visible range. The photoluminescence (PL) spectrum of the obtained solution is dominated by a strong and narrow, excitation independent, emission at 550 nm with a FWHM of about 35 nm. A 400 nm thick layer of this modified polymer is spun on a quartz wafer and on metal-coated quartz wafers and baked at 60 °C for 10 min before the NIL process. The stamp and the coated substrates are pressed together in a 2.5 inch Obducat nanoimprinter at 60 bar for 5 min at 90 °C. The pressure is sustained during the cooling phase until the temperature fell below 35 °C. As shown in Fig. 1(a), the patterns of the stamp are faithfully replicated in the modified polymer layer. Figure 1(b) presents the cross-section schematic of the studied system. Samples are excited with the 514.5 nm line of an Argon laser with a power of 5 μW, the incident beam is normal to the surface. The PL is collected through a ×20 microscope objective and dispersed in a spectrometer with 0.1 nm spectral resolution. The PL data is recorded at room temperature. The metallic substrates used have 50 nm thick layers of gold, aluminium and silver deposited by thermal evaporation on quartz substrates. The metal films were deposited using NFC 2000 Temescal 6 kW electron beam guns with a deposition rate of 10 Angstroms per second. The control of the deposition rate allows the tuning of the surface plasmon frequency of the film throughout the visible. This method is usually used to provide substrates for surface enhanced Raman spectroscopy [16].

3. Results and discussion

To test the impact of the NIL process on the optical properties of the active layer, the PL spectrum of the spin-coated dye-doped polymer on a quartz substrate has been compared to that of the same substrate imprinted with a flat stamp. A decrease of less than 3% in the PL intensity is found after patterning. This effect is due to the temperature sensitivity of rhodamine 6G during the prebaking and imprinting process, while the imprinting pressure is not found to be a critical parameter [17]. The use of a printable polymer (mr-NIL 6000) with a glass transition temperature of about 40°C, which is lower than that of, e.g., PMMA, is an advantage to avoid thermal degradation.

 

Fig. 2. Photoluminescence spectra of a nanoimprinted unpatterned dye-doped polymer film on a quartz substrate (black line), on a 50 nm Al quartz substrate (green line), on a 50 nm Ag quartz substrate (blue line), on a 50 nm Au quartz substrate (red line).

Download Full Size | PDF

A comparison of the emission spectra of the imprinted unpatterned dye-doped polymer on a quartz substrate without metal and on three quartz substrates coated with 50 nm of Al, Au and Ag, respectively, is presented in Fig. 2. The unpatterned sample refers to a resist film imprinted under the same conditions but with a flat stamp. To determine the plasmon resonance frequencies of the different substrates, normalized extinction spectra were measured and are presented in Fig. 3(a) for silver and gold. For aluminium, there is no plasmon resonance and the metal layer acts only as a mirror and reflects the pump light and the emitted light. In this case, a light extraction enhancement of about 2.5 is observed. A larger enhancement is observed for the Au and Ag films (Fig. 2). The surface plasmon resonances are found at 650 nm and 558 nm for Au and Ag, respectively. When the plasmon wavelength in the metallic film does not exactly match the emission wavelength of the dye, as in the case of Au, only a limited increase in the luminescence intensity is observed (Fig. 2). On the other hand if the plasmon resonance wavelength matches the emission wavelength of the dyes in the case of Ag, the excited electrons in the dye molecules can couple efficiently to the free charges at the metal surface. In this case, a 9.3 fold enhancement in the PL peak intensity is observed. It has been shown that the surface plasmon resonance wavelength can be tuned by controlling the Ag islands size [16, 18]. The Ag substrate surface was characterized using an atomic force microscope (Nanoscope IIIA, Veeco) in the tapping mode to determine the film morphology as presented Fig. 3(b). The area scanned in each image is 5 × 5 μm². The second advantage in using silver islands films apart from the tunability of the SP resonance wavelength is that the non-negligible surface roughness scatters the SP modes to radiated light [19].

 

Fig. 3. (a). Normalized extinction spectra of the different used substrates, presenting the surface plasmon wavelength tunability. (b). right upper image: AFM images (5 × 5 μm²) of a 50 nm thick Ag evaporated on quartz substrate, (black inset: the depth profile along the white line).

Download Full Size | PDF

It is known that emitters can couple to the SP mode only via their near-field, because SP modes are non-propagating evanescent waves. The typical interaction length of the SP mode is around a few dozens of nanometers for the silver films [13]. In order to further increase the light extraction of the 400 nm polymer films, they were structured by NIL with the PhCs stamp. The height of the stamp pillars is 350 nm. The residual layer can be estimated to be around 50 nm, which is the typical interaction length of the SP mode. PL spectra of a nanoimprinted unpatterned sample and of three different honeycomb lattice photonic crystals imprinted on a quartz substrate are shown in Fig. 4(a). Different enhancement factors for different lattice constant have been measured. The maximum enhancement is achieved with a lattice constant of 700 nm which corresponds to the normalized frequency of 1.27 since the emission wavelength centre is at 550 nm. This frequency region is situated well above the light line of the PhC band structure, so that all the modes are leaky and can couple to free space [4]. The ratio between PL intensities integrated from 520 to 650 nm of the patterned structure and the unpatterned imprinted surface has been calculated giving a maximum enhancement factor of 2.8 at room temperature for the sample with 700 nm lattice constant. The 700 nm lattice constant PhCs stamp was used to imprint the modified polymer spun on the Ag coated quartz substrate. A 27-fold enhancement in the peak PL intensity is observed in this sample [Fig. 4(b)]. A 200 nm layer of PMMA was spun between the Ag coating substrate and the active polymer layer. The enhancement due to surface plasmons was dramatically reduced; the surface plasmon penetration depth does not penetrate in the active layer.

 

Fig. 4. (a). Photoluminescence spectra of a nanoimprinted unpatterned dye-doped polymer film on a quartz substrate (black line), of a 2D photonic crystal with a 380 nm lattice constant (blue line), with a 500 nm lattice (green line) and with a 700 nm lattice (red line), (b). photoluminescence spectra of a flat surface imprinted on a quartz substrate (black line), of a 2D photonic crystal with a 700 nm lattice constant imprinted on a 50 nm Ag quartz substrate (blue line), of a 2D photonic crystal with a 700 nm lattice constant imprinted on a quartz substrate (red line), of a nanoimprinted unpatterned dye-doped polymer film on a 50 nm Ag quartz substrate (green line).

Download Full Size | PDF

4. Conclusion

In conclusion, a 27-fold enhancement of PL intensity at room temperature is achieved in a 2D PhC containing dye in the vicinity of Ag surface plasmons compared to an unpatterned sample on a glass substrate. This enhancement is attributed to the light coupling to the leaky modes of the PhC slab structure and to the coupling of the emitted photons to surface plasmons. Our results indicate that nanoimprint lithography is a well suited process to fabricate these challenging photonic structures and that the combination of surface plasmons and nanoimprinted structures in an active layer can lead to a new class of cost effective and high efficiency OLEDs. Furthermore, the metallic surface could be used as an electrical contact.