1. Introduction

In the past decade, solid-state organic lasers have attracted much attention for their application as compact light sources [1,2 ] because of the material advantages of light weight, ease of processing, low cost, flexibility, and spectral range that covers the blue to the far infrared [3]. Recently, lasing oscillations at various wavelengths in the whole visible region have already been demonstrated [3–7 ], indeed this is an attractive feature of solid-state organic laser that cannot be provided by inorganic semiconductor laser. For example, the size of flow cytometry system is quite large, in which a combination of large gas laser and solid-state lasers have been used as a multi-wavelength light source for spectroscopic analysis. As a result, a novel technique for the monolithic integration of multiple laser devices is strongly required [8, 9 ] to achieve a nanoscale analysis system based on the lab-on-a-chip concept [10–12 ], as well as the laser display technology [13].

Multicolor lasing oscillation in a single device could be a promising technique among the existing methods; however it is difficult to realize such fabrication in inorganic semiconductor laser system, because different inorganic laser materials are difficult to be grown on the same substrate. Comparing to organic light-emitting diodes [14, 15 ], multicolor lasing is too hard to be expected from the waveguide cavity with a laser dye blend resulted from the self-adjusting of the lasing wavelength to the wavelength with the maximum gain. Therefore, a parallel arrangement of channel waveguide cavities [16] or a stack of slab waveguide layers with various operation wavelengths could be a feasible device configuration. The multilayer structure of active waveguides is preferable due to the simplicity of the fabrication process, while two major challenges have to be considered [5]: 1) The emission generated in each active layer must be sufficiently confined to induce the lasing oscillation. 2) In each active waveguide, an effective optical feedback cavity must be fabricated with a simple method. Yamashita et al. have demonstrated that it is possible to obtain simultaneous emission at multi-wavelength from a two-dimensional (2D) multilayer structure of distributed feedback grating (DFB) fabricated by photo-nanoimprint lithography (photo-NIL) [17–20 ]. The lasing thresholds for green (G) and red (R) DFB lasers are 175 and 250 KW/cm², respectively. Photo-NIL was employed to avoid the thermal decomposition of the organic dye caused by thermoplastic NIL. However, without the thermal process, the blister in the photoresist cannot be efficiently removed, leading to the defects in the microstructure after the UV exposure. Furthermore, the UV exposure of the photo-NIL required a higher vacuum condition that increased the difficulty of fabrication, while different grating period of master molds were required to match single-mold laser oscillations. Additionally, such structure using poly (methyl methacrylate) (PMMA) with high transmittance as the spacer layer led to the unbalanced output peak intensities.

In this work, we reported on a simple fabrication process of a multi-wavelength amplified spontaneous emission (ASE) with a multilayer structure by vacuum thermal evaporation, and investigated the emission properties of such ASE. This multilayer structure can confine organic laser dye thin films in a planar waveguide, while silver (Ag) was applied as an intermediate modulating layer to confine the ASE in each active layer. Owing to the selectivity of the emission wavelength of organic dyes, ASE emission with the primary colors can be easily obtained. Furthermore, we explored the influence of the metal inserting layer on ASE and discussed the effect of lithium fluoride (LiF) as spacer layer. Two planar waveguides of G and R laser dyes were then integrated together, while each laser layer exhibited simultaneous operation under optical pumping. The thresholds of the R-G planar waveguides ASE were measured to be ~37.2 and ~39.7 KW/cm² for the emission at 503 nm and 662 nm.

2. Experimental section

Figure 1(a) shows the multilayer planar waveguides structure. Firstly, the tris(8-hydroxyquinoli-ne)aluminum (Alq₃) doped with 4-(dicyanomethylene)-2-t-but-yl-6(1,1,7,7-tetramethyljulolidyl-9-enyl)-4H-pyran (DCJTB) was used as the active layer fabricated for the R-planar waveguide. The thickness and the doping concentration of the active layer were ~200 nm and ~1.0%, respectively. A metal modulating layer of Ag was then deposited onto the R-active layer, while LiF was deposited on the both sides of the Ag film as the spacer layer in order to avoid the absorption to the pump energy caused by the metal layer. The N,N’-(4,4’-(1E,1’E)-2,2’-(1,4-phenylene)bis(ethene-2,1-diyl)bis(4,1-phenyl-ene))-bis(2-ethyl-6-methyl-N-phenylaniline) (BUBD-1) was then applied as the active layer of the G-planar waveguide and was stacked in the same manner as for the R-planar waveguide. The thickness of the BUBD-1 layer was 100 nm.

 

Fig. 1 (a) Schematic of G-R planar waveguides ASE. (b) Molecular structures of organic materials used in this work.

Download Full Size | PDF

The organic films were deposited onto clean glass substrates by thermal evaporation under a high vacuum (~10⁻⁶ mbar) to investigate the photoluminescence (PL) spectrum and ASE characteristics. The glass substrate was firstly degreased in an ultrasonic bath by the following sequence in detergent, deionized water, acetone and isopropanol. The deposition rates of organic material, LiF and Ag were 1.0, 1.0 and 0.2 Å/s, respectively.

The device was optically pumped from the sample surface, as it is shown in Fig. 1(b). The light source for the optical pumping was a Nd:YAG laser (FTSS 355-50, CryLaS) at an excitation wavelength of 355 nm, with a pulse width of about 1 ns and a repetition rate of 100 Hz. The beam profile of the pumping pulse was a rectangular stripe with dimension of 2.5 mm × 10 mm. A cylindrical lens and neutral density filters were used to adjust the excitation intensities. The emission lights were collected from the edge of the film into an optical fiber connected to a spectrometer. The PL spectra were measured by using FLSP 920 spectrometer series. The transmission properties and surface morphologies of the films were characterized with UV–visible spectra (U-3900H, HITACHI) and atomic force microscope (AFM, Nanonavi SPA-400 SPM). All the measurements were performed in the ambient atmosphere.

3. Results and discussion

3.1 The characteristics of active layers

Figure 2(a) shows the PL and ASE characteristics of the G and R active layers. As shown in Figs. 2(b) and 2(c), the peaks of ASE for the BUBD-1 and Alq₃: DCJTB are at 506 nm and 669 nm, respectively. The ASE thresholds of the G and R active layers were estimated to be at ~6.8 and ~4.9 KW/cm², which are relatively lower comparing to the recent reports in Ref [21, 22 ]. (100 KW/cm² for the green fluorescent complex and 66.7 KW/cm² for the red fluorescent film).

 

Fig. 2 (a) Normalized absorption spectrum of LiF, PL and ASE spectra of BUBD-1, PL spectrum of DCJTB and ASE spectra of Alq₃: DCJTB. (b) and (c) are the dependences of the output intensity and the full width at half maximum of the emission spectra on the pump intensity for the thin film of BUBD-1 and Alq₃: DCJTB.

Download Full Size | PDF

3.2 The choice of inserting metallic layer

In order to gain a high gain multi-wavelength ASE, it is essential to use an intermediate modulating layer to confine the ASE in one active layer. There are three requirements for the candidate materials: firstly, the absorptivity of this material should be as low as possible to ensure that the inserting layer has little effect on the threshold of laser dyes. Secondly, the transmission of the pump source through the inserting layer must be high enough to conserve sufficient energy to pump the next planar waveguide ASE. Finally, its refractive index must be lower than the organic film to form the asymmetric planar waveguide as glass/organic film/modulating layer, making the generation of ASE in this structure.

It is well known that the metallic layer has the ability to make the whole multilayer system having a high transmittance in the visible region [23]. Among the common metal such as gold (Au), Ag and aluminum (Al), Ag shows the lowest absorptivity (< 5%) in the visible region, while those of Au and Al are 8% and 30%, respectively [24]. The high absorptivity will impair the transmittance of whole multilayer system. Additionally, Ag has low refractive index (n = 0.35) which can form asymmetric slab optical waveguide with the glass substrate (n = 1.46). As a result, 10 nm Ag layer was used as the modulating layer in this work.

3.3 Effect of LiF spacer layer

Figure 3 shows the transmittance of 10 nm Ag layer for the pump energy, which is up to 78.9% at 355 nm, such layer is too thin to confine the pump energy in one planar waveguide. Additionally, the fluorescence quenching near to the interface of the active layer and the Ag layer should be considered. All of the issues resulted in no ASE in the planar waveguide under the optical pumping.

 

Fig. 3 The transmittance spectra of films for Ag(10 nm), LiF(150 nm)/Ag(10 nm), LiF(300 nm)/Ag(10 nm) and LiF(300 nm)/Ag(10 nm)/ LiF(300 nm).

Download Full Size | PDF

To overcome these problems, LiF was inserted as the spacer layer between the Ag and the organic layer. As it is shown in Fig. 3, the film of LiF can hardly absorb the light of the pump source at the wavelength of 355 nm, as well as the ASE of both G and R dyes. On the other hand, LiF has a relatively low refraction index n (~1.39), so it has the potential to confine the ASE. Bulović et al. found that the surface plasmons can decrease exponentially with distance from the electrode beyond 80 nm [25], so spacer layer with thicknesses of 150 and 300 nm were used in this work.

Figure 4 shows the emission spectra for the G-planar waveguide with (a) 150 nm and (b) 300 nm spacer layer fabricated by vacuum thermal evaporation under different pumping intensities, while ASE were clearly observed in all films. The inset plotted the relationship between the pump intensity and the peak intensity of spectra, as well as the full width at half maximum (FWHM). For the both G-planar waveguides, the ASE emission were obtained at 503 and 507 nm with thresholds at ~8.3 and ~4.6 KW/cm², indicating that the devices with a 300 nm spacer layer had a better performance.

 

Fig. 4 Optically pumped emission spectra of the structure: glass/BUBD-1(100 nm)/(a) LiF (150 nm); (b) LiF(300 nm)/Ag(10 nm); glass/Alq₃:1%DCJTB(200 nm)/(c) LiF (150nm); (d) LiF(300 nm)/Ag(10nm) at different pumping intensities. Insets are the dependences of the output intensity and the FWHM of the emission spectra on the pump intensity for the corresponding devices.

Download Full Size | PDF

Similarly, the ASE of R-planar waveguide with 150 and 300 nm LiF have been investigated, as it is shown in Figs. 4(c) and 4(d). The 644 nm and 664 nm ASE were obtained for the R-planar waveguide with the thresholds at ~13.9 and ~7.8 KW/cm², respectively. The results exhibit red shift of 20 nm with respect to the peak of ASE which can be attributed to the weak microcavity effect as the thickness of the whole planar waveguide increases [22].

In order to get a deeper insight into the reason why both G- and R-planar waveguides have better performance with 300 nm-thick LiF as the spacer layer, two aspects will be discussed. Firstly, the transmittance of LiF(150 nm)/Ag(10 nm) and LiF(300 nm)/Ag(10 nm) films for the pump energy at 355 nm are 73.6% and 59.2%, as it is shown in the Fig. 3. The transmittance of the former is too high to confine enough pump energy to stimulate the active layer, resulting in higher threshold. On the other hand, the threshold is related to the surface morphology of the LiF/Ag layers. Figure 5 shows the AFM measurements of 150 and 300 nm-thick LiF films grown on 10 nm Ag, as well as the corresponding line profiles. The molecular aggregation was observed on both surfaces, while the root mean square roughness of 150 and 300 nm-thick LiF films is 9.7 and 28.5 nm, respectively. The measurement of line profile implies that the surface of LiF(150 nm)/Ag(10 nm) film is irregular which could result in diffuse reflection of the emission light, leading to high thresholds. Conversely, the line profile of the LiF(300 nm)/Ag(10 nm) shows more regular bulges with larger height that can enhance the light adsorption due to light trapping effect [26]. Such surface contributes to reserve light in the waveguide and to decrease the loss of pumping source.

 

Fig. 5 AFM images of (a) LiF(150 nm)/Ag(10 nm); (b) LiF(300 nm)/Ag(10 nm) on the glass substrates and the corresponding line profiles.

Download Full Size | PDF

3.4 Simultaneous multi-wavelength ASE

The multi-wavelength structure can be obtained by integrating the G- and R-planar waveguides together. In the integrated structure, the LiF spacer layer was deposited at the both side of Ag modulating layer to prevent the fluorescence quenching. The transmittance of LiF(300 nm)/Ag(10 nm)/LiF(300 nm) for light of 355 nm is 47.9%, indicating that the inserting layers split the pump energy in half for the two planar multilayer waveguides devices. The different deposition positions of the single color planar waveguide have been demonstrated in Fig. 6 . The ASE thresholds of the G-R planar waveguide were measured to be ~42.6 and ~42.4 KW/cm² for the emission at 503 and 662 nm, respectively. The ASE thresholds of the R-G planar waveguide were ~37.2 and ~39.7 KW/cm² for the green and red ASE, respectively. The inset of Fig. 6(b) is a photo of the R-G device under optical pumping in which bright R and G emissions are clearly visible. It is found that the thresholds of the different multi-wavelength structures are slightly different. However, the imparity of the peak intensity of the G and R ASE light for the two structures was obvious. Comparing the different sequence of stacked structure, the R-G planar waveguides showed a more balanced ASE than that of the G-R planar waveguides.

 

Fig. 6 Emission spectra of (a) G-R, (b) R-G planar waveguides ASE under different optical pumping intensities. The inset shows output emission intensities of corresponding stacked structure at various pumping intensities and the photo of the stacked device.

Download Full Size | PDF

In the stacked structure, the active layer that deposited later is subjected to the effect from Ag layer in both sides, which has a stronger impact on the ASE peak intensity than the first deposited active layer. Therefore, the single color planar waveguide should be arranged in the order from the high threshold to low threshold to balance the emission intensity of both ASE.

4. Conclusions

In summary, we have reported multi-wavelength ASE with a multilayer structure of active planar waveguide fabricated by vacuum thermal evaporation. 10 nm Ag film was used as the intermediate modulating layer to confine the ASE in one active layer. 300 nm LiF was then inserted between the laser dye and metal layer to avoid fluorescence quenching and confine the pump energy in one waveguide, which was found to be the most important factor for the threshold of ASE. In the sample with G-R planar waveguides, ASE at 503 and 662 nm, corresponding to the ASE wavelength in the respective layers, were simultaneously observed. The thresholds of the R-G planar waveguides were measured to be ~37.2 and ~39.7 KW/cm² for the green and red ASE, respectively. The fabrication scheme proposed in this work is expected to have the potential to achieve a compact multicolor laser source using in organic photonic technologies.