1. Introduction

Since the theoretical prediction of inhibited spontaneous emission by a photonic crystal band gap by Yablonovitch [1], a photonic crystal has become a major topic in optical materials and has been extensively studied from various viewpoints. One of the most attractive points is the remarkable increase of a photonic density of state (DOS) at the edges of a photonic band gap (PBG), which can be applied to light amplification such as laser action. Cholesteric liquid crystals (CLCs) spontaneously form a periodical helical structure with a pitch of visible light wavelength, and its structure is regarded as a one-dimensional photonic crystal. Hence, a distributed feedback (DFB) type laser action was observed in dye-doped CLCs [2], opening extensive research field. Among various investigations on CLC lasing, lowering threshold is one of the urgent issues for practical applications, particularly for an ultimate goal toward current injection or continuous wave (cw) lasing. Amemiya et al. used reflection media for excitation and emission lights, and succeeded in decreasing threshold [3,4], and Zhou et al. also reported the similar results [5,6]. An attractive method in view of excitation was proposed by Belyakov; i.e., extremely long optical interaction length due to high photonic density of state is expected by exciting at a higher energy side of PBG, leading to lower threshold [7]. This idea was actually proved by Matsuhisa et al. [8] Lasing utilizing a defect-mode also provides one of the solutions, and was studied by many researchers using various techniques [9–12]. Moreover, Chee et al. succeeded in lowering threshold intensity using high-birefringent CLC cells with broader photonic band gap widths [13], and Blinov et al. proposed a new mode called the quasi-in-plane leaky mode, and indicated that the threshold of the quasi-in-plane leaky mode has 1/5 as small as that of the DFB mode [14]. Lee et al. discussed the threshold property depending on the incident angle of excitation light [15]. Naturally, laser dyes with high quantum efficiency are also studied to decrease the threshold [16,17].

Some years ago we constructed new cell structures consisting of a dye-doped low-molecular-weight CLC (low-CLC) layer sandwiched between polymer CLC (PCLC) films with opposite helical handedness to that of the low-CLC, and succeeded in realizing the decrease of the lasing threshold [18]. Following this study, we studied the lasing property in another three-layered structure with the same PCLC helical handedness to that of a low-CLC interlayer, and briefly discussed the difference of lasing characteristics between two 3-layered cells [19]. Recently we extended the fabrication technique to multilayers and succeeded in preparing a dye-doped low-CLC sandwiched by multilayered PCLC films. In this paper, we report the lasing property in the multilayered CLC structures. Lasing characteristics of a dye-doped low-CLC without and with sandwich structures by single and double PCLC layers were compared. The comparison between experimental lasing characteristics and calculated DOS was also made.

2. Experimental

The PCLCs used are mixtures of two aromatic polyester polymers (Nippon Oil Corporation), one of which has 25% chiral units in its chemical composition. We adjusted the mixing ratio of the two polymers in order to obtain an optical pitch of 500 nm. First, a 10 wt% chloroform solution of this polymer mixture was spin-coated onto 2-mm-thick quartz substrates with rubbed polyimide (AL1254, JSR). Then, the spin-coated PCLC films (first PCLC layer) were cured for 30 min at 180 °C and quenched to room temperature. Using this procedure, about 2-μm-thick well-aligned glassy PCLC films were obtained. In order to make multilayered systems, an aqueous solution of polyvinyl alcohol (PVA) was spin-coated on the PCLC-coated substrates. The thickness of PVA was about 0.5 μm. After rubbing the PVA surface, we spin-coated another chloroform solution of a PCLC mixture whose chiral sense is opposite to that of the first PCLC layer. By curing the substrate under the same condition as the first PCLC layer, 2-layered PCLC films were fabricated on the quartz substrates. PVA acts as not only an alignment layer for PCLC but also a blocking layer of the solvent (chloroform) for the second PCLC layer.

A low-CLC was a mixture of a host nemaitc LC (ZLI 2293, Merck) and a chiral dopant (MLC 6248, Merck) to form a left-handed helix. The optical pitch was ca. 480 nm. The PCLC has a larger anisotropy, Δn = 0.22 (n _e = 1.78 and n _o = 1.56), than that of the low-CLC, Δn = 0.14 (n _e = 1.63 and n _o = 1.49). Hence the photonic band of the low-CLC is within the photonic band of the PCLC. In the present experiments we adjusted the low energy PBG edge of the low-CLC to coincide to the middle of PBG of the PCLC. The low-CLC was doped with 3 wt% quarterthiophene as an emission dye, which has an emission peak at about 500 nm. To compare the efficiency of lasing and its threshold, we prepared five kinds of cell structures: dye-doped low-CLC cells (normal-cell), dye-doped low-CLC cells having PCLC films with the same handedness on both sides (LLL-cell), dye-doped low-CLC cells having PCLC films with the opposite handedness on both sides (RLR-cell), and dye-doped low-CLC cells having 2-layered PCLC films on both sides (LRLRL-cell and RLLLR-cell). Each cell structure is depicted in the inset of Figs. 1(a) –1(e). The thickness of the dye-doped low-CLC layer was about 16 μm.

 

Fig. 1 The emission spectra of (a) normal-cell, (b) LLL-cell, (c) RLR-cell, (d) LRLRL-cell, and (e) RLLLR-cell. Calculated (c/n)DOS of (f) normal-cell, (g) LLL-cell, (h) RLR-cell, (i) LRLRL-cell, and (j) RLLLR-cell. The refractive indices and helical pitch of the PCLC used for the simulation are n _e = 1.78, n _o = 1.56, and p = 299 nm, and those of the low molecular CLC are n _e = 1.63, n _o = 1.49, and p = 304 nm, respectively. The thicknesses of the PCLC and the low-CLC are 1.8 and 16.0 μm, respectively. Incident light is linearly polarized normal to the rubbing direction. Left-handed circular polarized was used for DOS calculation. The resolutions of wavelength (Δλ) were 0.001 nm (red broken line) and 0.0001nm (black bold line). Because of the extremely large amounts of data, the calculation was only made around the peak position (480~510 nm) in case of Δλ = 0.0001 nm. Peak positions are almost the same for both cases, but peak intensities for Δλ = 0.0001 nm is larger than those for Δλ = 0.001 nm. Green dotted line in (a) is the transmission spectrum of the active (low-CLC) layer.

Download Full Size | PDF

In the fluorescence and lasing emission measurements, we used a pulsed laser beam with a wavelength of 410 nm from an optical parametric oscillator (Surelite OPO, HOYA Continuum) pumped by third-harmonic light generated from a Nd:yttrium aluminum garnet (Nd:YAG) laser (Surelite II, HOYA Continuum). Linearly polarized (perpendicular to the rubbing direction) pumping beam was focused on the sample center. The beam diameter at the focus plane is estimated to be ca. 3.8 μm. The emission was collected using a lens to a multichannel spectrometer (HR-4000, Ocean Optics, Inc.) along the direction parallel to the helical axis, as illustrated in our previous paper [4]. The resolution of the spectrometer was 0.11 nm.

3. Results and discussion

Figure 1 shows lasing emission spectra [Figs. 1(a)–1(e)] in each cell together with corresponding DOS spectra [Figs. 1(f)–1(j)] for left-handed circular polarized light calculated based on Berreman’s 4 × 4 matrix method [20–22]. In all the spectra, the lasing emission peaks were observed. Multiple peaks were observed in all the lasing spectra except for a normal-cell and the spectral shapes did not change with increasing the excitation intensity. For the normal cell and LLL-cell, the highest emission peak appeared at the lower energy edge of photonic band gap of low-CLC [see Figs. 1(a) and 1(b)], being mainly attributed to DFB lasing. On the other hand, another three-layered cell, RLR-cell, gives multimode lasing, as shown in Fig. 1(c). The low-CLC layer acts as a defect layer, so that many DOS peaks corresponding to defect modes can be observed within the PBG of PCLC (460 ~530 nm).

Since the lasing peaks qualitatively correspond to DOS spectrum peaks, the lasing is attributed to defect mode lasing. Lasing spectra for five-layered cells, LRLRL- and RLLLR-cells, were more complicated; i.e., lasing emission appeared as multiple peaks within the wavelength range of 15 nm inside of the PBG of low-CLC [Figs. 1(d) and 1(e)]. Correspondence between lasing peaks and DOS peaks [Figs. 1(i) and 1(j)] is not good. The former is located in slightly higher energy side compared with the latter. Because of complicated structures, it is difficult to replicate the experimental result by calculation. This may be the reason of the lack of correspondence.

Figure 2 shows the log-log plot of the maximum emission peak intensity as a function of excitation energy in each cell. Threshold intensities for lasing determined from Fig. 2 and the full width of half maximum (FWHM) of the lasing peaks in each cell are listed in Table 1 . The emission intensity and the FWHM are determined from the highest peak and its profile among the multiple peaks. The order of the threshold intensity is normal >> LLL > RLR > RLLLR ≥ LRLRL. FWHMs also show the consistent order; i.e., FWHM decreases in this order. The threshold of the LRLRL-cell is very low; i.e., almost one order of magnitude smaller compared with that of the normal-cell. In our previous papers, we have already reported the lasing properties of the RLR-cell [18] and the LLL-cell [19]. However the direct comparison cannot be made because of different dyes and cell conditions used. According to Fig. 2, the threshold intensity and FWHM of the RLR-cell are clearly smaller than those of the LLL-cell, indicating that the RLR-cell has higher Q value than the LLL-cell. This fact can be also found in laser beam patterns. Figure 3 displays the far-field patterns of the lasing emission from (a) RLR- and (b) LLL-cells. The RLR-cell shows a clear diffraction fringe pattern, which proves that the emission light from the RLR-cell has very high coherence [23]. Moreover, the slope efficiency in the RLR-cell seems to be larger than that in the LLL-cell.

 

Fig. 2 Log-log plot of emission intensity in each cell as a function of excitation energy. Thickness of a low-CLC was 16 μm, and the wavelength of the excitation energy was 410 nm.

Download Full Size | PDF

Table 1. Threshold intensity and FWHM of laser emission peaks of each cell.

View Table

 

Fig. 3 Photographs of the lasing beam patterns obtained from (a) RLR-cell and (b) LLL-cell.

Download Full Size | PDF

Now we discuss the lasing characteristics in each cell. In the LLL-cell, left-circularly polarized (L-CP) laser light is generated in the middle left-handed low-CLC (L-low-CLC) active layer. The lasing wavelength is within the band gap of L-PCLC, and moreover the PCLC is not thick enough to act as leaky cavity mirrors for the low CLC. This is why the lower threshold is achieved in an LLL-cell compared with that in a normal cell. As for the RLR-cell, L-low-CLC acts as a defect layer for R-PCLC. Actually, many defect states are seen within the whole band gap of PCLC in the DOS spectrum of the RLR-cell, while no defect states exist in the LLL-cell in this range. Together with the experimental fact that lasing occurs at wavelengths corresponding to the defect state within the band gap of low-CLC, we can conclude that the lasing from the RLR-cell is due to a defect mode. This is the essential difference between the LLL- and RLR-cells and explains the different threshold and cavity behavior. In general, a defect mode can realize high Q cavity [24]. Furthermore, RLR-cell shows DOS peaks even for right circularly polarized (R-CP) light (not shown) as high as those for L-CP light, whereas LLL-cell shows only negligible DOS peaks for R-CP light. These are the reason why the RLR-cell gives lower threshold, smaller FWHM, and higher coherence than the LLL-cell.

The threshold energy of the LRLRL-cell is much lower than that of the normal-cell by about a factor of 10. The FWHMs of the LRLRL- and RLLLR-cells are on the average 0.24 nm and 0.26 nm, respectively, and are smaller than those in the normal-cell (0.47 nm) and 3-layered cells (0.32~0.39 nm). Since PCLC films in the 5-layered cells reflect both the left- and right-handed circularly polarized lights of emission, a more effective feedback will occur and the optical gain will increase, compared with the normal- and 3-layered cells (RLR and LLL). The RLLLR-cell has slightly higher threshold energy than the LRLRL-cell, while the calculated DOS of the RLLLR- and LRLRL-cells are almost the same [see Figs. 1(i) and 1(j)]. In DOS calculations, the maximum values are very sensitive to the resolution of wavelength [25] and cell conditions such as helical pitch and the thicknesses of layers [18,19]. In addition DOS also depends on light polarization state used for calculation. Moreover, the polarization state of lasing emission is not simple except for normal-cells, since there is no distinct eigen mode in such complicated structures. Hence it is almost impossible to obtain DOS responsible for lasing emission from each cell exactly. Therefore, we cannot comment on the most advantageous structures for the threshold property from the viewpoint of DOS.

4. Conclusions

In conclusion, we developed a novel hybrid structure composed of a dye-doped low-molecular-weight cholesteric liquid crystal sandwiched by multi-layered polymer cholesteric liquid crystal films. In the 5-layered cell, the defect-mode lasing was observed with an extremely reduced threshold by a factor of 10 compared with that in a simple dye-doped low-CLC cell, 12 nJ/pulse. The threshold behaviors between low-CLC, 3-layered and 5-layered cells were discussed based on the calculated density of state (DOS).