1. Introduction

Random lasers have attracted considerable attention over the last decade due to their interesting fundamentals and potential applications in photonics and bio-medicine [1–6]. Many disordered materials, such as TiO₂ and ZnO powders [2], polymers [7], human tissues [5], dye-doped liquid crystals (DDLCs) [8–13,16], and dye-doped polymer-dispersed LCs (DDPDLCs) [14,15] can be employed to generate random lasing. A coherent random lasing can be obtained by either extended or localized modes via the multi-scattering process with or without a coherent feedback effect. When the time photons remain in the gain medium is sufficient, the amplification of fluorescence can exceed optical loss, such that the random lasing can occur [17].

The diffusion constant of photons in a disordered medium plays a key role in the occurrence of random lasing [4,6–8,14]. For instance, in a DDPDLC system, the diffusion constant strongly depends on the difference between the refractive indices of the LC droplets and the polymer [14,15]. As this difference increases, the diffusion constant decreases, thereby increasing scattering strength, and, in turn, random lasing can be enhanced.

In the above mentioned materials, only those associated with LCs can be used to control lasing characteristics of a random laser because of the externally flexible controllability of LC orientation and, thus, the refractive index of LC. Some thermally and electrically controllable random lasers have been developed using LC-associated materials [8,9,11,13,14]. However, no study has investigated optically controllable random lasers. Thus, this study for the first time investigates an all-optically controllable random laser based on a DDPDLC with nano-sized LC droplets. Experimental results demonstrate that the obtained random lasing intensity can be controlled to decrease with increasing the irradiation time/intensity of one green beam, and increased by increasing the irradiation time of one red beam. This all-optical controllability of random lasing can be attributed to the green-beam-induced isothermal nematic→isotropic (N→I) phase transition and red-beam-induced isothermal isotropic→nematic (I→N) phase transition in LC droplets by trans→cis and cis→trans back isomerizations of azo dyes, respectively. The former (latter) mechanism can decrease (increase) the difference between the refractive indices of the LC droplets and the polymer, thereby increasing (decreasing) the diffusion constant (or transport mean free path), which finally causes the decrease (increase) of scattering strength and, thus, random lasing intensity.

2. Sample preparation and experimental setups

The materials used in this study are 34.47wt% nematic LC (NLC) E7 (n_e = 1.7462 and n_o = 1.5216 at 20°C for λ = 589nm; n_i≅(n_e + 2n_o)/3 = 1.5965 in the isotropic phase) (Merck), 45.24wt% monomer trimethylolpropane triacrylate, 7.88wt% cross-linking monomer N-vinylpyrrollidone, 0.60wt% photo-initiator rose bengal, 0.93wt% coinitiator N-phenylglycine, 9.50wt% surfactant octanoic acid, 0.25wt% azo dye D2 (all from Aldrich), and 1.13wt% laser dye 4-dicyanmethylene-2-methyl-6-(p-dimethylaminostyryl)-4H-pyran (DCM) (Exciton). These materials are uniformly mixed and injected into an empty cell, which is fabricated using two indium-tin-oxide-coated glass slides separated by two plastic spacers, each 6μm thick. After the mixture fills the entire cell via the capillary effect, the cell is then illuminated by a uniform green laser beam from a diode-pumped solid state laser (532nm) with an intensity of 400mW/cm² (irradiated radius of 0.15cm) for 4min to form a DDPDLC. The inset on the bottom of Fig. 1 shows a SEM photograph of the formed DDPDLC, in which the size of LC droplets occupying the black regions, which are randomly dispersed in the polymer matrix (grey regions), are primarily distributed at 30–70nm. These droplets serve as good scatterers for fluorescence photons inside the cell because of the strong mismatch between the refractive indices of the LC droplets and the polymer.

Figure 1 shows the experimental setup for examining the all-optically controllable random lasing emission of the DDPDLC. One pumped laser beam, derived from a Q-switched Nd:YAG second harmonic generation (SHG) pulse laser (wavelength, 532 nm) with a pulse duration of 8ns, repetition rate of 10Hz and pumped energy, E, is focused by a cylindrical lens (focal length, 15cm) on a stripe region of the DDPDLC. The excited stripe is 3mm long and 0.3mm wide. A fiber-optic probe of a fiber-based spectrometer (HR4000, Ocean Optics, resolution: ∼1nm) is placed such that it faces the bottom edge of the cell to record the lateral random lasing output of the DDPDLC. A half-wave plate (λ/2 for 532 nm) and polarizing beam splitter (PBS) are placed in front of the lens for varying incident pulse energy. During the experiments in which the random laser was all-optically controlled, one CW circularly polarized green beam (from a diode-pumped solid-state laser, wavelength: 532nm, output power≤1W) and one CW circularly polarized red beam (from a He-Ne laser, wavelength: 633 nm, output power≤35mW) are installed to, or not to, pre-illuminate the excited cell stripe. The incident angles relative to the cell normal for the green and red beams are both 12°. The pre-measured absorption spectra (not shown) of the DDPDLC cell indicates that it is distributed in the 400–610nm visible region.

 

Fig. 1. Top view of experimental setup for examining the all-optically controllable random lasing emission of the DDPDLC cell (with a 6μm-thickness). The green and red beams are both circularly polarized (λ/2, half waveplate for 532nm; PBS, polarizing beam splitter). The inset on the bottom is an SEM photograph of the DDPDLC cell, in which the length of the white bar is 100nm and the black regions are places occupied by nano-sized LC droplets in the polymer matrix (grey regions).

Download Full Size | PDF

3. Results and discussion

Before conducting the experiments investigating the all-optical controllability of random lasing, the energy threshold of incident pumped pulses for generating random lasing in the cell must be determined. At this stage, both green and red beams are turned off. Figure 2(a) shows the variation in measured fluorescence spectra with a pumped energy of E = 11–22μJ/pulse. Figure 2(b) plots and summarizes experimental data, in which variations in peak intensity of fluorescence output and the corresponding full widths at half-maximum (FWHM) with the pumped energy are presented. Notably, the peak intensity of fluorescence output increases nonlinearly as pumped energy increases. An energy threshold (E_th) ∼16μJ/pulse can be obtained, which is indicative of gain narrowing. The inset in Fig. 2(b) is a photograph of the emission pattern at E = 22μJ/pulse. In Fig. 2(a), several narrow random lasing peaks with ≤1nm-FWHM (the narrowest is Δλ∼0.8nm at E = 22μJ/pulse in the inset) appear at the top of the fluorescence envelopes when pulse energy exceeds E_th. The grey dotted-curve of the fluorescence spectrum of the DDPDLC cell between 580nm and 650nm is also displayed in Fig. 2(a). Those discrete spikes of random lasing emission are distributed around the region of 612–625nm which is near the wavelength (∼600nm) of the maxima of the fluorescence emission. The absorption spectrum of the cell (not shown) indicates that the re-absorption effect may cause suppression effect of the amplification of the fluorescence emission at ≤610nm. This is the possible reason that the strongest lasing spike does not occur at 610nm. To ascertain the underlying mechanism of the random lasing shown in Fig. 2, this work performs the coherent backscattering (CBS) experiment by probing the DDPDLC cell with the use of a weak 633nm laser beam [18]. The measured coherent cone width of backscattering is ∼9mrad. The coherent cone θ is relative to the transport mean free path l ^* (defined as the average distance a photon travels before its direction of propagation is completely randomized) by θ ≅ λ/(2πl ^*) [19–21]. Substituting λ = 633nm and θ = 9mrad into this relation, the transport mean free path is calculated to be approximately 11.2μm. With the satisfactory of the condition l ^*/λ > 1, the random lasing observed in this work (Fig. 2) is attributable to the weak localization of the fluorescence photons via multiple scattering of the LC droplets with coherent feedback in our DDPDLC cell [10,12,15–17].

 

Fig. 2. Variations of (a) measured fluorescence spectra and (b) peak intensity of fluorescence output and corresponding full-width at half-maxima (FWHM) with incident pumped energy. The insets in (b) and (a) are the emission pattern and magnified fluorescence spectrum of the narrow random lasing peaks at E = 22μJ/pulse, respectively.

Download Full Size | PDF

The following experiments investigate all-optical controllability of the random lasing emission with a fixed E = 22μJ/pulse by irradiation via one green and one red beam on the DDPDLC cell. Figure 3(a) (3(b)) show different random lasing emissions generated after the DDPDLC is irradiated by the green beam at different irradiation times t_G = 0–100s (different irradiation intensities of I_G = 0–450mW/cm²) and a fixed irradiation intensity of I_G = 450mW/cm² (a fixed irradiation time of t_G = 2.5min). The intensity of the random lasing signal can be controlled to decrease by increasing t_G (or I_G). Moreover, Fig. 4(a) shows experimental results for the effect of irradiation by the red beam on the random lasing emission via the following three steps. (I) Without illumination by the green and red beams (I_G = I_R = 0), the cell is excited by pumped pulses and a strong random lasing emission is acquired (black curve in Fig. 4(a)). (II) The green beam with I_G = 450mW/cm² is turned on to illuminate the cell for t_G = 2.5min (I_R = 0); the cell is then excited by pumped pulses to generate a weak fluorescence emission, which is not a lasing emission (red curve in Fig. 4(a)). (III) The green beam is turned off (I_G = 0), and the red beam at I_R = 1000mW/cm² is turned on to irradiate the cell. After irradiation by the red beam for t_R = 3, 6, and 9min, the cell is then excited by pumped pulses and thereby generates different random lasing signals (orange, pink, and green curves, respectively, in Fig. 4(a)). The intensity of the obtained random lasing signal can be controlled to increase back by increasing t_R. Figure 4(b) summarizes and plots the variations of normalized peak intensity of the random lasing output with t_G and t_R. Clearly, the DDPDLC random laser has one all-optically controllable feature—the random lasing emission can decay gradually and increase again when irradiated by the green and red beams for an increasing duration, respectively.

 

Fig. 3. Variations of the random lasing emission with (a) irradiation time (fixed at I_G = 450mW/cm²) and (b) irradiation intensity (fixed at t_G = 2.5min) of one circularly polarized green beam. The pumped pulse energy is 22μJ/pulse.

Download Full Size | PDF

 

Fig. 4. (a) Variation of the random lasing signal as irradiation time (t_R) of the red beam (fixed irradiation intensity of I_R = 1000mW/cm²) increases. (b) Variations in normalized peak intensity of random lasing output with t_G and t_R.

Download Full Size | PDF

The all-optical controllability of decreasing (increasing) random lasing intensity is mainly attributable to green (red)-beam-induced isothermal N→I (I→N) phase transition in the LC droplets [22] via trans→cis (cis→trans back) isomerization of the D2 dyes. The details are as follows. All D2 dyes are stable trans-isomers in darkness. The rod-like trans-D2 dyes can be aligned with LCs via the guest-host effect in each LC droplet of cell. In reference to the absorption spectrum of D2 dye reported in our previous study [23], azo dyes can absorb green light (circularly polarized) and typically transform to curve cis-state and then disturb the order of the LCs in each droplet. As t_G or the strength of I_G increases, the concentration of azo dyes converting into cis-isomers increases. This can cause LCs to gradually change from the nematic to the isotropic phase isothermally, and the difference between the refractive indices of the LC droplet and the polymer can markedly decrease from (n_N-n_p) to (n_i-n_p), where n_N and n_i are the refractive indices of an LC droplet in the nematic and isotropic phase, respectively, and n_p is the refractive index of the polymer (≅1.52–1.53) [15]. The diffusion constant of photons (D) in a disorder medium can be represented by D = (vl ^*)/3 [8], where l ^* is the transport mean free path and v the transport velocity of photons in the DDPDLC sample. Decreasing the difference between the refractive indices of the LC droplets and the polymer can increase the diffusion constant (or transport mean free path), thereby decreasing the scattering strength of the photons in the cell. In turn, random lasing intensity can decrease as t_G or I_G increases. As mentioned, most D2 dyes can lie in a cis-state when they absorb green light, which can decay the random lasing emission of the DDPDLC cell. Once the green beam is turned off and the cell is irradiated by the red beam, the cis-D2 dyes can transform rapidly back to the trans-state [23]. An increasing number of dyes can become rod-like trans dyes and the LC droplets will gradually return to the nematic phase as the irradiation time of the red beam increases. The difference in the refractive indices of the LC droplets and the polymer will then gradually increase, causing the diffusion constant to decrease and, thus, scattering strength increases. Therefore, the intensity of random lasing recovers as t_R increases.

To demonstrate that the all-optical controllability of the random lasing emission does not result from the thermal effect-induced phase transition in LC droplets under illumination of the CW laser beam, two separate experiments are conducted. Figures 5(a) and 5(b) present and compare experimental results. The black and red curves in Fig. 5(a) represent the measured random lasing signals at cell temperatures of T = 23 and 80°C, respectively, with I_G = I_R = 0. The random lasing signal decays considerably when cell temperature is increased from 23°C to 80°C, which exceeds the clear point of the E7 LC (∼60°C). This is due to the thermal effect-induced N→I phase transition in LC droplets, causing the decrease in the difference between the refractive indices of the LC droplets and the polymer. This can lead to the decay of scattering strength and thus random lasing intensity. Cell temperature then naturally decreases to 23°C within 4min. The random lasing signal is then measured again (green curve in Fig. 5(a)). The peak intensity of the random lasing signal almost recovers to its original value. This is because the LC droplets all return back to the nematic phase via thermal relaxation. On the other hand, the black and red curves in Fig. 5(b) represent the obtained random lasing signals without and with pre-irradiation of the green beam with I_G = 450mW/cm² for t_G = 2.5min, respectively, at 23°C. The random lasing signal decays significantly when the cell has been pre-irradiated by the green beam. Afterwards, the green beam is turned off (I_G = 0) for 4min, and then the cell is re-excited by pumped pulses. The obtained signal (green curve in Fig. 5(b)) cannot recover to its original state prior to irradiation by the green beam. The discrepancy between experimental results (Figs. 5(a) and 5(b)) implies that the all-optical controllability of the random lasing signal (Figs. 3 and 4) is not induced by the thermal effect, but rather the photoisomerization effect-induced phase transition of LC droplets.

 

Fig. 5. (a) Random lasing signals measured in order at cell temperatures of 23, 80, and 23°C (black, red and green curves, respectively). The time required for the cell temperature to naturally relax from 80 to 23°C is roughly 4min. (b) The random lasing signals measured at 23°C in order as the pre-irradiation intensity of the green beam is set at 0, 450 (for 2.5min) and 0 (for 4min) mW/cm² (black, red and green curves, respectively).

Download Full Size | PDF

4. Conclusion

In summary, an all-optically controllable random laser based on a DDPDLC with nano-sized LC droplets is investigated and reported for the first time. Experimental results reveal that irradiation of one CW green and red beam can all-optically control the lasing intensity of the generated DDPDLC random laser. The causes of the controllability of the random lasing emission are attributable to the green-beam-induced isothermal N→I phase transition via trans→cis isomerization and the red-beam-induced isothermal I→N phase transition via cis→trans back isomerization in LC droplets. The former (latter) may cause the difference between the refractive indices of the LC droplets and the polymer to decrease (increase) and, thus, increase (decrease) the diffusion constant (or transport mean free path) and decrease (increase) scattering strength, which in turn decreases (increases) random lasing intensity. The thermal effect is excluded from possible mechanisms causing controllability of the random laser. Valuable applications for such an all-optically controllable random laser can be found in integrated photonics.