1. Introduction

Over the past few decades, random lasers have attracted extensive attention due to their interesting physical mechanism and their potential applications in photonics, speckle free imaging and biomedicine [1–7]. Mirrorless random lasers have been realized in many systems, such as semiconductor powders [8,9], dye-doped polymer films [10,11], biological tissues [6,7], quantum dots [12,13], dye-doped liquid crystals (DDLCs) [14,15] and dye-doped polymer dispersed liquid crystals [16,17]. In 2008, Conti and Fratalocchi [18] studied the three-dimensional (3D) photon strong localization and provided the first-principle analysis of random lasers composed by NPs suspended in dye solutions. They reported the relation between the diffusion constant D and the strength of the system disorder. Later, Fratalocchi et al. studied the on-chip lasers [19,20]. In 2020, they reported an on-chip hyperuniform laser for controllable transitions in disordered systems [20]. Their researches provide a new way to realize on-chip lasers that combine the advantages of classical and random lasers into a single platform. Recently, noble metal NPs, such as silver (Ag) NPs and gold (Au) NPs, are often added in random systems to improve the performance of the random laser [21–26]. However, except the price, noble metal NPs have some disadvantages for practical applications. For example, noble metal NPs are not compatible with standard silicon manufacturing process. The optical properties of noble metal NPs can not be tuned easily. Additionally, silver is not stable in the air and it oxidizes quickly when exposed to water or oxygen.

Titanium nitride (TiN), as a new plasmonic material, has attracted extensive attention due to its metallic behavior [27–29]. The plasma frequency of TiN can exist in visible region due to its high carrier concentration, which is close to Au [30]. And the intensity of the localized surface plasmon resonance (LSPR) mode of TiN NPs is comparable with that of Au NPs [31]. TiN, as a plasmonic material, has some advantages over Au and Ag. TiN, as a refractory plasmonic material, can be used in tough operational conditions due to its high melting point, strong hardness and good chemical stability [32,33]. The optical properties of TiN can be manipulated by changing the deposition conditions [34]. Additionally, the TiN material is cheap, abundant, corrosion resistant and CMOS-compatible. Therefore, TiN is a good plasmonic alternative to the Au and Ag for the realization of the high-quality and low-threshold random lasers [35].

In this article, low-threshold random lasers formed by the DCM dye solution with randomly suspended TiN NPs have been studied. Results show that the lasing threshold of the DCM dye solution with TiN NPs is much lower than that of the DCM dye solution with TiO₂ NPs and its reasons are discussed detailed. The performance of the random lasers depends on the cooperative effect of the LSPR of TiN NPs and the scattering strength of random systems. It’s worth pointing that the LSPR of TiN NPs plays a leading role. The lasing threshold can be further decreased by using the NLC as the host material instead of the ethanol solution. Only when the number density of TiN NPs and the scattering strength of the random system are optimum, the random laser threshold is minimum.

2. Sample preparation and experimental setup

The samples used in this experiment mainly include laser dye DCM (from Exciton), ethanol solution and TiN NPs. The TEM image of a TiN NP is shown in Fig. 1(a). The particle size of TiN NPs is about 40 nm in diameter. Firstly, the DCM dye is dissolved in ethanol solvent to form 0.3wt% DCM ethanol solution as the laser gain media. Then, TiN NPs are dispersed in the DCM ethanol solution to form DCM ethanol solution with different final number densities of TiN NPs ($1.162 \times {10^{10}}m{l^{ - 1}}$, $5.813 \times {10^{10}}m{l^{ - 1}}$, $2.915 \times {10^{11}}m{l^{ - 1}}$, $1.468 \times {10^{12}}m{l^{ - 1}}$, $7.296 \times {10^{12}}m{l^{ - 1}}$ and $3.648 \times {10^{13}}m{l^{ - 1}}$). Ultrasonic dispersion process is applied about 20 minutes at room temperature for uniform dispersion of TiN NPs in the mixture. As we immerse glass capillary tubes into the mixture solution, the uniform mixture solution fills into capillary tubes via the capillary effect to form a DCM dye ethanol solution with TiN NPs sample. The length of the capillary tube is about 5 cm. The inner diameter of the capillary tube is about 300 µm.

 

Fig. 1. (a)TEM images of TiN NPs. The scale bar is 50 nm. (b) Schematic diagram of experimental setup.

Download Full Size | PDF

Figure 1(b) displays the schematic diagram of the experimental setup. The sample is pumped by a frequency doubled Q-switched Nd:YAG pulsed laser ($\lambda = 532{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} nm$), with a 10 Hz repetition rate and 8 ns pulse duration. A half wave plate (${\lambda / 2}$) and a polarizer (P) are used to vary the pump energy by rotating the half wave plate. The polarization of the pump beam is parallel to the capillary tube by setting the transmission axis of the polarizer along the capillary tube. A neutral beam splitter (NBS) is used to split the pump beam to two sub-beams with half the energy. One sub-beam is detected by the energy meter (EM). The other sub-beam is focused by a cylindrical lens (CL) to form an excited stripe with about 0.2 mm of width and about 1 cm of length on the side of the capillary tube. A fiber spectrometer (FS) with the spectral resolution of less than 0.13 nm is placed to face the side of capillary tube to collect the emission lasing from the sample.

3. Results and discussion

Figure 2 shows the absorption spectrum (black line) and emission spectrum (blue line) of DCM dye, the LSPR spectrum (red line) of TiN NPs and the pump spectrum (green line). As we can see in Fig. 2, the LSPR spectrum of TiN NPs is very broad, and the LSPR spectrum significantly overlaps with both the absorption spectrum of the DCM dye and the emission spectrum of the DCM dye. In this way, there is a possibility of energy transfer between the TiN NPs and the DCM dye. Thus, the plasmonic enhanced random laser can be fabricated. It is worth noting that the pump wavelength (532 nm) lies within both the wavelength range of the absorption spectrum of DCM dye and the wavelength range of the LSPR spectrum of TiN NPs, and the pump wavelength is near the peak wavelength of the LSPR spectrum. Thus, the TiN NPs can absorb efficiently the pump energy and enhance the energy transfer between the TiN NPs and the DCM dye molecules, then improve the performance of random lasers.

 

Fig. 2. Absorption spectrum (black line) and emission spectrum (blue line) of DCM dye, LSPR spectrum (red line) of TiN NPs and pump spectrum (green line). The blue background indicates the overlap between the LSPR spectrum of TiN NPs and the absorption spectrum of DCM dye. The red background indicates the overlap between the LSPR spectrum of TiN NPs and the emission spectrum of DCM dye.

Download Full Size | PDF

Figure 3(a) shows the emission spectrum from the pure DCM dye solution (black line), the DCM dye solution with TiO₂ NPs (red line) and the DCM dye solution with TiN NPs (blue line), at the same pump energy of 67.2 µJ/pulse. The number densities of TiO₂ NPs and TiN NPs both are $1.468 \times {10^{12}}m{l^{ - 1}}$. As we can see in Fig. 3(a), the emission spectrum of the DCM dye solution is a broad spontaneous emission with the central wavelength of about 626.8 nm. The full width at half maximum (FWHM) of the spontaneous emission spectrum is about 68 nm. Although, compared with the emission spectrum of the DCM dye solution, the emission spectrum of the DCM dye solution with TiO₂ NPs slightly increases in the intensity and narrows slightly in the FWHM (about 46 nm), the emission spectrum of the DCM dye solution with TiO₂ NPs still remains a spontaneous emission. However, the emission spectrum of the DCM dye solution with TiN NPs has an essential change. There are some discrete sharp peaks above the spontaneous emission spectrum at the central wavelength of 630 nm. The intensity of the spectrum is about 5 times larger than that of the DCM dye solution. The FWHM of the peak is about 0.5 nm, which is the spectral characteristic of random lasers. These phenomena discussed above illustrate that TiN NPs are effective to obtain a low threshold random laser.

 

Fig. 3. (a) The emission spectrum from the pure DCM dye solution (black line), DCM dye solution with TiO₂ NPs (red line) and DCM dye solution with TiN NPs (blue line), when the pump energy is 67.2 µJ/pulse. (b) The electric field intensity distribution of single TiO₂ NP (top) and single TiN NP (bottom) simulated by the FDTD method. (c) The schematic illusrarion of the formation process of the random lasers enhanced by TiN NPs. (d) The scattering cross section of the TiN NP (black line) and TiO₂ NP (red line) calculated by using FDTD method. The refractive index of surrounding environment is 1.33.

Download Full Size | PDF

The enhancement effect of TiN NPs on the emission spectrum of the DCM dye solution can be explained by two processes. One is that the electric field around the TiN NPs is localized and enhanced due to LSPR of TiN NPs. Figure 3(b) shows the electric field intensity distributions around the TiO₂ NP (top) and the TiN NP (bottom) simulated by the FDTD method, respectively. Compared with the electric filed intensity around the TiO₂ NP, the electric field intensity around the TiN NP is significantly enhanced, which enhances the absorption efficiency of the pump energy and ultimately strengthens the fluorescence amplification efficiency of the DCM dye molecules near the TiN NP as shown in the inset of Fig. 3(c). The other reason is that the scattering strength of the random system increases due to the presence of TiN NPs. For many TiN NPs, the light will be confined in the random system and interacts with the gain medium before it escapes from the random system due to multiple scattering as shown in Fig. 3(c). The multiple scattering depends on the scattering strength of the NPs. The scattering strength of a particle is evaluated by the scattering cross section ${\sigma _s}$. It can be expressed as ${\sigma _s}(\omega ) = {P_{sca}}(\omega )/{I_{in}}(\omega )$, where ${P_{sca}}(\omega )$ is the total scattering power by the particle and ${I_{in}}(\omega )$ is the incident intensity at frequency $\omega $. The scattering strength of the TiN NP is much larger than that of the TiO₂ NP, especially at the wavelength of 500 nm to 650 nm, as shown in Fig. 3(d). The scattering mean free path ${l_s}$, the average length between two scattering events, can be expressed as ${l_s} = 1/\rho \sigma {}_s$, where $\rho $ is the number density of the particles. The dwelling time of the light generated from the DCM dye increases due to the increase of the scattering strength, and then the gain of the light is increased. When the gain exceeds the loss, the random laser occurs. Therefore, the random lasing from the DCM dye solution with TiN NPs depends on the cooperative effect of the LSPR and the multiple scattering of TiN NPs. When the number density of TiN NPs is $1.468 \times {10^{12}}m{l^{ - 1}}$, the scattering mean free path ${l_s}$ can be approximately estimated about 2.21 cm at the central wavelength of 630 nm, which is much larger than the size of the pump area. This indicates that the scattering strength of the sample is weak. The LSPR of TiN NPs plays a dominant role in the coherent random lasing from the DCM dye solution with TiN NPs.

In order to gain insight into the random lasing action from the DCM dye solution with TiN NPs, the time-resolved photoluminescence (TRPL) measurements are studied using a picosecond laser, as shown in the Fig. 4. The black dots are the data points from the DCM dye solution without TiN NPs. The red dots are the data points from the DCM dye solution with TiN NPs. The solid lines are the corresponding exponential decay fitting. It can be seen from Fig. 4, the luminescence lifetime of the DCM dye solution with TiN NPs is about 4.874 ns, which is much longer than that (0.877 ns) of the DCM dye solution without TiN NPs. The energy resonant coupled transitions between TiN NPs and the DCM dye molecules and the multiple scattering of TiN NPs extend the luminescence lifetime of the DCM dye molecules, which is very beneficial to improve the fluorescence efficiency of the DCM dye molecules and promote the generation of random lasers.

 

Fig. 4. The TRPL measurements from the DCM dye solution without TiN NPs and the DCM dye solution with TiN NPs.

Download Full Size | PDF

In order to study the influence of the number density of TiN NPs on the random lasing from the DCM dye solution with TiN NPs, we studied the emission spectrum of the DCM dye solution with different number densities of TiN NPs as shown in Fig. 5(a, c and e). The insets of Fig. 5(a, c and e) show the emission spectrum of the high intensity area from 620 nm to 660 nm. Figure 5(b, d and f) show the peak intensity and FWHM of the corresponding emission spectrum as a function of the pump energy. As we can see in Fig. 5(a, c and e), when the pump energy is low, the emission spectrum is a single broad spontaneous emission. The intensity of the emission spectrum increases and the FWHM of the emission spectrum decreases with the increase in the pump energy, which is consistent with the results reported in the previous literatures [36,37]. When the pump energy exceeds the threshold, which is given from the slope changing of the peak intensity and the sharp narrowing of the spectrum as the pump energy increases as shown in Fig. 5(b, d and f), many discrete sharp peaks can be observed above the spontaneous emission spectrum. The FWHM of the discrete sharp peaks is less than 0.5 nm. The laser threshold is about 130.2 µJ/pulse, 39.3 µJ/pulse and 58.5 µJ/pulse, when the number density of TiN NPs is $5.813 \times {10^{10}}m{l^{ - 1}}$, $1.468 \times {10^{12}}m{l^{ - 1}}$ and $3.648 \times {10^{13}}m{l^{ - 1}}$, respectively. Figure 5(g) shows the random laser threshold of the DCM dye solution with TiN NPs as a function of the number density of TiN NPs. We can see that the random laser threshold decreases from 180 µJ/pulse to 39.3 µJ/pulse with increasing the number density of TiN NPs from $1.162 \times {10^{10}}m{l^{ - 1}}$ to $1.468 \times {10^{12}}m{l^{ - 1}}$. This can be attributed to two reasons. (I) The scattering mean free path ${l_s}$ decreases with increasing the TiN NPs. The dwelling time of the light in the random system increases with the decrease in the scattering mean free path, and then the gain of the light increases. (II) The number of the DCM molecules in the LSPR area increases with the increase in the number density of TiN NPs. The fluorescence amplification efficiency of the DCM dye is strengthened strongly due to LSPR. When we continue to increase the number density of TiN NPs, the random laser threshold increases. This can be explained by two reasons. (I) Meng et al. [22] has demonstrated that the fluorescence enhancement of the dye molecules depends on the distance between the dye molecules and metallic NPs. When the dye molecules are very close to the metallic NPs, the fluorescence quenching of the dye molecules occurs [22]. The distance between the dye molecules and TiN NPs is very close due to the high number density of TiN NPs. This leads to the occurrence of the fluorescence quenching of the dye molecules. (II) The number of dye molecules in samples decreases because the TiN NPs occupy the positions of the dye molecules. This leads to the decrease in the system gain. As we can see in Fig. 5(g), there is an optimum number density of TiN NPs for the lowest threshold and the optimum number density of TiN NPs is about $1.468 \times {10^{12}}m{l^{ - 1}}$.

 

Fig. 5. (a), (c) and (e) The emission spectrum of the DCM dye solution with different number densities of TiN NPs as a function of the pump energy. The number densities of TiN NPs are (a) $5.813 \times {10^{10}}m{l^{ - 1}}$, (c) $1.468 \times {10^{12}}m{l^{ - 1}}$, (e) $3.648 \times {10^{13}}m{l^{ - 1}}$, respectively. The insets show the emission spectrum of the high intensity area from 620 nm to 660 nm. (b), (d) and (f) The peak intensity and FWHM of the corresponding emission spectrum as a function of the pump energy. (g) The random laser threshold of the DCM dye solution with TiN NPs as a function of the number density of TiN NPs.

Download Full Size | PDF

As comparison, we studied the emission spectrum of the DCM dye solution with TiO₂ NPs as shown in Fig. 6(a) and Fig. 6(c). Figure 6(b) and Fig. 6(d) show the peak intensity and FWHM of the corresponding emission spectrum as a function of the pump energy. The number density of TiO₂ NPs studied in Fig. 6(a) is $1.468 \times {10^{12}}m{l^{ - 1}}$, which is same with the number density of TiN NPs studied in Fig. 5(c). The scattering mean free path of the DCM dye solution with TiO₂ NPs studied in Fig. 6(c) is 2.21 cm, which is same with the scattering mean free path of the DCM dye solution with TiN NPs studied in Fig. 5(c). Although, the random lasing also occurs in the DCM dye solution with TiO₂ NPs, the random laser threshold of the DCM dye solution with TiO₂ NPs is much larger than that of the DCM dye solution with TiN NPs and the random lasing performance of the DCM dye solution with TiO₂ NPs is much weaker than that of the DCM dye solution with TiN NPs. The random laser threshold of the DCM dye solution with TiO₂ NPs, under the number density of TiO₂ NPs of $1.468 \times {10^{12}}m{l^{ - 1}}$, is about 129.2 µJ/pulse as shown in Fig. 6(b), which is about 3.3 times larger than that of the DCM dye solution with TiN NPs. The random laser threshold of the DCM dye solution with TiO₂ NPs, under the scattering mean free path of the system of 2.21 cm, is about 85.2 µJ/pulse as shown in Fig. 6(d), which is about 2.2 times larger than that of the DCM solution with TiN NPs. In addition, we studied the emission spectrum of the DCM dye solution with Ag NPs as a function of the pump energy as shown in Fig. 6(e). The number density of Ag NPs is $1.468 \times {10^{12}}m{l^{ - 1}}$, which is same with the number density of TiN NPs studied in Fig. 5(c). It can be seen from Fig. 6(e), the performances of the random lasing from the DCM dye solution with Ag NPs are similar to that of the DCM dye solution with TiN NPs as shown in Fig. 5(c). Figure 6(f) shows the peak intensity and FWHM of the corresponding emission spectrum as a function of the pump energy. The laser threshold of the DCM dye solution with Ag NPs is about 44.3 µJ/pulse, which is larger than that of the DCM dye solution with TiN NPs. These results indicate that TiN NPs are very effective for decreasing the random laser threshold.

 

Fig. 6. (a) and (c) The emission spectrum of the DCM dye solution with TiO₂ NPs as a function of the pump energy. (a) The number density of TiO₂ NPs is $1.468 \times {10^{12}}m{l^{ - 1}}$. (c) The scattering mean free path of the system is 2.21 cm at the central wavelength of 630 nm. (e) The emission spectrum of the DCM dye solution with Ag NPs as a function of the pump energy. The number density of Ag NPs is $1.468 \times {10^{12}}m{l^{ - 1}}$. (b), (d) and (f) The peak intensity and FWHM of the corresponding emission spectrum as a function of the pump energy.

Download Full Size | PDF

Above studies indicate that the random laser threshold can be decreases with increasing the random system’s scattering strength and improving the fluorescence amplification efficiency of the dye. Therefore, besides increasing the fluorescence amplification efficiency of the dye by choosing the optimum number density of TiN NPs, increasing the scattering strength of random systems is also an effective method for the low-threshold random laser. Because nematic liquid crystals (NLCs) are strong scattering fluids and their refractive index and dielectric tensor can be easily controlled by the external thermal, electric field and optical field, NLCs are widely used in the random laser studies [14–17]. To increase the scattering strength of the system, we use the nematic liquid crystal (P0616A) [35] as the host material instead of the ethanol solution. The other conditions are same with the sample detected in Fig. 5(c). Figure 7(a) shows the emission spectrum of the dye-doped NLC with TiN NPs as a function of the pump energy. Figure 7(b) shows the peak intensity and FWHM of the corresponding emission spectrum as a function of the pump energy. Compared with the emission spectrum of the DCM dye solution with TiN NPs shown in Fig. 5(c), the discrete sharp peaks of the dye-doped NLC with TiN NPs shown in Fig. 7(a) are more obvious. The ratio $R = {{{I_p}} / {{I_s}}}$ can be used to evaluate the quality of random lasers, where ${I_p}$ expresses the peak intensity of the emission spectrum, and ${I_s}$ expresses the intensity of the spontaneous emission as shown in the inset of Fig. 7(a). The ratio R of the dye-doped NLC with TiN NPs can reach up to 3.12 when the pump energy is 13.04 µJ/pulse, which is about 2.7 times larger than the ratio of the DCM dye solution with TiN NPs. The random laser threshold of the dye-doped NLC with TiN NPs is about 5.11 µJ/pulse, which is about 7.7 times lower than that of the DCM dye solution with TiN NPs. In order to gain insight into the observation of the low-threshold random lasing, we calculated the cavity path length of random lasers by performing the power Fourier transform (PFT) analysis for the emission spectra, as shown in Fig. 7(c). The cavity path length of the random laser is proportional to the dwelling time of photons in the gain system. Therefore, the longer cavity path length supports the lower threshold of random laser. The cavity path length of random lasers can be expressed by ${L_c} = {d_m}\pi /mn$[38], where m is the order of the Fourier harmonic, ${d_m}$ is the Fourier components. The first Fourier component is 8.93 µm for the DCM dye-doped NLC with TiN NPs when the pump energy is 13.04 µJ/pulse, as shown in the top of Fig. 7(c). Based on n = 1.72 for NLC, the cavity path length ${L_c}$ of the DCM dye-doped NLC with TiN NPs is about 16.3 µm. The first Fourier component is 3.74 µm for the DCM dye ethanol solution with TiN NPs when the pump energy is 67.2 µJ/pulse, as shown in the bottom of Fig. 7(c). Based on n = 1.33 for ethanol, the cavity path length ${L_c}$ of the DCM dye ethanol solution with TiN NPs is about 8.83 µm. The cavity path length ${L_c}$ of the DCM dye-doped NLC with TiN NPs is much larger than the ${L_c}$ of the DCM dye ethanol solution with TiN NPs, which can explain that the quality of the random lasing from the DCM dye-doped NLC with TiN NPs is better than that from the DCM dye ethanol solution with TiN NPs.

 

Fig. 7. (a) The emission spectrum of the dye-doped NLC with TiN NPs, when the number density of TiN NPs is $1.468 \times {10^{12}}m{l^{ - 1}}$. The inset shows the enlarged view of the emission spectrum when the pump energy is 13.04 µJ/pulse. (b) The peak intensity and FWHM of the corresponding emission spectrum as a function of the pump energy. (c) Top: the power Fourier transform (PFT) of the emission spectrum of the DCM dye-doped NLC with TiN NPs when the pump energy is 13.04 µJ/pulse. Bottom: the power Fourier transform of the emission spectrum of the DCM dye solution with TiN NPs when the pump energy is 67.2 µJ/pulse.

Download Full Size | PDF

4. Conclusion

In conclusion, low-threshold random lasers enhanced by TiN NPs suspended randomly in DCM dye solution are demonstrated. The absorption spectrum and emission spectrum of DCM dye, the LSPR spectrum of TiN NPs, the electric field intensity distribution of TiO₂ NP and TiN NP, the scattering cross section of TiN NP and TiO₂ NP, and the time-resolved photoluminescence measurements are studied to gain insight into the random lasing action from the DCM dye solution with TiN NPs. Results show that the underlying mechanism of the low-threshold random laser is mainly based on the cooperative effect of the LSPR of TiN NPs and the multiple scattering of random systems. The LSPR of TiN NPs strengthens the fluorescence amplification efficiency of the DCM dye molecules near the TiN NPs. The multiple scattering of the random systems extends the dwelling time of light in random systems, which provides more possibilities for the light amplification in the gain medium. In the same situation, the random laser threshold of the DCM dye solution with TiN NPs is much lower than that of the DCM dye solution with TiO₂ NPs. These results indicate that TiN is a good plasmonic alternative to the Au and Ag for the realization of the high-quality and low-threshold random lasers. There is an optimum number density (about $1.468 \times {10^{12}}m{l^{ - 1}}$) of TiN NPs for the lowest-threshold random laser. In addition, the random laser threshold can be further decreased by using the NLC as host material instead of the ethanol solution. The random laser threshold of the dye-doped NLC with TiN NPs is about 7.7 times lower than that of the DCM ethanol solution with TiN NPs under the same experiment conditions. These results are very useful for designing low-threshold random lasers with low-cost and high-quality.