1 Introduction

Plexcitonic nanostructures as a combination of plasmonic and excitonic buildings offer small mode volumes due to plasmonic arm and strong coupling even at the level of a few emitters at room temperature owed to efficient selection of excitonic ones [1, 2]. The strong coupling in these structures can act as an efficient tool to control the lasing parameters, because of this fact that the spontaneous emission and other optical properties can be influenced due to the adjacency of an emitter in excitonic media near a plasmonic surface [3, 4].

One may know that this control process in lasing needs to highlight matter interaction in gain media, which is satisfied by plasmonic media usage, nevertheless, there is main tradeoff between confinement and losses, which must be solved using hybrid modes instead of plasmonic ones [5]. To find these hybrid modes, there was some reports on this kind of lasers based on surface lattice resonance (SLR), as resonance states in plasmonic lattices of nanoparticles and also nanowires, due to localized surface plasmon resonances and diffracted orders of the periodic structure, to cover the abovementioned tradeoff [6]. In that systems, the strong coupling by excitonic states has been reported [1]. In addition, there were nice works on the random lasing as a kind of lasers in the infrared and also visible region [7,8,9,10,11,12,13,14]. For example, by considering main dark and bright modes of Fano resonances in plasmonic lattices, Hakala and his coworkers reported random lasing in two-dimensional plexcitonic structures [4, 15], also by two-photon absorption (TPA) and two-photon pumped (TPP) phenomena to convert the pump laser light to the higher frequency in metal organic frameworks [16]. Also as it is reported in our earlier paper, one can use core shell nanoparticles as the gain media for random lasing [17].

According to the abovementioned reports, plexcitonic structures can be used as an efficient random laser, but there was an important open question about this fact that can we reduce the input power of these lasers? For this purpose, we want to suggest our previous proposed one-dimensional [18] and two-dimensional [19] plexcitonic structures as low-power plasmonic random lasers and investigate the footprint of plexcitonic states in those structures onto the laser.

2 Plexcitonic structures

Our one-dimensional and two-dimensional plexcitonic structures have been prepared by nanoimprint lithography as a convenient technique of fabricating nanometer-scale patterning with low cost and high resolution. In this method, a thin layer of polydimethylsiloxane (PDMS) mixed with curing agent, which is considered as a polymer in silicon elastomer and characterized by its high flexibility, was used as a substrate placed on a DVD (1D- sample) and camera’s CCD (2D sample) forming the 1D or 2D nano-grating and allowing the deposition of a dye layer of over 30 nm of gold layer on it (the details of these samples’ construction have been reported in [18]).

Coating dyes, as exitonic subsystems, on both 1D and 2D samples, is done by spin coating. By this method, Rhodamin B (RhB) used here can be combined with polyvinylpyrrolidone (PVP), which can act as a polymer for increasing the viscosity.

We produced three samples at the different concentrations of RhB with 1 % concentration of PVP. These solutions, (composition) 0.44 mL of RhB, were dissolved in methanol of concentrations 10–3 mg/ml, 10–4, and 10–5, with 4.4 mg/ml of PVP. In this situation, spin coating take place optimally with 3000 rounds per minute (rpm), 30 s concomitant 20 μL of dye solutions. The schematic diagram of final 1D and 2D sample is shown in Fig. 1. We use conventional field emission scanning electron microscope (FE-SEM) and also cross-section SEM to extract the thickness of dye layer.

Fig. 1
figure 1

Schematic diagram of construction processes a 1D and b 2D plexcitonic structures

Full size image

Finally, to get random lasing, we use second harmonic generation of Nd: YAG laser to pump the samples and collect the lasing by our Avantes spectrometer.

3 Results and discussion

The top surface and also cross-sectional FE-SEM image of the fabricated 1D and 2D plasmonic samples are shown in Fig. 2a–d, respectively. Multi-dye layers onto the one-dimensional grating which is used in this paper is clarified by the image shown in Fig. 2c.

Fig. 2
figure 2

The FE-SEM image of 1D and 2D plasmonic structures top surface (a, b), respectively and FE-SEM cross-section images, (c) 1D two layers' dye sample (d) 2D one-layer dye sample

Full size image

The prepared interface between two dye layers is evident from this figure and also our coated 30 nm gold layer and also 100 micron of dye layers are obvious in these two figures.

3.1 2D plasmonic structure as a gain media

The measured reflection spectrum of our plasmonic structure (without any dye layer) is shown in Fig. 3 for 2D structure. This figure shows the reflectance spectrum of the sample, at different wavelength, for both TM and TE polarizations, specifically at 56°. For that lattice, two SLRs at 539 nm and 596.5 nm for p-polarization case, can be observed in the reflectance spectrum of the cubic array.

Fig. 3
figure 3

Measured reflectance spectrum for a 2D plasmonic sample (without dye) and b 2D plexcitonic sample (with dye)

Full size image

The same is the case for s-polarization, but the second SLR occurred at somewhat different wavelength (603.5 nm). Even in this array, the dip related to the diffracted order occurred at 568 nm and 570 nm for p- and s-polarizations, respectively. The reason for this small peak shift is that the suggested sample is autonomous for p- and s-polarization due to the symmetrical design of plasmonic structure. In addition, there is a main change in the reflectance spectra, and thus the phase shift between two different polarizations, when we add the dye in the plasmonic structure and reach the plexcitonic ones. This fact comes from new plexcitonic modes in the sample, which are covered by the exciton host medium in the vicinity of the plasmonic structure.

As explained above in the introduction part, we want to use plasmonic lattices (nanorods array) as lasing gain media due to this fact, these structures can support dark and bright surface lattice resonance modes. One may know this important fact that when we use of grating structure, the diffractive patterns of metallic nanoparticles (that periodically arranged) leading to hybridization of localized plasmonic mode and the diffractive modes of structure. These collective modes due to interaction of localized surface plasmon and diffraction orders are introduced as SLR.

This is established that the period of the structure is an important factor in determining the response of the array [6]. Position of the SLR is determined by the point of intersection of the real part of the inverse single particle polarizability 1/α and the real part of the array factor S that depends on the distance between particles. The strength and the width of the SLR depend on the difference between the imaginary part of 1/α and the imaginary part of S [20]. Certainly, interferences between two modes in grating structure occur in different wavelength, proportional to the lattice constant. Therefore, we will observe the absorption especially at SLR range and can expect the random laser in a certain wavelength range.

As we discussed above, the exciton–plasmon for both 2D and 1D gold grating covered by dye, can be observed. When the resonant frequency of the surface plamon (SP) is very close to that of the exciton in different systems, such as organic semiconductors (Frenkel excitons) and dye J-aggregates, an interaction between two modes can be noticed. The exchange energy between them is an important factor affecting their interaction. Then the strength of this coupling can be controlled by tuning the parameters related to the plasmonic and excitonic subsystem.

But, we want to investigate this fact that what happens for our lasing gain in this 2D plexcitonic structure due to different kinds of SLR modes such as bright and dark modes. Bright modes are actually standing wave antinode at each particle, induces a large dipole moment and lead to radiation to the far field. While radiation fields result to a standing wave with node for dark modes. Darks modes induce a quadrupole moment (charge distribution) into each particle leading to zero net dipole moment and negligible far-field radiation [4].

Now to investigate the lasing, the sample (gain medium) was pumped with a second harmonic of Nd: YAG nanosecond pulsed laser and the reflected light from the sample was collected into a spectrometer at specified angle under different pump energy.

Below threshold, the emission closely follows the dispersion of the SLR modes; after that at the threshold pump power, we have excited higher energy lasing modes. For the SLR modes (dark and bright), the higher and lower energy lasing mode in fact considered the dark and bright modes, respectively [7].

The lasing performance of the abovementioned 2D samples with different concentrations of RhB 10–3, 10–4, and 10–5 M, at low, 0.13 mJ, and high pumping energy, 5.11 mJ, are shown in Fig. 4a, b, respectively. Sharp-enhanced intensity in these samples by change in the concentrations can be explained by Förster resonance energy transfer (FRET) and surface energy transfer (SET), fluorophores implicated in resonance energy transfer affect the donor and acceptor of emission spectral properties. The enhancement in emission intensity could be achieved by modification of the rate of fluorophore radiative with plasmonic structure. The results indicated that decay rates of radiative and non-radiative are significantly affected by fluorophores’ dipole relative to the plasmonic surface. The increase of pump energy for the gain media could be proportional to the enhancement of the localized electromagnetic field in the vicinity of plasmonic structures; which proportionality may lead to enhancement of the probability of the RhB molecules’ excitation to the higher energy level. The recorded broader absorption and multiple spikes are due to the overlap of localized surface plasmon resonance of Au nanorods with both absorption and emission spectra of the donor–acceptor of the gain medium. In other words, this fact exactly can take place in plexcitonic structure due to plasmon and exciton coupling.

Fig. 4
figure 4

The lasing performance of different concentrations (10–3, 10–4, and 10–5 M) of 2D sample at a low (0.13 mJ) and b high (1.95 mJ) pumping energy

Full size image

Our gain materials in 1D and 2D structure have high two-photon absorption (TPA) coefficients due to choosing RhB and gold nanorods as the gain media. In a TPA excitation phenomenon, two photons are absorbed together by molecules in the gain media. One photon excites the electron to a virtual state with a life time of about 10–18 s and the second photon has to be absorbed within this timeframe to complete the transition into a real excitation state. In fact, the up-conversion can take place in florescence materials in the vicinity of metallic structure.

RhB belongs to Xanthene family in which by this fact that the chemical structure comprises of conjugated bonds and delocalized π electrons, which forms an electron cloud along the molecule; it is liable for high nonlinearity onto intense laser interactions.[21]. It is a known fact that the TPA cross-section depends on the concentration of RhB. As in higher concentrations, due to agglomeration, the nonlinearity of the dye medium decreases and so yield to the lower TPA cross-section [22]. In 1D and 2D structures, with dye concentrations of 10–3 and 10–4 M, the quantum efficiency of TPA decreases and observed that with low concentration of RhB reduces the overall number of TPA, so the strongest coherence blue shift lasing appears for random lasing emission from the sample with 10–5 M concentration. We can be explain the blue shift random lasing by the difference of the re-absorption coefficient at different RhB concentrations.

When we use lower energy pumping (0.13 mJ) onto the sample (with different RhB concentrations) with high dye concentration (10–3 M), there appears blue shift at spikes’ wavelengths from 449.6 to 470.7 nm in blue region and also in the same manner in the green region of this sample from 531.5 to 532.8 nm. After that in the middle state of concentration, 10–4 M, we have the blue wavelength shift range with multi spikes from 451.8 to 468.5 nm and also green range from 531.5 to 532.8 nm can be observed. Lower dye concentration can enhance intensity by coupling of the RhB dye and nanorod dipoles, the radiative rate increased by constructive interference of the dipoles. Now we want to improve the number of spikes in blue spectrum region and also slightly enhance in green shifting region accompanied with lower intensity by the final concentration state set to 10–5 M.

By increase pumping energy to 1.95 mJ, the interaction between the emitted photons of RhB and localized surface plasmon of the gold nanorods is increased. Thus, we have an efficient resonance coupling that is likely to increase the effective scattering cross-section of the gold nanostructure and also increases coupling of the RhB dye and NPs dipoles. For this reason, the random lasing enhances by increasing pumping energy, as we note that at three different dye concentrations, at the first concentrations of 10–3 M, we observed higher spikes’ number and also emission intensity as compared with the lower pump intensity. In this case, the wavelengths of blue shift range changed from 450 to 469.6 nm with green shift range in 531.3 to 533.3 nm. When we use middle concentration, with the same pumping energy 1.95 mJ, the gain per unit length increases, which yields condensed narrow spikes and also slightly broader blue and green regions. Finally, at the dye concentration set to 10–5 M, we get the blue random laser with very nice spikes with width less than 0.5 nm and the wavelengths range at 454 to 468 nm with the green random laser by multi spikes at the wavelength range from 530.8 to 536.3 nm. We choose this concentration as the final state to prevent reabsorption in the dye medium.

Another important parameter to check the lasing status is the change in the intensity as a function of pumping power as shown in Fig. 5, for this best sample by 10–5 M. This graph indicates the logic enhancement in the lasing intensity by enhancing the pumping intensity in blue spectral region.

Fig. 5
figure 5

Recorded intensity as a function of pumping energy from the low to the high energy

Full size image

As shown in this figure, in the blue line region, we have lasing due to the coherent and incoherent feedback mechanisms in the green region. The flat region due to the spontaneous emission takes place below the 0.1 mJ energy per pulse. After that the random effect is stimulated when multiple light scattering becomes significant by enhancing the pump energy to provide the feedback mechanism necessary for optical gain to occur. In the excitation laser energy below 0.8 mJ, we have two-photon absorption inducing photo luminescence. As the excitation laser energy was increased above this point, several spikes evolved due to coherent feedback mechanism.

The next part also explores the characterization of blue–green random laser using 1D nano-grating, which is divided into two parts as different concentrations and thicknesses of the dyes.

3.2 1D plasmonic structure as gain media

As explained before, we want to introduce new 1D grating covered by different thickness and concentrations of dye medium, as the low-cost blue and green random laser systems. In this case, we observed multiple interference between different dye layers with concentration (1 × 10−5 M) as reflectors to satisfy the necessary optical feedback to produce the amplified stimulated emission.

It appears as enhanced gain for the RhB random laser by the gold layer as a first layer onto the 1D grating by their surface plasmon resonance, which yields to the blue random laser shift at wavelengths ranging 446.3–471.2 nm with lower emission intensity and some spikes over nice distinct peaks in this region. In addition, advent of green random laser with some spikes at wavelength range from 531.4 to 533 nm in the vicinity of our source pump (Fig. 6).

Fig. 6
figure 6

The lasing performance of different thickness of RhB onto the 1D sample at low (0.13 mJ) pumping energy in the (left) blue and (right) green region of the (middle) main spectrum

Full size image

Going ahead to the larger thickness of dye layer by repeating the coating process to two layers, the intensity and number of spikes increases in blue–green random laser, because the emission wavelength of the grating can be tuned by changing the thickness of dye, we note such case in which the blue–green random laser range at 451.8–469.3 nm and 530.9–534.2 nm, respectively. Finally, by third layer of dye over this grating, enhancing the number of fine spikes about 0.6 nm and the intensity was repeated in the area of 451.1–467.5 nm and 530.5–536.6 nm.

After we study the characteristics of emission spectrum of blue–green random laser at fixed concentration of RhB dye (10–5) and change in the thickness by deposition multilayers, now we will study the behavior of emitted spectrum using 1D nano-grating with different concentrations of dye (10–3, 10–4, and 10–5 M) as shown in Fig. 7.

Fig. 7
figure 7

The lasing performance of different concentrations of RhB onto the 1D sample at high (1.95 mJ) pumping energy in the (left) blue and (right) green region of the (middle) main spectrum

Full size image

Abovementioned blue–green random lasing was repeated in these samples also, but by the main difference in the green response of the middle concentrations, which reaches to about 10 nm. Higher intensity of blue–green random laser and number of 0.4 nm spikes can be reached with upper gain per unit length at the final dye concentration as shown in this figure.

Also we see with final dye concentration that the range of the green region is smaller than the green region at the middle concentration, because with higher concentration a reabsorption phenomenon in the dye medium is observed, causing an increasing gain per unit length in this region.

4 Conclusions

We have demonstrated a new green–blue plasmonic random laser by 1D and 2D plasmonic structure as gain media deposited with Au nanostructures and different concentrations and thickness of RhB dye. We want to increase the light confinement by more enhances of light scattering in the medium. We observe with 2D plasmonic structure at low energy pumping (0.13 mJ) the output intensity and blue–green range enhances dramatically as the dye concentrations increased. Also more sharp spikes with width less than 0.5 nm appeared and higher random lasing emission intensity can be achieved when increasing pump energy to 1.95 mJ due to the interaction between the emitted photons of RhB and localized surface plasmon of the Au nanorods is increased. These results in the enhanced spectral behavior by the effect of both localized electromagnetic field and scattering, also electromagnetic field localized at the edges of metal nanostructures.

Also to check the effect of thickness on the lasing performance, we use 1D plasmonic structure with different thickness layers and concentrations to enhance gain for the RhB random laser by their surface plasmon resonance, which also yield to the blue–green random laser. With lower energy pumping (0.13 mJ), we observe when increasing thickness of layers, the intensity and the number of spikes increases in blue and green range, because the emission wavelength of the grating can be tuned by changing the thickness of dye.