1. Introduction

The definition of a RL system is not a straightforward issue, however, a RL should satisfy the following two criteria: (1) light is multiply scattered owing to randomness and amplified by stimulated emission, and (2) there exists a threshold, due to the multiple scattering, above which total gain is larger than total loss [1]. Since the pioneering work of Letokhov and associates in 1968 [2], various randomly dispersive hosts such as semiconductor powders [3–5], silver nanopowders [6], polymers [7], liquid crystalline materials [8–12] and even human tissues [13] have been employed for fabricating RLs. The relatively low building costs, sample specific lasing frequency, small size, flexible shape, high emission efficiency and substrate compatibility lead to a plethora of potential technological applications involving these systems.

Light localization and interference effects which survive the multiple scattering events have been invoked to explain random lasing observed in many exotic and complex systems. Optical scattering phenomena (whether week or strong) inside a random medium is thus capable of inducing a phase transition in the photons transport behavior. The optical emission from a RL system often exhibits narrowband spectra composed of ultranarrow modes with subnanometer bandwidth (under certain conditions). Such emission, namely coherent random lasing, was demonstrated to derive from resonant feedback in localized modes [14], two-scatterer cavities [15], or particular nonresonant feedback that generates modes extended over the system [16].

In a random laser discrete modes, amplified by the disordered medium, have distinct resonant wavelengths, while the spectral interval between adjacent modes is usually in the range of a few nanometers. The mélange of light localization and random lasing is especially attractive since each individual random laser source would have a unique emission spectrum given by the specific localized modes in each sample. The lasing phenomena in random systems and also in photonic crystals have promising properties that the conventional laser systems cannot provide [17]. The pioneer research in this area carried out by D. Wiersma, A. Lagendijk, N.M. Lawandy and H. Cao [14, 18–26] has been of great stimulus for the design of efficient and even free shape lasing materials.

In this paper we present a new type of random laser system built from dye doped ionic liquid materials. Ionic liquids (ILs) are presently defined as being organic salts that melt at or below 100° C [27]. Their success arises primarily from the thermo-physical and phase-equilibria properties, while the versatility of their synthesis makes them perfect candidates as “designer materials” [28] for task specific requirements. ILs (and in particular room temperature ionic liquids) have been intensively studied during last two decades and have drawn plenty of attention in numerous scientific areas including organic chemistry [29,30], catalysis [31], electrochemistry [32,33], physical chemistry and engineering [34,35] due to their special physical and chemical characteristics, such as: low vapor pressure, low inflammability, high inherent conductivities, wide range of potential density and viscosity values, liquidity over a wide temperature range, recyclability, high thermal and chemical stability [36].

Due to their tunable and intriguing properties they present themselves as suitable greener alternatives to conventional volatile organic solvents for liquid phase reactions, separation processes, as lubricants and solvents for cleaning or purification [31, 37–39]. Particularly important and promising development routes for the ILs also include their use as electrolytes for lithium batteries, dye-sensitized solar cells, light emitting electrochemical cells, LEECs, colloids and electrochemical processes (as deposition or metal recovery) [40, 41].

Taking advantage of all these interesting features, we demonstrate in this paper, for the first time, the successful use of various types of ionic liquid materials as dielectric hosts for organic dyes in creating highly efficient random laser systems.

2. Experimental section

Here we describe a series of experimental results for selected dyes dissolved in ILs, demonstrating the ability to obtain laser emission (Fig. 1). However we wish to emphasize that the finding is very general and can be extended to almost any IL and dye molecule.

 

Fig. 1 Laser action in ionic liquid systems. Laser emission in various system confinement geometries: (a) wedge cell ; (c) discotic cavity; (e) freely suspended thin film and (g) freely suspended droplet. (b,d,f,h) Emission spectra in case of low pump power (black line, when only fluorescence is obtained) and also for “above threshold” energy conditions (red line, lasing is achieved).

Download Full Size | PDF

We introduce some of the investigated active systems consisting of ILs doped with 0.2-0.5% by weight organic dyes, namely:

Mixture M1: 0.5% wt. Coumarin 540A dye (C540A, 2,3,5,6-1H,4H-Tetrahydro-8-trifluormethylquinolizino-[9,9a,1-gh]coumarin) in Ionic Liquid Methyl-trioctylammonium bis(trifluoromethylsulfonyl)imide), denoted IL1.

Mixture M2: 0.5% wt. Pyrromethene 597 dye (PM597, 4,4-Difluoro-2,6-di-t-butyl-1,3,5,7,8-pentamethyl-4-bora-3a,4a-diaza-s-indacene2,6-Di-t-butyl-1,3,5,7,8-pentamethyl pyrromethene difluoroborate Complex) in Ionic Liquid Trihexyltetradecylphosphonium bis(trifluoromethylsulfonyl)amide, denoted IL2.

Mixture M3: 0.5% wt. DCM dye (4-Dicyanmethylene-2-methyl-6-(p-dimethylaminostyryl)-4H -pyran) in Ionic Liquid 1-Butyl-3-methylimidazolium hexafluorophosphate, denoted IL3.

The low concentration of the organic dyes in the ILs lead to a liquid solution, in which the dye is perfectly dissolved as evidenced by the complete absence of dye aggregation when investigating the samples by means of optical microscopy.

The various samples were optically pumped by means of a pulsed Nd:YAG laser (frequency of 10Hz, pulse duration 8 ns) coupled to a MOPA (Master Oscillator Power Amplifier) system (Spectra Physics 200PRO) that allowed for operating with selective wavelengths (laser line ca. 0.8 nm wide) for the input excitation (in the range 420-680 nm). The optical set-up includes also a series of polarizers, waveplates and lenses for controlling the input intensity, input polarization and for collimating the laser beam onto our sample.

A first sample consisted of a wedge cell containing IL1 and C540A dye (Mixture M1). The wedge cell is made up of two microscope glass plates (0.15mm or 1mm thickness) separated by DuPont Mylar spacers with a thickness of 100 μm at one edge and 1.5 μm at the other one (for avoiding the formation of a Fabry-Perot resonator cavity) and closed at the sides by means of either metallic clips or a glue resin. These wedge-cells were then filled by capillarity our prepared active mixtures. To make sure we eliminate the etalon cavity effect in our wedge cells, we also calculated the free spectral range and this was found to be absolutely incompatible with the wavelength region of the lasing modes and with the spectral spacing of the modes (so the Fabry-Perot etalon effect can be ruled out from our measurements).

Our wedge cell sample filled with M1 mixture shows at low pump energy densities (at excitation wavelength of 460 nm) the typical Coumarin dye fluorescence spectrum (Figs. 1(a) and 1(b)). Above a certain input energy threshold highly intense random laser emission is obtained, in the shape of speckle-like patterns. The phenomenon is clearly noticeable in the actual picture of the experiment (Fig. 1(a)) and the accompanying laser emission spectrum (Fig. 1(b)). We confirm the presence of discrete sharp lasing peaks on top of the residual fluorescence (emission wavelength at 510 nm; each spike has the full width half maximum, FWHM, of about 0.4 nm). Moreover, a decrease in the spectrum width of the emitted light is triggered by the increase of the pump energy. A measure for the optical gain is the so-called narrowing factor (NF), defined as NF = FWHMbelow / FWHMabove, where the FWHMbelow refers to the emitted light below threshold and the FWHMabove to the emission spectrum of the random laser above threshold. In the case of our sample, the narrowing factor is about 250 for each laser mode.

The lasing phenomenon is observed as well for various other types of samples, in different confined geometries and combinations of ILs and fluorescent dyes. Figure 1(c) portrays laser emission obtained in a boundaryless M2 mixture system. The solution is filled in a small hole in a plastic substrate, and stays inside the discotic cavity (due to capillary surface tension forces) that is open at both sides (thickness about 1mm, Fig. 1(c) inset). The emission for both below and above threshold energies are swown in Fig. 1(d) (random lasing peaks are obtained around 582 nm because of the employed PM597 dye).

Figures 1(e) and 1(g) represent random laser emission from special system configurations. The M3 mixture was freely suspended by means of a fluid spreader on a PVC net creating a squared-comb shape of 2 mm x 2 mm and having a thickness of about 450 microns (Fig. 1(e) inset). The profile of the solution menisci driven by competitive forces (gravity, viscosity, and surface tension) across the medium present a slightly thinner central region estimated to be around 300 microns. The associated emission spectra are shown in Fig. 1(f) (lasing at 640 nm related to the emission of the DCM dye). A particular IL droplet lasing system is presented in Fig. 1(g). Mixture M3 is used to create a freely suspended droplet, of 2mm diameter, at the end of a syringe needle. Upon optical pumping, the system lases and a fluctuating aspect random laser pattern is obtained on the background screen. By analyzing the emission spectra we clearly acknowledge that random lasing occurs above a given pump energy threshold value (Fig. 1(h)).

The idea to have laser action in absence of mirrors and boundaries thrusts towards unparalleled tunable and moldable laser source possibilities. By employing a novel system design we make a critical step in demonstrating random laser action in freely suspended dye doped ionic liquid thin films. The mismatching of refractive indices, the irregular shape of the air–liquid interface and the scattering cross sections are some of the parameters accounted for obtaining the presented outcome. Several systems have been studied by considering various dye lasers and ILs and the geometry to suspend the gain complex fluid, with the aim to investigate how the menisci and air–liquid interfaces modify the emission properties. These investigations consented to exclude the influence of any etalon cavity effect for our presented studies.

For higher energies the narrow lasing lines (FWHM 0.4 nm) are always accompanied by other lateral modes, which fluctuate shot by shot, corroborating the idea of an intrinsically stochastic process. The mode-selection mechanisms imply significant interference effects, enhanced by the presence of the gain medium, and related to the random walk of multiple-scattered light waves inside the system. Jiang and Soukoulis modeled a simplistic monodimensional dye-doped random media by the finite-difference time domain method, predicting that localized and extended modes are expected [42]. The former is responsible of random lasing with such a high quality factor of the “cavity”.

The lasing phenomena is confirmed [43] for our systems, by analyzing different parameters, such as the shape and width of the emission spectra under different conditions, input vs. output energy dependence, spatial intensity etc. These studies were performed for all the samples, but herein we present only some of the results. The optical emission behavior depends highly on the input pump energy. At low regimes, one can notice the typical spontaneous emission curve of the Pyrromethene597 dye, indicating that the system does not greatly modify the fluorescence spectrum (Fig. 2(a), black spectrum). When increasing the pump power above a given threshold value (about 3.5 microJ/pulse), discrete narrow peaks emerge from the residual fluorescence spectrum. The linewidth of these sharp peaks was measured to be less than 0.5 nm (Fig. 2(a), red spectrum). When the incident pump energy clearly exceeds the threshold value, the lasing peak intensity increases much more rapidly with the pump power and more sharp peaks appear, owing to the fact that now the balance gain-loss of these lossiest modes becomes positive (Fig. 2(a), blue spectrum). Note that all three spectra (black, red, and blue) were normalized individually for a better comparison representation. In Fig. 2(b) we plot the typical graph for a laser system, that is the so called “Input vs Output” [43]. We can observe two distinct behaviors (and graphic trends). The first left part represents a slow linear increase dependence, characterizing the system being pumped below the lasing threshold (where only the fluorescence contributes to the output intensity value). For a certain input energy the emission intensity drastically changes and starts rapidly increasing. This point is defined as the lasing threshold and for our wedge cell system (that is filled with mixture M1 and also presented in Fig. 1(a)) it has a very low value of about 3.5 microJ/pulse.

 

Fig. 2 Emission properties and Lasing Conditions. (a) Emission spectra at different excitation energies. (1)Typical fluorescence spectrum of the dye obtained for low pump energies. (2) Above a given pump threshold discrete lasing peaks are emerging on top of the residual fluorescence. (3) Multiple laser peak emission is achieved when further increasing the input power. Note that each of the emission spectra (1, 2, 3) was rescaled on the Y Intensity scale for a better representation. (b) Input energy vs output emission intensity curve for the wedge system cell filled with M1 mixture. The lasing threshold is determined at ca. 3.5 microJ/pulse.

Download Full Size | PDF

The energy output for all of the systems in Fig. 1 is obtained in the forward direction (the lasing beam is in the same direction as the pump beam), but is also achieved in the completely back direction, as previously observed [12, 19]. We calculated the efficiency of our systems (as being the output lasing energy divided the input beam energy), and even with part of the incident beam not being absorbed into the gain medium, we still obtain an extremely high value, roughly around 35%. Additionally, we should consider that this already high yield value can be further increased by optimizing various optical and geometrical parameters.

The presented systems are very stable and robust. They exhibit the same emission characteristics after prolonged optical pumping exposure periods and even after months from fabrication date.

In this manuscript we show only some of the experimental results obtained for our systems. We tested many ionic liquids and dyes and we wish to acknowledge that many of the mixtures are working as laser emitters. After we shall perform a through characterization of these systems, a more rigorous systematic interpretation of the results and the potential connections between the main physical parameters (such as emission threshold conditions, lasing linewidth, emission wavelength central position, dye concentration influence, ionic liquid - dye polarity matches etc) will be published elsewhere. However, we can say so far that the same dye behaves differently when placed in different polarity ionic liquids while keeping the same confinement of the system: (a) The lasing emission can be shifted in wavelength, sometimes as much as 40-50nm; (b) The lasing threshold can vary from very low values around microJ/pulse up to tens or even hundreds of microJ/pulse (this implies a change in the amplification efficiency of the mixture dye-ionic liquid); (c) The lasing wavelength spectrum can present only 2-3 principal spikes, or the spectrum can be formed by multiple peaks. The linewidth of each lasing spike usually remains (when above threshold conditions) in the same range (of FWHM ca. 0.5 nm). When considering the same dye - ionic liquid mixture in different confinement geometries we can also notice some important changes in the emission parameters (spectral differences can be observed in Fig. 1(b), 1(d), 1(h) and 1(f)). More investigations are in progress, as well as a model for correlating all these particular behaviors.

In order to get more information about our IL lasing systems we also investigated the far field intensity spatial profile of the mirrorless laser emission, in well above lasing threshold energy conditions, in the case of the wedge cell sample from Fig. 1(a). The overall shape for our lasing systems (above threshold) is close to the pump laser Gaussian profile, as it normally should be (Fig. 3(a)). But still, in accurate detail, one can notice that it is formed by a series of bright tiny spots spatially overlapped which create a richly structured pattern (Fig. 3(b)).

 

Fig. 3 Far field emission intensity profile from a wedge cell system. (a) The common shape in our lasing systems is generally similar to the pump laser Gaussian (in well above threshold conditions). However the intensity distribution (zoomed in (b)) clearly demonstrates the presence of sharp and richly structured intensity peaks of various heights and spatial silhouettes, confirming random laser action. The x,y size of the two images represent the plane spatial dimension of the sensor of the Thorlabs CCD camera beam profiler, that is 8.77 mm x 6.6 mm. Please note that this is not the real excitation spot size on our sample (which was, depending on the employed optical set-up, between 50 and 200 μm), but a far field intensity profile diagram.

Download Full Size | PDF

Additionally, we acquired a sequence of consecutive images for successive pump pulses and we can clearly observe that the optical emission changes in time in terms of both spatial positioning and emission intensity. Figure 4 depicts the measured field emission intensity profile from our freely suspended system in Fig. 1(e). In Figs. 4(a)-4(d) we represent lasing emission for four successive pump pulses, in perfectly stable input energy conditions, for a value corresponding to the lasing threshold situation. As expected, for a characteristic behavior random laser, the spectrum shows slightly different features from shot to shot (the spatial lasing peaks change in optical intensity, spatial distribution and also in number). Since the condition for lasing comes from a sensible balance between gain and loss, it is clear that not for every pump pulse the random walk inside our system is the same and the system does not always gather enough gain to lase efficiently. Sometimes, the gain and loss ratio allows even for the amplification and generation of additional adjacent narrow-banded laser peaks. These measurements represent yet another evidence that the observed phenomena is definitely random lasing. The measurements in Fig. 3 and Fig. 4 were performed by means of a Thorlabs CCD camera beam profiler (1360 x 1024 Pixels, Sensor Area 8.77 mm x 6.6 mm, 6.45 µm x 6.45 µm pixel size) and a Thorlabs Hi-Res Scientific Grade CCD camera (3296 x 2472 pixels, 5.5 µm x 5.5 µm pixel size).

 

Fig. 4 Field spatial laser emission from a freely suspended thin film system. Laser emission has been recorded for four successive pump pulses, a,b,c,d. The input energy is kept stable and the system is slightly above lasing threshold pumping conditions. An interesting random lasing scenario is revealed where, for each input pulse, the emission laser peaks fluctuate in number, spatial placement, distribution, and intensity. The size of the x,y axes is ca. 200μm for each. The region x,y of the four images represent a crop of the same plane space area of the sensor for the Thorlabs Hi-Res Scientific Grade CCD camera in the case of four successive excitation pulses. Please note that this is not the real excitation spot size on our sample (which was, depending on the employed optical set-up, between 50 and 200 μm), but a far field intensity profile representation.

Download Full Size | PDF

The presented random laser action in dye doped IL materials is far from being deciphered and more studies are needed to gain further understandings on this very interesting phenomenon. Some ILs possess liquid crystal properties (lately, lasing effects were also revealed in dye doped nematic liquid crystals [12]), which indicate that these compounds may present large nonlinear optical responses [44] that can enhance the optical gain. Recently, it was also found that the fluorescence intensity is greatly enhanced in ILs as compared to the organic solvents [45, 46]. We advance a hypothesis that a network of cavities is formed due to the salt nature of the solvent and the ability to solvate in different manner the more polar dyes with the smaller and more ionic liquid. On the other hand, the less polar dyes are being “confined” better from less polar ionic liquids (e.g. IL2) and, therefore, the light amplification can be higher and directly related in this micro/nanoconfinement (these results are in preparation and will be presented elsewhere). The nonpolar and polar domains are interconnected and expand into a nanostructurally organized network that can localize light and lead to photon enhancement and amplification. Previous studies also suggested that, in some imidazolium-based ionic liquids, a hydrogen bond resulted in their structural heterogeneity leading to an ordered microscopic network, instead of random intermolecular interaction [47].

3. Conclusion

We have introduced, for the first time to our knowledge, and characterized laser action in a new category of materials, namely dye doped ionic liquids. Owing to their low vapor pressure, low inflammability, liquidity over a wide temperature range, recyclability, elevated solvation capabilities, high thermal and chemical stability ionic liquids are excellent nominees for building, as demonstrated, a series of exotic boundaryless or confined laser systems. Ranging from standard wedge sandwich cell design to freely suspended thin active media films and liquid droplets, these lasers are ultra-compact hazard free narrow banded (FWHM ca. 0.4 nm for each laser mode) emitters that excel due to their low lasing threshold, elevated efficiency, long term durability, easiness in fabrication and vast variety of possible configurations. A detailed analysis of the far field modal profiles for the optical random laser emission reveals a sequence of spatially overlapped lasing peaks, producing an alternating speckled pattern which is temporally fluctuating in position and intensity. Ionic liquids prove to be exceptional hosts for organic dyes and we believe they can be employed as optimal near future replacements of conventional flammable solvents in already available dye laser instruments. These results definitely trace very promising routes for fundamental studies and hold important technological consequences and immediate possible applications in material science, nanotechnologies, optics and photonics.