We demonstrate Au-nanoparticle-assisted random lasing from a powdered GaN sample. In the presence of Au nanoparticles on GaN powder surfaces, several lasing lines are observed in photoexcited luminescence spectra near the center of the GaN band-edge emission peak. The random lasing is considered to arise from a decrease in the lasing threshold due to the suppression of crystal defect loss by surface plasmon excitation on Au. From spatially resolved lasing emission spectra and their FT analysis results, the formation of random lasing cavities at different spatial positions is confirmed. The size of the random lasing spot is determined to be larger than that of the scattered light speckle of the pumping source on a thin powdered GaN sample.
© 2011 Optical Society of America
Random lasers can be used as a new light source without any precise external laser cavities. Such random lasers have been extensively studied by a number of groups because of their unique characteristics, such as spatial emission and a relatively high spontaneous emission factor. Random lasing action is caused by multiple scattering of lights in a gain medium consisting of highly concentrated and randomly shaped particles or column scatterers . Since the coherent random lasing action in ZnO was reported by Cao et al. , such a study has been performed on various semiconductor materials, such as SnO2 , GaN , GaAs , and ZnSe .
For improvement of the random lasing properties of semiconductor materials, it is necessary to suppress radiative recombination loss. One origin of this loss is the presence of surface or internal defects in semiconductors. In particular, the surface defects are easily introduced during grinding bulk semiconductor materials for preparing powdered samples to achieve random lasing. These defects lead to a decrease in quantum efficiency of the emitters. Recently, it has been reported that the emission efficiencies from nanorods are markedly improved by depositing metal nanoparticles on these nanorods [7–9]. This improvement is considered to be due to the effect of surface plasmons on the metal nanoparticles. It is natural to consider that the same strategy can be applicable to powdered semiconductors and may lead to the realization of a random laser. It is interesting to develop a random laser using a coupled system of semiconductor gain materials and metal nanoparticles. Note that the surface plasmon accompanies strong localized fields and that the light scattering cross section in a coupled system of metal nanoparticles and gain materials is strongly enhanced [10, 11]. Concerning the surface-plasmon-enhanced random laser, most studies use dye molecules as a gain material (see, e.g., Refs.  and ), but only one study uses ZnO as a gain material .
In this work, we demonstrate a random laser using powdered GaN combined with Au nanoparticles. Several lasing lines are observed from a powdered GaN/Au nanoparticle system near the center of the GaN band-edge emission peak by excitation above the threshold (≥2 MW/cm2). Note that the spatially resolved lasing emission profile yields useful information on random lasing modes [13–15]. Therefore, we measure such a profile in the present random laser system. The obtained result suggests that random lasing cavities are located at different spatial positions and that the size of the random lasing spot is larger than that of the scattered light speckle of the excitation source.
2. Experimental procedure
Commercially available GaN powder (Furuuchi Chemical Co., Ltd) with a mean diameter of 300 nm was dispersed in methanol. To obtain a very thin powdered GaN film, this methanol solution was dropped on a silicon substrate and then rotated at 5000 rpm on a spin coater. Figure 1(a) shows the scanning electron microscope image of the sample surface. As shown in Fig. 1(a), such a thin GaN film consists of aggregates of GaN particles. These aggregates are well separated with each other and inter-distances between them are found to be in the range of 10–100 μm. In Fig. 1(b), we show the scanning electron microscopy image of a GaN powder aggregate. The size of the aggregate is several tens of micrometers. A Au layer with a mass thickness of several nanometers was deposited on the powdered GaN film by vacuum evaporation. The Au layer may cover only one side of the GaN aggregates. Thus, the deposited Au layer is regarded as island-type nanoparticles and may act as a light absorber in the red wavelength range, owing to the excitation of localized surface plasmons . The Au nanoparticle size was controlled by the deposited film thickness monitored using quartz microbalance system.
Photoluminescence measurements were performed by excitation with a continuous light of 325 nm from a He-Cd laser (Kimmon IK3302R-E). For lasing experiments, a frequency tripled light pulse of 355 nm from a Nd:YAG laser with 5 ns pulse duration was used. The excitation area was about 300 μm in diameter. The excited light was incident at an angle of 45° to the sample surface. Emission spectra were measured using a single monochromator equipped with a charge-coupled device (CCD, Princeton Instruments, PIXIS:100B). Spatially resolved lasing emission spectra were measured using the same CCD in the imaging mode, where a spatial resolution of about 6 μm was obtained from the vertical dimension (2 mm) of the CCD chip. Emission from the sample was collected using two double convex lenses, 74 and 220 mm in focal length and 50 mm in diameter.
3. Results and discussion
Figure 1(c) shows the photoluminescence spectra of powdered GaN with and without the deposition of Au nanoparticles measured by excitation with a continuous He-Cd laser. The broad emission band with a peak at 650 nm observed for the sample without Au deposition corresponds to the defect-related emission band previously reported . A small peak is also observed at around 390 nm (∼3.2 eV). This peak corresponds to the band-edge emission in GaN. The deposition of a Au layer with a mass thickness of 2 nm decreases the broad emission band intensity and simultaneously greatly enhances the band-edge emission at 390 nm.
The same enhancement as that shown in Fig. 1(c) is observed in ZnO nanowires and nanorods deposited with Au or Ag2O nanoparticles [7–9]. The two independent mechanisms of this type of photoluminescence enhancement are proposed [7, 9]. One mechanism is the resonance absorption of defect emission by the excitation of surface plasmons on metal nanoparticles and subsequent charge back-transfer process from metal to ZnO . The other mechanism is the local electric-field-induced enhancement of the band-edge emission due to the surface plasmon excitation on metal nanoparticles . The resonance energy of the surface plasmons should be the defect emission peak energy in the former case and the band-edge emission energy in the latter case. In the present study, the embedded Au nanoparticles are found to show an absorption peak at around 660 nm. This wavelength exactly coincides with the defect emission peak observed in the powdered GaN. Therefore, at least for our sample, the former mechanism is concluded to be dominant. Indeed, such an enhancement was not observed when the deposited Au layer thicknesses were larger or smaller than 2 nm shown in Fig. 1(c) (dashed lines).
Figure 2 shows the evolution of emission spectra from powdered GaN with Au nanoparticles excited by a pulsed Nd:YAG laser. The spectra are plotted after integration over 50 excitation pulses. At low excitation powers, only a weak emission band is observed at 390 nm (3.2 eV). At excitation powers above the threshold (≥ 2 MW/cm2), sharp emission peaks appear near the center of the broad spontaneous emission band. With a further increase in excitation power, the emission intensity increases and becomes slightly broader. In the inset of Fig. 2, we show the lasing output intensity as a function of excitation power density. We observed no lasing spectrum at lower excitation power densities. At above lasing threshold powder density (≥ ∼2 MW/cm2), the output intensity steeply increased with increasing excitation power density. Because the GaN aggregates are well separated with each other [see Fig. 1(a)], single aggregate may act as a random-mode cavity, which is formed due to refractive index inhomogeneity . Thus, our observed sharp peak in Fig. 2 may arise from random lasing due to the resonant feedback in each cavity.
It should be noted that such a lasing emission never occurs in powdered GaN without Au even at the highest excitation power of our pulsed laser. In the case of GaN without Au nanoparticles, almost all the excited electrons may be trapped at the defect sites and recombine with the emitted light in the defect emission peak region (∼660 nm). This leads to a decrease in band-edge emission intensity. Such loss of the band-edge emission inhibits random lasing. On the other hand, in the case of GaN with Au the loss can be compensated by the surface plasmon absorption of the defect-emitted light and the subsequent charge back-transfer process from Au nanoparticles to powdered GaN. Therefore, the lasing emission is attained via the surface-plasmon-related processes above a threshold excitation power.
The lasing emission lines observed in the present study are broader than those in other semiconductor random lasers [2, 4, 6]. The typical full-width at half maximum (FWHM) value of ZnSe random laser is 0.4 nm , while that observed in the present study is 2.5 nm. The large broadening is caused by the fact that each spectrum in Fig. 2 was obtained after integration over 50 Nd:YAG laser excitation pulses. In Fig. 3(a), we show the pulse-to-pulse spectra excited sufficiently above the threshold power (∼20 MW/cm2) at a fixed excitation spot. Several narrow spikes with an FWHM value of ∼0.3 nm are observed on each spectrum. This FWHM value is nearly the same as that observed in a ZnSe random laser . As shown in Fig. 3(a), the observed lasing spike pattern fluctuates from pulse to pulse. This fluctuation is a typical feature of random laser and has been observed in other random laser systems [15,18,19]. Note that no lasing spike appears in the spectrum P1, although the lasing probability is very high because of sufficient lasing excitation power.
We performed the Fourier transformation (FT) of the emission spectra to estimate the cavity length of the random laser . The cavity length is obtained from the relation
Figure 3(b) shows the FT spectra of the three pulse-to-pulse spectra, P1, P2, and P3, in Fig. 3(a). The spectral resolution of the FT spectra is about 0.3 μm, which is estimated from the spectral resolution of lasing spectra in Fig. 3(a), ∼0.1 nm. No clear FT component is observed in the FT spectrum of P1. This is because no clear lasing signal is detected in its emission spectrum. The main peak at d = 7 μm and its higher-order peaks are clearly observed in the FT spectrum of P2. In P3, peaks at d = 11 and 17 μm as well as the FT component at d = 7 μm are observed. By introducing the refraction index value of n = 2.617 at 3.1 eV (387.5 nm) , the cavity diameters are determined to be D = 5 μm for P2 and 5, 8, and 13 μm for P3. These values are comparable to that obtained in a ZnSe random laser (∼6 μm) . Note that these cavity diameters are much smaller than the spot size of a Nd:YAG excitation source (∼300 μm).
The FT analysis indicates the formation of lasing cavities with different diameters by pulse-to-pulse excitation. These cavities are considered to be located at spatially different positions in the lasing medium. To investigate the spatial-dependent random lasing properties in detail, spatially resolved lasing emission spectra were measured. Figure 4(a) shows spatially (vertical axis) and spectrally (horizontal axis) resolved lasing emission profiles. The color scale bar represents the emission intensity. Each spectrum is obtained after integration over 50 excitation pulses.
Several lasing spots are observed in Fig. 4(a). In the rectangular area A1, the large lasing spot is observed at the 400 to 450 μm CCD position. This lasing spot exhibits the highest intensity. A small lasing spot is also observed at around 350 μm in the rectangular area A2. This spot emits a relatively weak lasing light. Note that the lasing intensity fluctuates in space due to pulse-to-pulse fluctuation, and thus the reproducibility of the spatial mode map is low in short-range order. This spatial intensity fluctuation was also observed in the literature . However, the overall (long-range order) feature, i.e., the positions of large lasing spot A1 and small lasing spot A2, is not changed from pulse to pulse.
In the inset of Fig. 4(b), we show spatially resolved lasing emission spectra in the rectangular areas A1 and A2. Because these spectra are obtained after integration over 50 excitation pulses, a broad peak together with spikelike structures is observed in both spectra. Figure 4(b) shows the FT spectra obtained from these emission spectra. The broad FT component in the spectrum of A2 peaks is observed at around d = 14 μm. This component is the second-order peak of the main component observed at d = 7 μm in the pulse-to-pulse FT spectra [Fig. 3(b)]. In the FT spectrum of A1, the weak FT component at 22 μm as well as the component at 14 μm are observed. The 22 μm component may be the second-order peak of the main FT component at 11 μm observed in Fig. 3(b).
The spatially resolved FT analysis revealed that the A1 area has several cavities, at least two cavities with D = 5 and 8 μm. This result can be easily understood from the fact that the A1 area has widely extended and strongly lasing spots [see Fig. 4(a)]. It should be noted that the size of the smaller cavity in the A1 area is the same as that in the A2 area, D = 5 μm. This may indicate that there is a threshold cavity size for lasing, which is determined, for example, from the gain and loss values, the mean free path of scattered light, and the lasing wavelength.
Finally, we determine the difference between the lasing spatial profile and that of the scattered excitation light at the random medium (i.e., GaN aggregates). Figure 5(a) shows the spatial intensity profile for the scattered excitation light (355 nm) in the same area and using the same CCD system as that in Fig. 4(a). The intensity profile in Fig. 5(a) shows a highly irregularly pattern. The irregularity is due to the inhomogeneous scattering event of coherent light at the random medium. The observed irregular pattern, which is termed the “speckle pattern,” contains much information on the random medium . The speckles in Fig. 5(a) appear in the CCD position range from 200 to 500 μm. This pattern is clearly different from the intensity profile of the lasing emission in Fig. 4(a).22]. Figure 5(b) shows the calculated correlation intensities C(Δr) as a function of the correlation length Δr for the lasing emission (390 nm) and scattered light (355 nm). The calculated data in the random laser was normalized to the value at Δr = 6 μm because the value at Δr = 0 μm has large experimental noise .
As shown in Fig. 5(b), the lasing correlation length plot decays more slowly than that of the scattered light. The decay lengths were estimated by fitting the correlation intensity plots with a single exponential function. The obtained decay lengths are 33 and 11 μm for the lasing emission and scattered light, respectively. The longer decay length for the lasing emission indicates that the size of the lasing spots is larger than that of the speckles. The same tendency was reported in Ref. . The size of the lasing spots is also confirmed to be smaller than that of the excitation area (∼300 μm). Note that the size of the speckles (∼ 11 μm) is much larger than that predicted on the basis of a theory (∼0.5 μm) . This may be due to the low spatial resolution of our measurement setup (∼ 6 μm).
We observed random lasing emission from powdered GaN with Au nanoparticles and determined the spatial extent of lasing spots by analyzing the spatially and spectrally resolved lasing emission spectra. In the presence of Au nanoparticles on GaN aggregates, several lasing lines were observed near the center of the GaN band-edge emission above the excitation threshold (≥2 MW/cm2), while GaN aggregates without Au exhibited only spontaneous emission. The lasing was due to loss compensation by surface plasmon excitation on Au and to the resonant feedback in each random-mode cavity, i.e., single GaN aggregates. From the spatially resolved lasing emission spectra and their FT analysis results, the formation of random lasing cavities at different spatial positions was confirmed. The size of the random lasing spots for the GaN/Au aggregates was found to be larger than that for the scattered light of the excitation source using the intensity autocorrelation analysis.
This work was partly supported by a Grant-in-Aid for Young Scientists (B) (22760050) from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
References and links
1. M. A. Noginov, Solid-State Random Lasers (Springer, New York, 2005).
2. H. Cao, Y. G. Zhao, S. T. Ho, E. W. Seelig, Q. H. Wang, and R. P. H. Chang, “Spatial confinement of laser light in active random media,” Phys. Rev. Lett. 82, 2278–2281 (1999). [CrossRef]
3. H. Y. Yang, S. F. Yu, S. P. Lau, S. H. Tsang, G. Z. Xing, and T. Wu, “Ultraviolet coherent random lasing in randomly assembled SnO2 nanowires,” Appl. Phys. Lett. 94, 241121 (2009). [CrossRef]
4. M. Sasaki, Y. Inose, K. Ema, T. Ohtsuki, H. Sekiguchi, A. Kikuchi, and K. Kishino, “Random laser action in GaN nanocolumns,” Appl. Phys. Lett. 97, 151109 (2010). [CrossRef]
5. M. A. Noginov, G. Zhu, I. Fowlkes, and M. Bahoura, “GaAs random laser,” Laser Phys. Lett. 1, 291–293 (2004). [CrossRef]
9. C. W. Cheng, E. J. Sie, B. Liu, C. H. A. Huan, T. C. Sum, H. D. Sun, and H. J. Fan, “Surface plasmon enhanced band edge luminescence of ZnO nanorods by capping Au nanoparticles,” Appl. Phys. Lett. 96, 071107 (2010). [CrossRef]
10. O. Popov, A. Zilbershtein, and D. Davidov, “Random lasing form dye-gold nanoparticles in polymer films: Enhanced at the surface-plasmon-resonance wavelength,” Appl. Phys. Lett. 89, 191116 (2006).
11. G. D. Dice, S. Mujumdar, and A. Y. Elezzab, “Plasmonically enhanced diffusive and subdiffusive metal nanoparticles-dye random laser,” Appl. Phys. Lett. 86, 13110 (2005). [CrossRef]
12. A. Kumar, S. F. Yu, and X. F. Li, “Random laser action in dielectric-metal-dielectric surface plasmon waveguides,” Appl. Phys. Lett. 95, 231114 (2009). [CrossRef]
13. A. Tulek, R. C. Polson, and Z. V. Vardeny, “Naturally occurring resonators in random lasing of π-conjugated polymer films,” Nat. Phys. 6, 303–310 (2010). [CrossRef]
14. J. Fallert, R. J. B. Dietz, J. Sartor, D. Schneider, C. Klingshirn, and H. Kalt, “Co-existence of strongly and weakly localized random laser modes,” Nat. Photonics 3, 279–248 (2009). [CrossRef]
15. R. G. S. El-Dardiry, A. P. Mosk, O. L. Muskens, and A. Lagendijk, “Experimental studies on the mode structure of random lasers,” Phys. Rev. A 81, 043830 (2010). [CrossRef]
16. T. Karakouz, A. B. Tesler, T. A. Bendikov, A. Vaskevich, and I. Rubinstein, “Highly stable localized plasmon transducers obtained by thermal embedding of gold island films on glass,” Adv. Mater. 20, 3893–3899 (2008). [CrossRef]
17. M. A. Reshchikov and H. Morkoc, “Luminescence properties of defects in GaN,” J. Appl. Phys. 97, 061301 (2005). [CrossRef]
18. S. Mujumdar, V. Türck, R. Torre, and D. S. Wiersma, “Chaotic behavior of a random laser with static disorder,” Phys. Rev. A 76, 0338071 (2007). [CrossRef]
20. S. Adachi, Handbook on Physical Properties of Semiconductors (Springer, New York, 2004).
21. T. Takahashi, T. Nakamura, and S. Adachi, “Blue-emitting ZnSe random laser,” Proc. SPIE 7597, 75971T (2010). [CrossRef]