Subscriber access provided by Massachusetts Institute of Technology
J. Phys. Soc. Jpn. 92, 064710 (2023) [4 Pages]
FULL PAPERS

Thickness Dependence on Wavelength Range of Random Laser in Ultrathin ZnO Crystals Grown by Mist-CVD on C-plane Sapphire Substrate

+ Affiliations
1Faculty of Engineering, Shizuoka University, Hamamatsu 432-8561, Japan2Interdisciplinary Graduate School, University of Yamanashi, Kofu 400-8511, Japan

The ultrathin ZnO crystals were grown by mist chemical vapor deposition technique with potentially low-cost and low-environmental impact on commercially available c-plane sapphire substrates. The photoluminescence spectra of the ultrathin ZnO crystals were evaluated, and we observed the random lasing behaviors. In addition, the range of random laser wavelengths of photoluminescence can be shifted with a constant trend by changing the structural parameters of the ultrathin ZnO crystals. Here, the structural parameter was the thickness of the ultrathin ZnO crystals. The results indicate that the oscillation wavelength range of the random laser may be potentially fine-tuned by changing the structural parameters of the random laser medium.

©2023 The Physical Society of Japan
1. Introduction

Random lasers have been investigated as one of the optical microcavities, and random lasers have been realized using various materials and structures such as nanoparticles, nanocrystals and thin layers.17) A large number of microcavity optical paths can be realized in such media, and low or in-coherence random lasers can be potentially realized. It is expected that speckle-free laser imaging will be performed with random lasers.8) Basically, it is difficult to precisely adjust the oscillation wavelengths of random lasers, but the oscillation wavelength-range can be tuned within a certain range by, for example, changing the materials (optical gain) used as media for random lasers.911) Realizing the precise control of the oscillation wavelength of random lasers, one can expect their wide optical applications similarly to photonic crystals and whisper-gallery mode resonators. Moreover, it is desirable for the fabrication techniques of such optical materials for advanced optical applications to be fully considered for sustainable development. Here, mist chemical vapor deposition (mist-CVD) and related techniques,1217) which may potentially realize low-cost and low-environmental-impact fabrication techniques, are candidates for meeting these challenges. Thus far, we have demonstrated the fabrication of random lasers using ZnO and ZnO based crystals grown by mist-CVD.11,18)

In this study, we fabricated ultrathin ZnO crystals by mist-CVD on a c-plane sapphire substrate as a medium for random lasers. Then, we investigated the optical properties of random lasers with different thickness of ultrathin ZnO crystals, and considered the possibility of the precision control of the oscillation wavelength range of random lasers.

2. Experimental Results and Discussion

Using a mist-CVD crystal growth system, we fabricated ultrathin ZnO crystals on a c-plane sapphire substrate. Our mist-CVD system has been described in our previous report.18) Mist was generated from an aqueous solution prepared by dissolving zinc acetate dihydrate in pure water at a concentration of approximately 0.03 mol kg−1. The generated mist was carried by nitrogen gas at approximately 4 L min−1 flow rate. The distance between the nozzle exit and the substrate was approximately 1 cm. The crystal growth temperature (i.e., the temperature of the tube furnace) and crystal growth time were approximately 960 °C and 3 min, respectively. The short growth time resulted in the formation of ultrathin ZnO crystals. As previously described, owing to the positional relationship between the central axis of the nozzle and the substrate, the crystal growth did not proceed with in-plane uniformity on the substrate. This enabled the growth of the ultrathin ZnO crystals with different thicknesses in-plane on the substrate. By using the ultrathin ZnO crystals with such geometries, we investigated the relationships between random laser properties and the ZnO crystal thickness. The thickness range we evaluated in this study was approximately 43–91 nm.

Room-temperature photoluminescence (RT-PL) spectra of ultrathin ZnO crystals were evaluated using a high-power excitation source of a Nd:YAG pulse laser [266 nm, pulse duration of 6–9 ns, pulse repetition rate of 2 Hz, roughly elliptical excitation spot (the diameters of the excitation area were approximately 225 and 75 µm)]. Figure 1 shows examples of the RT-PL spectra obtained from ultrathin ZnO crystals with thicknesses of 43 and 91 nm, and they originated from a single excitation laser pulse (1.5 MW/cm2). As shown in Fig. 1, each RT-PL spectrum showed multiple sharp peaks (and/or spikes) with different shapes, which were randomly obtained. Such features of the RT-PL spectra indicate that random lasers were realized1,2,6) in ultrathin ZnO crystals, and the approximate oscillation wavelengths of the ZnO crystals with thicknesses of 43 and 91 nm were 375 and 395 nm, respectively. As shown in the SEM images in Fig. 2, there were many nanometer order irregularities on the surfaces of the ultrathin ZnO crystals. We speculate that these nanometer order irregularities were the origin of multiple scattering (random lasers). The size order of the irregularities was generally similar to that of the 43 and 91 nm samples. Figure 3 shows examples of the RT-PL spectra obtained from the ultrathin ZnO crystals with thicknesses from 43 to 91 nm, which were excited by a single excitation laser pulse (1.5 MW/cm2). As shown in Fig. 3 when the thickness became smaller than approximately 55 nm, the oscillation wavelengths of the random lasers shifted toward the short wavelengths. Their shapes changed with each excitation laser pulse, and it was difficult to evaluate their properties quantitatively. Therefore, we evaluated the sums of the many RT-PL spectra of the ultrathin ZnO crystals excited by a single pulse. We consider that the sums show the average photoluminescence properties of the random lasers and that the spectra enable the quantitative evaluation of the average oscillation-wavelength range of the random lasers. Figure 4 shows the crystal thickness dependence of the RT-PL peak wavelengths determined from the many RT-PL spectra accumulated (approximately 50 excitation pulses) with the excitation power densities in the cases of 0.5, 1.0, and 1.5 MW/cm2. As a result, when the crystal thickness became smaller than approximately 55 nm, the trend of the shift toward the shorter wavelength side for the oscillation-wavelength range of the random lasers became clear. The thickness decrease of approximately 10 nm resulted in an oscillation-wavelength-range shift of approximately 15–20 nm at RT.


Figure 1. (Color online) Examples of series of RT-PL spectra of ultra-thin ZnO crystals with (a) 43 and (b) 91 nm thicknesses, which were excited by a single excitation pulse of a Nd:YAG pulse laser.


Figure 2. Typical surface SEM images of near ultrathin ZnO crystals with (a) 43 and (b) 91 nm thicknesses, which were excited by a single excitation pulse of a Nd:YAG pulse laser as shown in Fig. 1.


Figure 3. (Color online) Examples of series of RT-PL spectra of ultra-thin ZnO crystals from 43 to 91 nm thicknesses, which were excited by a single excitation pulse of a Nd:YAG pulse laser.


Figure 4. (Color online) Thickness dependence of the RT-PL peak wavelengths determined from the many RT-PL spectra accumulated (approximately 50 excitation pulses) with the excitation power densities with the cases of 0.5, 1.0, and 1.5 MW/cm2.

Figure 5(a) shows the RT-PL spectra obtained from the ultrathin ZnO crystals with thicknesses from 43 to 91 nm, which were excited by a low-power excitation source of a He-Cd CW laser @ 325 nm wavelength. Using the low-power excitation source, we can generally evaluate PL spectra. As shown in Figs. 5(a) and 5(b), the emission intensity decreased with decreasing thickness, which was due to the decrease in optical material volume. The emission wavelength peaks and shapes of the RT-PL spectra remained almost unchanged with decreasing thickness. This indicates that the RT-PL properties of the ZnO crystals remained almost unchanged with the decrease in their thickness. Therefore, the oscillation-wavelength range shift may have originated from the effects of crystal geometries, specifically, the thickness itself. These results indicate that the oscillation-wavelength range of random lasers may be potentially fine-tuned by changing their structural parameters, particularly the thickness.


Figure 5. (Color online) The crystal thickness dependence of (a) RT-PL spectra and (b) RT-PL peak intensities and wavelengths when the ZnO crystals excited by a He-Cd CW laser.

Here, we discuss the origin of the shift. An oscillating light based on a random laser is believed to propagate along a small closed loop with multiple scatterings in ultrathin ZnO crystals. Then, we estimate the propagation properties of an asymmetric waveguide [sapphire substrate (substrate)/ultrathin ZnO crystals (core layer)/air (cladding layer)] for TE0 modes (electric field parallel to the c-plane sapphire surface) of 375 and 395 nm wavelengths. In our analysis, we employ relatively reasonable values of the sapphire and ZnO refractive indices as 1.75(n\(_{\text{sapphire}}\)) and 2.05–2.25(n\(_{\text{ZnO}}\)), respectively. Figure 6 shows the estimated confinement factor of the ultrathin ZnO crystals, which indicates that the confinement factor decreases with decreasing crystal thickness and is lower on the long wavelength side than on the short wavelength side. The difference in confinement factor for the wavelengths of 375 and 395 nm increases with decreasing thickness of ultrathin ZnO crystals. In the case of \(\mathrm{n}_{\text{ZnO}} = 2.15\) and \(\text{thickness}= 45\) nm, the confinement factor for the wavelength of 375 nm is about 2.2 times that for the wavelength of 395 nm. Then, in the case of \(\mathrm{n}_{\text{ZnO}} = 2.15\) and \(\text{thickness}= 55\) nm, the confinement factor for the wavelength of 375 nm is about 1.2 times that for the wavelength of 395 nm. In addition, in the case of \(\mathrm{n}_{\text{ZnO}} = 2.15\), light with the wavelength of 395 nm cannot be guided if the thickness is 43 nm or less (cutoff). The lasing wavelength of optical microcavities formed with such optical waveguide is considered to be proportional to the confinement factor and the optical gain of the ultrathin ZnO crystals. Therefore, it is assumed that widening the difference in the magnitude of the confinement factor (the confinement factor is larger on the short wavelength side than on the long wavelength side) or the cutoff on the long wavelength side makes it easier to induce laser oscillation at shorter wavelengths than at longer wavelengths. The confinement factors do not differ significantly for a crystal thickness of 55 nm or more, and the laser oscillation wavelength is expected to depend on the wavelength at which the optical gain of the ultrathin ZnO crystal is greatly obtained. Therefore, we consider that this is the reason why no oscillation-wavelength range shift of random lasers was observed for a crystal thickness of 55 nm or larger. Since this discussion focuses on only one aspect (properties of the optical waveguide), further discussion (e.g., including optical gain properties under high excitation condition) will be required to clarify the origin of this phenomenon. The estimated confinement factor of TM0 modes is lower than that of TE0 modes in such an ultrathin ZnO crystal model, and we omitted this from the discussion above.


Figure 6. (Color online) Estimated confinement factors for TE0 modes at 375 and 395 nm wavelengths for sapphire/ZnO/air asymmetric slab optical waveguides.

3. Summary

In summary, we grew ultrathin ZnO crystals by mist-CVD, and evaluated the properties of random lasers by RT-PL measurements using high-power and low-power excitation lasers. We observed that the oscillation wavelengths of the random lasers were shifted by changing the thickness of the ZnO crystals. This indicates that the oscillation wavelength range of random lasers may be potentially fine-tuned by changing the structural parameters of the random lasing medium without changing the random laser medium material at RT. The possibility of tuning the oscillation wavelengths of random lasers with the mist-CVD-grown ZnO crystals can contribute to the development of environmentally friendly and advanced nano- and microphotonics and elements.


References

  • 1 H. Cao, Y. G. Zhao, S. T. Ho, E. W. Seelig, Q. H. Wang, and R. P. H. Chang, Phys. Rev. Lett. 82, 2278 (1999). 10.1103/PhysRevLett.82.2278 CrossrefGoogle Scholar
  • 2 S. F. Yu, C. Yuen, S. P. Lau, W. I. Park, and G. C. Yi, Appl. Phys. Lett. 84, 3241 (2004). 10.1063/1.1734681 CrossrefGoogle Scholar
  • 3 H. Y. Yang, S. F. Yu, S. P. Lau, S. H. Tsang, G. Z. Xing, and T. Wu, Appl. Phys. Lett. 94, 241121 (2009). 10.1063/1.3157842 CrossrefGoogle Scholar
  • 4 M. Sakai, Y. Inose, K. Ema, T. Ohtsuki, H. Sekiguchi, A. Kikuchi, and K. Kishino, Appl. Phys. Lett. 97, 151109 (2010). 10.1063/1.3495993 CrossrefGoogle Scholar
  • 5 B. Q. Sun, M. Gal, Q. Gao, H. H. Tan, C. Jagadish, T. Puzzer, L. Ouyang, and J. Zou, J. Appl. Phys. 93, 5855 (2003). 10.1063/1.1568533 CrossrefGoogle Scholar
  • 6 S. F. Yu, C. Yuen, S. P. Lau, and H. W. Lee, Appl. Phys. Lett. 84, 3244 (2004). 10.1063/1.1719279 CrossrefGoogle Scholar
  • 7 X. Yang, G. Feng, K. Yao, J. Yi, H. Zhang, and S. Zhou, AIP Adv. 5, 067160 (2015). 10.1063/1.4923317 CrossrefGoogle Scholar
  • 8 B. Redding, M. A. Choma, and H. Cao, Nat. Photonics 6, 355 (2012). 10.1038/nphoton.2012.90 CrossrefGoogle Scholar
  • 9 H. Y. Yang, S. P. Lau, S. F. Yu, M. Tanemura, T. Okita, H. Hatano, K. S. Teng, and S. P. Wilks, Appl. Phys. Lett. 89, 081107 (2006). 10.1063/1.2338525 CrossrefGoogle Scholar
  • 10 H. Y. Yang, S. F. Yu, and S. P. Lau, J. Cryst. Growth 312, 16 (2009). 10.1016/j.jcrysgro.2009.10.004 CrossrefGoogle Scholar
  • 11 T. Kanai, K. Fujiwara, M. Sakai, K. Hara, and T. Kouno, Jpn. J. Appl. Phys. 61, 055502 (2022). 10.35848/1347-4065/ac5e8f CrossrefGoogle Scholar
  • 12 Y. Kamada, T. Kawaharamura, H. Nishinaka, and S. Fujita, Jpn. J. Appl. Phys. 45, L857 (2006). 10.1143/JJAP.45.L857 CrossrefGoogle Scholar
  • 13 M. T. Htay, Y. Hashimoto, and K. Ito, Jpn. J. Appl. Phys. 46, 440 (2007). 10.1143/JJAP.46.440 CrossrefGoogle Scholar
  • 14 T. Kawaharamura, H. Nishinaka, and S. Fujita, Jpn. J. Appl. Phys. 47, 4669 (2008). 10.1143/JJAP.47.4669 CrossrefGoogle Scholar
  • 15 K. Kaneko, T. Onuma, K. Tsumura, T. Uchida, R. Jinno, T. Yamaguchi, T. Honda, and S. Fujita, Appl. Phys. Express 9, 111102 (2016). 10.7567/APEX.9.111102 CrossrefGoogle Scholar
  • 16 K. Uno, Y. Yamasaki, and I. Tanaka, Appl. Phys. Express 10, 015502 (2017). 10.7567/APEX.10.015502 CrossrefGoogle Scholar
  • 17 K. Ishii, M. Ono, K. Kaneko, T. Onuma, T. Honda, and S. Fujita, Appl. Phys. Express 12, 052011 (2019). 10.7567/1882-0786/ab10e2 CrossrefGoogle Scholar
  • 18 J. Iwata, M. Sakai, K. Ohashi, K. Hara, and T. Kouno, Jpn. J. Appl. Phys. 60, 100901 (2021). 10.35848/1347-4065/ac21b2 CrossrefGoogle Scholar