記住我
In this study, we have attempted to enhance the bandgap emission from GaN epilayer in the quadrupole oscillation mode of LSPR using fabricated Ag NPs. Moreover, we have attempted to reproduce the obtained PL enhancements using FDTD calculations.
The fabrication procedure for the samples is shown in Fig. 1. GaN epilayers with 4-μm-thick were grown on sapphire substrates via metal-organic chemical vapor deposition (MOCVD). Ag thin films were deposited on GaN substrates using resistor heating vapor deposition (Sanyu electron SVC-700TM). The samples were annealed in an electric furnace in a nitrogen atmosphere. Ag nano-hemispherical structures were obtained by depositing 5, 10, and 15 nm thick Ag on GaN substrates, followed by annealing at 300 °C for 10 min, 350 °C for 15 min, and 400 °C for 15 min in a nitrogen atmosphere. The size of the Ag nano-hemispherical structures was tuned by changing the thickness of the Ag thin film. It was difficult to form Ag NPs for Ag thin films that were less than 5 nm in size. SiO2 spacer layers of 5 nm thickness were fabricated between Ag and GaN using high-vacuum RF sputtering (Sanyu Electron, SVC-700RF).The reflection spectra of the samples were measured using a reflectance measurement attachment (5° incidence angle, Shimadzu UV-1800, Japan). The surface morphologies of the Ag nanostructures were observed by atomic force microscopy (AFM) (Nano Wizard, Bruker-JPK Instruments AG). The PL spectra were measured using a multi-channel spectrometer (Roper Scientific, PIXIS 100B-3). A He-Cd laser (Kimmon, IK3202R-D) was used as the excitation light (λ = 325 nm) with a power of 200 mW. The spot size is 1 mm.
The electromagnetic field and resonance spectra of the sample structure model were analyzed by FDTD calculations (Poynting for Optics, Fujitsu, Japan). A nonuniform mesh was used with a grid size of 0.5–5 nm. A periodic boundary condition was set in the X- and Y-directions, and an absorbing boundary condition was set in the Z-direction. A pulsed light composed of a differential Gaussian function with a pulse width of 0.5 fs with an x-polarized electric field of 1 V m−1 was used as the excitation light. The dielectric function of Ag was approximated by the Drude formula based on the values reported by Johnson and Christy.2121. P. B. Johnson and R. W. Christy, Phys. Rev. B 6(12), 4370–4379 (1972). https://doi.org/10.1103/PhysRevB.6.4370 The diameter of the Ag NPs was set as 100 nm, and the thickness of the SiO2 thin film was set as 5 nm. The Ag NPs were placed on top of the GaN substrate. The refractive indices (n) of SiO2 and GaN were set to 1.5 and 2.4, respectively, without dispersion. The dipole light source of the modulated Gaussian function (duration = 50 fs, center = 25 fs, width = 10 fs, period = 1.22 fs, and λ = 365 nm) was positioned 5–100 nm below the contact point of GaN and the Ag NPs. This was done because the penetration depth of the excitation light, that is, the He-Cd laser light (λ = 325 nm), on GaN was 100 nm.The obtained AFM images and reflectance spectra of the fabricated samples are shown in Fig. 2. Ag NPs' size is about 90–180 nm range when the Ag film thickness is 15 nm. It is mostly distributed in the range of 90–150 nm with 126 nm as the average value. When the Ag film thickness was 5 nm, Ag NPs' size average was 58.9 nm. The Ag NPs we have fabricated were independent, and we could not observe any dimer effects in Fig. 2 as shown in other papers.22,2322. Y. Li, B. Liu, R. Zhang, Z. Xie, Z. Zhuang, J. Dai, T. Tao, T. Zhi, G. Zhang, P. Chen, F. Ren, H. Zhao, and Y. Zheng, J. Appl. Phys. 117, 153103 (2015). https://doi.org/10.1063/1.491855523. H. Imadate, T. Mishima, and K. Shiojima, Jpn. J. Appl. Phys., Part 1 57(4S), 04FG13 (2018). https://doi.org/10.7567/JJAP.57.04FG13A hemispherical structure was clearly observed in the AFM image of the Ag NPs immediately after thermal annealing of the 5 nm thick Ag film directly located on the GaN substrate [Fig. 2(a)]. The reflection spectrum in panel (b) contains one peak representing the dipole mode in the range of approximately 500–700 nm and another peak representing the quadrupole mode at approximately 350 nm. Within 2 days, the structures of the Ag nano-hemispheres were found to be destroyed, and the resonance peaks disappeared. This is likely due to oxidation and sulfurization of Ag NPs on GaN in air. Ag nanoparticles are known to be chemically unstable and have been reported to corrode when exposed to ambient laboratory air.2424. M. D. Mcmahon, R. Lopez, H. M. Meyer III, L. C. Feldman, and R. F. Haglund, Jr., Appl. Phys. B 80, 915–921 (2005). https://doi.org/10.1007/s00340-005-1793-6The main peak of the LSPR resonance spectra corresponding to the dipole mode was not suitable for enhancing the blue and green light emissions of InGaN/GaN QWs in the range 450–550 nm or for enhancing UV light emissions from GaN epilayer. To enhance these emissions, the LSPR peak wavelength of the dipole mode needed to be shortened. The LSPR peak of the quadrupole mode was located at approximately 360 nm that was useful for enhancing the band edge emission of GaN around 365 nm. In addition, the reflectance around 320 nm was low and enabled the penetration of the excitation light with a wavelength of 325 nm into the Ag NPs and excited the GaN epilayer.
Figure 2(b) shows the AFM images and the reflectance spectra of the Ag NPs formed from the 15 nm thick Ag film directly located on the GaN substrate. The Ag NPs were formed after thermal annealing increased with the increase in the initial Ag film thickness. In fact, as the Ag NP sizes increased, the LSPR peak of the dipole mode was redshifted. Simultaneously, the LSPR peak of the quadrupole did not change much with the change in wavelength, but its intensity increased. Therefore, a thickness of 15 nm was required for the Ag film to enhance the bandgap emission from GaN epilayer.When we set the initial thickness of the Ag film to 15 nm, the Ag nano-hemisphere structures were retained in 2 days. There was no significant change in the shape, but the LSPR peak became slightly smaller [Fig. 2(b)]. The Ag nano-hemisphere structures were finally destroyed and became smaller after 10 months. A technique to prevent Ag NPs from deteriorating is by coating the Ag NPs located directly on the GaN substrate with an SiO2 thin film.17,2217. C. Langhammer, M. Schwind, B. Kasemo, and I. Zorić, Nano Lett. 8(5), 1461–1471 (2008). https://doi.org/10.1021/nl080453i22. Y. Li, B. Liu, R. Zhang, Z. Xie, Z. Zhuang, J. Dai, T. Tao, T. Zhi, G. Zhang, P. Chen, F. Ren, H. Zhao, and Y. Zheng, J. Appl. Phys. 117, 153103 (2015). https://doi.org/10.1063/1.4918555 Therefore, we fabricated samples with SiO2 coated Ag NPs located directly on GaN to prevent the deterioration of Ag NPs [the fabrication process is shown in Figs. 1(a-i)–1(a-iii)].Figure 2(c) shows the AFM images and the reflectance spectra of the Ag NPs formed from the 15 nm thick Ag film coated with a 5 nm thick SiO2 thin film directly located on GaN. The Ag nano-hemisphere structures were protected by the SiO2 thin film; however, the reflection spectra became much smaller in 2 days. This suggested that the Ag nano-hemisphere structures may have oxidized from the inside. The SiO2 coating can protect the Ag NPs from oxidation in the air, but it cannot prevent the fine Ag NPs from ionizing and oxidizing from the inside. Oxidation may have occurred through the interface between GaN and Ag. Therefore, we inserted a 5 nm thick SiO2 film on the GaN substrate and subsequently formed Ag nano-hemispherical structures on it [the fabrication process is shown in Figs. 1(b-i)–1(b-iii)]. In Figs. 2(a)–2(c), the peak of the quadrupole resonances hardly shifted while the peak of the dipole resonances shifted more than 100 nm because dipole resonance wavelength is a function of the particle geometry and the dielectric constant unlike quadrupole oscillation. Figure 3 in the supplementary material shows that the peak of the quadrupole oscillation mode is independent of the size of the NPs. These results also seem to be due to the fact that these two types of resonances are fano-resonant.The AFM images and spectra of the samples are shown in Fig. 2(d). The inserted SiO2 thin film prevented the Ag NPs from aggregating and maintained the effectiveness of the LSPR intensity even after 10 days. In addition, we coated the Ag NPs with a 5 nm SiO2 film to protect them from oxidation [the fabrication process is shown in Figs. 1(b-i)–1(b-iv)]. The inserted SiO2 spacer and SiO2 coating prevented the Ag NPs from aggregating and maintained the LSPR intensity [Fig. 2(e)].The electron transfer from Ag to the GaN may have been caused by the ionization of Ag owing to the large difference in the Fermi level between Ag and GaN. However, a sufficiently high Schottky barrier was formed in the undoped GaN and, especially, in Si-doped GaN (n-type GaN), thereby making the transfer of electrons difficult from Ag to the GaN.23,2523. H. Imadate, T. Mishima, and K. Shiojima, Jpn. J. Appl. Phys., Part 1 57(4S), 04FG13 (2018). https://doi.org/10.7567/JJAP.57.04FG1325. T. U. Kampen and W. Mönch, Metal Contacts on a-GaN ( Cambridge University Press, 2014). We obtained similar results using n-type GaN to those obtained using undoped GaN. Subsequently, we investigated the existence of Ga vacancies on the GaN surface.2626. K. Saarinen, T. Laine, S. Kuisma, J. Nissilä, P. Hautojärvi, L. Dobrzynski, J. M. Baranowski, K. Pakula, R. Stepniewski, M. Wojdak, A. Ysmolek, T. Suski, M. Leszczynski, I. Grzegory, and S. Porowski, Observation of Native Ga Vacancies in Gan by Positron Annihilation ( Cambridge University Press, 2011). Ga vacancies can assume the characteristics of acceptors and attract electrons owing to the large electronegativity of nitrogen.27,2827. J. Neugebauera and C. G. van de Walle, Appl. Phys. Lett. 69, 503 (1996). https://doi.org/10.1063/1.11776728. J. L. Lyons and C. G. Van de Walle, npj Comput. Mater. 3, 12 (2017). https://doi.org/10.1038/s41524-017-0014-2 Electron transfer at the interface without crossing the Schottky barrier can promote Ag ionization and destroy the shape of Ag NPs. Such electron transfer to the Ga vacancies on the surface of the GaN substrate was a possible reason for the shape changes of the Ag NPs on GaN.Figure 4 shows the PL measurements of the samples with Ag NPs formed from the 15 nm thick Ag. Since the absorption components of Ag NPs at 325 nm of the excitation light were small in all spectra in Fig. 2, we concluded that the absorption and scattering of the excitation light by NPs have little effect on the PL spectra. The PL intensity of the sample with Ag NPs directly located on the GaN substrate was immediately quenched (purple line). The PL spectrum was also quenched similarly when only the outside of the NPs was coated with SiO2 (green line). The observed quenching effect of the PL spectra of GaN when the Ag NPs were placed directly on GaN was caused by the destruction of the shapes of the Ag NPs and by the Förster resonance energy transfer (FRET) from GaN to Ag.2929. J. Bohlen, Á. Cuartero-González, E. Pibiri, D. Ruhlandt, A. I. Fernández-Domínguez, P. Tinnefeld, and G. P. Acuna, Nanoscale 11, 7674–7681 (2019). https://doi.org/10.1039/C9NR01204D Simultaneously, the sample with an inserted SiO2 film between the GaN and Ag NPs resulted in a PL enhancement of 1.8-fold higher than that of GaN (blue line). The PL intensity of the sample that has both the SiO2 spacer between GaN and Ag and SiO2 coating outside the Ag NPs was about 2.2 times higher than that of GaN (red line). We concluded that the SiO2 thin film not only maintained the shape of the fabricated Ag NPs but may also have played an effective role in preventing contact with the metal surface to reduce emission quenching. From the above, we demonstrated in enhancing the bandgap emission from GaN epilayer by using LSPR in the quadrupole oscillation mode using Ag NPs on GaN.Hereafter, we would like to predict the emission enhancement due to surface plasmon resonance by FDTD calculation. Since the Purcell effect is a quantum phenomenon due to the interaction between excitons and resonators, it cannot be reproduced by FDTD based on classical electromagnetism. On the other hand, when dipole oscillator is placed near metallic nanostructures, the optical radiation from the dipole oscillator is enhanced by resonance. This phenomenon is called an electromagnetic optical antenna and can be calculated with FDTD. We compare the quantum mechanical Purcell effect with the electromagnetic optical antenna effect and note their similarity.
Figure 5(a) shows the calculated enhancement factors of the PL intensities through dipole oscillations (625 nm) and quadrupole oscillations (365 nm) plotted against the distance between the point dipole light source and the metal interface. The optical antenna effect increased the radiation from the point dipole and maybe an electromagnetic equivalent of the Purcell effect. The enhancement factor owing to the optical antenna effect in the dipole oscillation mode was very high because this mode was a bright mode. The enhancement factor due to the optical antenna effect in the quadrupole oscillation mode still existed in the dark mode where the interaction with light was weak. This is an important result of the study. This figure also suggests that the effective enhancement effect was not observed when the light source was located further than 30 nm from the Ag NPs. We calculated the distance dependence between the point light source and NPs of enhancement factor to understand what range of point sources act on the optical antenna effect. The effect of optical antenna and the inserting oxide film was reduced, and dark modes, which are the original nature of quadrupole oscillation modes, become dominant after 30 nm. That is why there is the degradation of the enhancement factor after 30 nm.To verify the role of the thin film between the GaN and Ag NPs, we evaluated various models with thin films of SiO2 (n = 1.5), ITO (n = 2.2), ZnO (n = 2.0), and air (n = 1.0). The results are presented in Fig. 5(b). There was a dip in the PL enhancement ratio around a distance of 50 nm between the Ag NPs and the point dipole source. At this distance, the quadrupole oscillation was very symmetrical and canceled each other out, thereby entering a fully dark mode. In contrast, when the distance was 10 nm, the quadrupole oscillation became so asymmetrical that a relatively large amount of light was emitted. There was no noticeable LSPR or quadrupole oscillation on Ag NPs further than 60 nm in the presence or absence of a thin film. This suggested that only when the point dipole source were in close to the NPs. The Ag NPs were influenced by LSPR and the oxide film.Figure 5(c) shows the extinction spectra for each sample using the FDTD calculations. The peak positions of the quadrupole mode around 350 nm did not change, whereas the peak position of the dipole mode around 600–700 nm was redshifted with increasing refractive index of the spacer layers. In contrast to dipole resonances, the quadrupole resonances unchanged in both FDTD simulation and experiment when the Ag NPs' size changed. In addition, it almost unchanged when the refractive index was changed. These results are due to fano-resonance.3030. K. Okamoto, K. Okura, P. Wang, S. Ryuzaki, and K. Tamada, Nanophotonics 9, 3409–3418 (2020). https://doi.org/10.1515/nanoph-2020-0118 Figure 5(d) shows the calculated PL intensities for the various spacer layers. The enhancement factor increased with a decrease in the refractive index of the layers. When the refractive index of the inserted thin film was significantly different from that of GaN, the environment around the point dipole light source became asymmetric, and the source was more likely to couple into the quadrupole oscillation mode, which is a dark mode. This can increase both the local enhancement of the electromagnetic field and the light extraction from the LSPR mode. This result suggested that the PL enhancement factor can be modified and optimized by the refractive index of the spacer layer.In this study, we enhanced the bandgap emission from GaN epilayer by LSPR in the quadrupole oscillation mode using Ag NPs on GaN. The shape of the Ag NPs directly located on the GaN substrate was found to be destroyed over time, and the Ag NPs directly fabricated on the GaN substrate did not enhance the PL through SP resonance. We proposed the insertion of an SiO2 thin film between the Ag NPs and GaN to solve these problems. We demonstrated in suppressing the shape change of Ag NPs and the absorption loss of the SPs, thereby resulting in a high efficiency of light emitted in the UV region. This technique can be very effective for application in SP-enhanced efficiency light-emitting devices. In addition, we reproduced the high-efficiency PL of the GaN substrate through the optical antenna effect of the dipole point source located near the metal nanostructures using the FDTD method. In the simulation, a localized electric field was found to be induced in the SiO2 thin film and the localized electric field of the metal NPs also increased. The simulation of various thin films as spacer layers revealed that oxide thin films other than SiO2 were also useful for PL enhancement.
See the supplementary material for histogram of the size of Ag NPs' size measured from the AFM image, observed extinction spectra for various inter-particle distances, simulated extinction spectra with and without a 5 nm thick SiO2 layer, and extinction spectra with various radii in FDTD simulation and experiment.The authors wish to thank Professor K. Tamada of Kyushu University and Professor M. Terazima of Kyoto University for their valuable discussions and support. This work was supported by the JSPS Grants-in-Aid for Specially Promoted Research (No. JP20H05622) and Scientific Research (S) (No. JP19H05627), and exploratory research (No. JP21K19218).
Conflict of Interest
The authors have no conflicts to disclose.
Author Contributions
Seiya Kaito: Data curation (lead); Formal analysis (lead); Software (lead); Visualization (equal); Writing – original draft (lead). Tetsuya Matsuyama: Validation (equal); Writing – review & editing (equal). Kenji Wada: Validation (equal); Writing – review & editing (equal). Mitsuru Funato: Conceptualization (equal); Funding acquisition (equal); Methodology (equal); Validation (equal); Writing – review & editing (equal). Yoichi Kawakami: Conceptualization (equal); Funding acquisition (equal); Methodology (equal); Validation (equal); Writing – review & editing (equal). Koichi Okamoto: Conceptualization (lead); Funding acquisition (lead); Methodology (lead); Project administration (lead); Resources (lead); Supervision (lead); Writing – review & editing (lead).
The data that support the findings of this study are available from the corresponding author upon reasonable request.
REFERENCES
1. S. Kusanagi, Y. Kanitani, Y. Kudo, K. Tasai, A. A. Yamaguchi, and S. Tomiya, Jpn. J. Appl. Phys., Part 1 58(SC), SCCB28 (2019). https://doi.org/10.7567/1347-4065/ab0f11, Google ScholarCrossref2. Q. Wang, K. Li, and M. Liu, Superlattices Microstruct. 122, 46–56 (2018). https://doi.org/10.1016/j.spmi.2018.08.023, Google ScholarCrossref3. D. Dobrovolskas, J. Mickevičius, S. Nargelas, H. S. Chen, C. G. Tu, C.-H. Liao, C. Hsieh, C. Y. Su, G. Tamulaitis, and C. C. Yang, Plasmonics 9(5), 1183–1187 (2014). https://doi.org/10.1007/s11468-014-9729-9, Google ScholarCrossref4. K. Okamoto, M. Funato, Y. Kawakami, and K. Tamada, J. Photochem. Photobiol. C 32, 58–77 (2017). https://doi.org/10.1016/j.jphotochemrev.2017.05.005, Google ScholarCrossref5. K. Okamoto, I. Niki, A. Shvartser, Y. Narukawa, T. Mukai, and A. Scherer, Nat. Mater. 3(9), 601–605 (2004). https://doi.org/10.1038/nmat1198, Google ScholarCrossref6. K. Okamoto and Y. Kawakami, IEEE J. Sel. Top. Quantum Electron. 15(4), 1199 (2009). https://doi.org/10.1109/JSTQE.2009.2021530, Google ScholarCrossref7. E. M. Purcell, H. C. Torrey, and R. V. Pound, Phys. Rev. 69(1–2), 37–38 (1946). https://doi.org/10.1103/PhysRev.69.37, Google ScholarCrossref, ISI8. M. Agio and D. M. Cano, Nat. Photonics 7(9), 674–675 (2013). https://doi.org/10.1038/nphoton.2013.219, Google ScholarCrossref9. D. M. Yeh, C. Y. Chen, Y. C. Lu, C. F. Huang, and C. C. Yang, Nanotechnology 18(26), 265402 (2007). https://doi.org/10.1088/0957-4484/18/26/265402, Google ScholarCrossref10. M. K. Kwon, J. Y. Kim, B. H. Kim, I. K. Park, C. Y. Cho, C. C. Byeon, and S. J. Park, Adv. Mater. 20(7), 1253–1257 (2008). https://doi.org/10.1002/adma.200701130, Google ScholarCrossref11. K. Kolwas and A. Derkachova, Opto-Electron. Rev. 18(4), 429–437 (2010). https://doi.org/10.2478/s11772-010-0043-6, Google ScholarCrossref12. H. Y. Ryu, Opt. Quantum Electron. 48(1), 6 (2016). https://doi.org/10.1007/s11082-015-0276-1, Google ScholarCrossref13. T. Karakouz, D. Holder, M. Goomanovsky, A. Vaskevich, and I. Rubinstein, Chem. Mater. 21(24), 5875–5885 (2009). https://doi.org/10.1021/cm902676d, Google ScholarCrossref14. F. Tam, G. P. Goodrich, B. R. Johnson, and N. J. Halas, Nano Lett. 7(2), 496–501 (2007). https://doi.org/10.1021/nl062901x, Google ScholarCrossref15. P. Mulvaney, Langmuir 12(3), 788–800 (1996). https://doi.org/10.1021/la9502711, Google ScholarCrossref16. A. V. Akimov, A. Mukherjee, C. L. Yu, D. E. Chang, A. S. Zibrov, P. R. Hemmer, H. Park, and M. D. Lukin, Nature 450(7168), 402–406 (2007). https://doi.org/10.1038/nature06230, Google ScholarCrossref17. C. Langhammer, M. Schwind, B. Kasemo, and I. Zorić, Nano Lett. 8(5), 1461–1471 (2008). https://doi.org/10.1021/nl080453i, Google ScholarCrossref, ISI18. L. W. Jang, D. W. Jeon, T. Sahoo, D. S. Jo, J. W. Ju, S. J. Lee, J. H. Baek, J. K. Yang, J. H. Song, A. Y. Polyakov, and I. H. Lee, Opt. Express 20(3), 2116–2123 (2012). https://doi.org/10.1364/OE.20.002116, Google ScholarCrossref19. R. Bardhan, N. K. Grady, and N. J. Halas, Small 4(10), 1716–1722 (2008). https://doi.org/10.1002/smll.200800405, Google ScholarCrossref20. W. Wang, Z. Li, B. Gu, Z. Zhang, and H. Xu, ACS Nano 3(11), 3493–3496 (2009). https://doi.org/10.1021/nn9009533, Google ScholarCrossref21. P. B. Johnson and R. W. Christy, Phys. Rev. B 6(12), 4370–4379 (1972). https://doi.org/10.1103/PhysRevB.6.4370, Google ScholarCrossref, ISI22. Y. Li, B. Liu, R. Zhang, Z. Xie, Z. Zhuang, J. Dai, T. Tao, T. Zhi, G. Zhang, P. Chen, F. Ren, H. Zhao, and Y. Zheng, J. Appl. Phys. 117, 153103 (2015). https://doi.org/10.1063/1.4918555, Google ScholarScitation, ISI23. H. Imadate, T. Mishima, and K. Shiojima, Jpn. J. Appl. Phys., Part 1 57(4S), 04FG13 (2018). https://doi.org/10.7567/JJAP.57.04FG13, Google ScholarCrossref24. M. D. Mcmahon, R. Lopez, H. M. Meyer III, L. C. Feldman, and R. F. Haglund, Jr., Appl. Phys. B 80, 915–921 (2005). https://doi.org/10.1007/s00340-005-1793-6, Google ScholarCrossref25. T. U. Kampen and W. Mönch, Metal Contacts on a-GaN ( Cambridge University Press, 2014). Google Scholar26. K. Saarinen, T. Laine, S. Kuisma, J. Nissilä, P. Hautojärvi, L. Dobrzynski, J. M. Baranowski, K. Pakula, R. Stepniewski, M. Wojdak, A. Ysmolek, T. Suski, M. Leszczynski, I. Grzegory, and S. Porowski, Observation of Native Ga Vacancies in Gan by Positron Annihilation ( Cambridge University Press, 2011). Google Scholar27. J. Neugebauera and C. G. van de Walle, Appl. Phys. Lett. 69, 503 (1996). https://doi.org/10.1063/1.117767, Google ScholarScitation28. J. L. Lyons and C. G. Van de Walle, npj Comput. Mater. 3, 12 (2017). https://doi.org/10.1038/s41524-017-0014-2, Google ScholarCrossref, ISI29. J. Bohlen, Á. Cuartero-González, E. Pibiri, D. Ruhlandt, A. I. Fernández-Domínguez, P. Tinnefeld, and G. P. Acuna, Nanoscale 11, 7674–7681 (2019). https://doi.org/10.1039/C9NR01204D, Google ScholarCrossref
留言 (0)