Semiconductor heterostructures represent the most important building block for current optoelectronic devices. One of the common features of semiconductor heterostructures is the existence of internal strain due to lattice mismatch. The internal strain can tilt the band alignment and significantly alter the physical properties of semiconductor heterostructures, such as reducing the internal quantum efficiency of a light emitter. Here, we provide a convenient route to release the internal strain by patterning semiconductor heterostructures into nanotip arrays. The fabrication of the nanotip arrays was achieved by self-masked dry etching technique, which is simple, low cost and compatible with current semiconductor technologies. By implementing our approach to InGaN/GaN multiple quantum wells, we demonstrate that the light emission can be enhanced by up to 10 times. Our approach renders an excellent opportunity to manipulate the internal strain, and is very useful to create highly efficient solid state emitters.
© 2007 Optical Society of America
It is well known that strain is able to significantly change the physical properties of a solid. Semiconductor heterostructure is the most important component used in current optoelectronic devices, which inevitably contains strain due to lattice mismatch of the constituent epilayers. One of the drawbacks of the existence of the internal strain is to tilt the band gap and reduce the radiative recombination efficiency. Let us take solid-state light emitting diodes (LEDs) as an example. LEDs with high external efficiency are currently in high demand for a variety of applications including flat panel displays, printers, optical interconnects in computers, and general lighting [1-2]. InGaN compound semiconductors have been attracting much attention as potential materials for the fabrication of LEDs [1-2] and laser diodes (LDs) [1-3] due to their tunability of the energy gap covering the visible to near ultraviolet light spectra. One of the distinct features of wurtzite nitride semiconductors is that they are good piezoelectric materials. Together with the existence of internal strain due to lattice mismatch, nitride heterostructures process large spontaneous and piezoelectric polarization, and therefore a large built-in internal electric field [4-8]. A strong built-in electric field will reduce the internal quantum efficiency (IQE) due to the less overlap of electron-and-hole wave function. Several groups have pointed out that the quantum confined Stark effect (QCSE) due to the piezoelectric (PZ) field can drastically alter the energy states and dominates the recombination mechanism [9-10]. Since the need to save energy resources has become an important issue in modern times, how to make highly efficient lighting is more and more requisite nowadays, which is also a major obstacle to be overcome for the application of nitride semiconductors in optoelectronic devices. Much effort has been exerted in improving LED quantum efficiency, by increasing IQE and light extraction efficiency; such as modifying spontaneous emission by resonant cavity, top surface rough process,[12-13] nonpolar GaN growth, two-dimensional (2D) photonic crystals (PCs) , nanoporous structures , nanocolumn , nanorod arrays , nanopillar and nanostripe arrays , and nanopost structures . However, these techniques do not provide a significant improvement of the luminescence, and most of them require complex pre-growth procedure. For example, GaN nanopillar and nanostripe arrays with embedded InGaN/GaN multiple quantum wells have been fabricated by holographic lithography and subsequent reactive ion etching . It was found that the fabrication process can seriously damage sample surface and lead to a significant decrease in the multiple quantum well luminescence. Even after the optimized annealing condition at 900 °C for 30 minutes, it can only slightly enhance the luminescence intensity in comparison to the planar wafers. Meanwhile, electron-beam lithography and inductively coupled plasma (ICP) reactive ion etching (RIE) have also been used for fabricating 10-40 nm InGaN/GaN nanoposts . However, it is well-known that the electron-beam lithography can only produce nanostructures with a rather limited area. Therefore, an efficient nanofabrication process over large area on varied substrates and compatible with existing process technologies is still needed. Besides, both of significant blue shifts and no measurable effect on the Stokes shift have been observed in the pillar relaxation processes.[19,20] This is quite a contradictory result, which needs a further investigation in order to establish a generalized understanding in nitride semiconductors.
Here, we demonstrate an alternative approach where the IQE can be enhanced considerably by a simple fabrication of nanotip arrays, which were achieved by one step and self-masked dry etching (SMDE) technique. It should be stressed that SMDE technique is simple, low cost, and compatible with current semiconductor technologies over a large area. It is found that integrated photoluminescence (PL) intensity of InGaN/GaN nanotips can be remarkably enhanced by 10 times compared with that of as-grown InGaN/GaN multiple quantum wells (MQWs). Based on our structural and optical studies, we clearly show that the internal strain of InGaN/GaN nanotips can be greatly reduced due to a large surface-to-volume ratio of nanotip. We firmly establish that in addition to the increase of light extraction efficiency, strain release is responsible for the enhancement of emission from InGaN/GaN nanotips. Our results not only evidence the possibility of the emission enhancement by the reduction of internal strain, but also provide a convenient way to manipulate the optical properties of a material by using nanoscale tip structures, which is very useful for the creation of highly efficient optoelectronic devices.
The samples studied in this work were prepared by low-pressure metalorganic chemical vapor deposition (MOCVD). A series of 10 periods of 20 Å thick In0.18Ga0.82N wells and 90 Å thick GaN barriers were grown on (0001) sapphire. The multiple-quantum-well layers were sandwiched with 6 µm n-GaN layer and 0.2 µm p-GaN capping layers, which is the typical structure for LEDs. The In content in the wells is estimated by comparing the PL position of the bulk materials grown in similar conditions. The fabrication of InGaN/GaN nanotip arrays was followed by a process named SMDE, which has been depicted schematically elsewhere [21-23]. Here, we only briefly describe the main procedures. Before fabricating the nanotip arrays, Au nanoparticles were coated on the top of InGaN/GaN MQWs by using a sputtering system (JFC-1600, JEOL) to protect the surface of as-grown sample. The diameter of the deposited Au nanoparticles is about 10 nm. The pre-cleaning of the substrate by hydrogen (H2) plasma, followed by simultaneous process of self-masking and reactive etching of the unmasked area for nanotip formation, was done in the reaction chamber in one step and without interrupting the vacuum . Reactive gases comprising of argon (Ar), H2, methane (CH4), and silane (SiH4) (10%, diluted in helium) with typical flow rates of 3 sccm, were activated by a microwave power of typically 1200 W at a chamber pressure of 5.8 mTorr, during the nanotip fabrication process. The gas-mixture consisting of SiH4, CH4, Ar, and H2 present in the electron cyclotron resonance (ECR) plasma is supposed to react and play two different roles: formation of SiC nanomasks and etching of the substrate to develop nanotips. It is believed that in the ECR plasma, SiC nanosized clusters are formed from the reaction of SiH4 and CH4 plasma and uniformly distribute themselves over the surface of as-grown InGaN/GaN MQWs. Meanwhile, Ar and H2 plasma are responsible for the physical etching and chemical etching, respectively. The SiC nanoclusters deposited on the surface of asgrown InGaN/GaN MQWs then act as nanomasks against etching, because of their higher hardness and chemical inertness. Arrays of aligned InGaN/GaN nanotips with a tip diameter typically in the range of ~2 nm, base diameter of 20 nm, and tip length of ~1 µm were obtained by the scanning electron microscope (SEM) image shown in Fig. 1. The SEM image was taken by using a field emission based SEM (JEOL JSM-6500F). Figure 1 depicts the tilted top-view SEM image of the InGaN/GaN nanotip arrays. We can clearly see that nanotips were well aligned and uniformly distributed over the entire surface with a high density. For the optical measurements, the photoluminescence (PL) spectra were performed with a SPEX 0.85 m monochromator. A cw He-Cd laser with a wavelength of 325 nm was used for PL excitation and a cooled GaAs photomultiplier tube interfaced with a lock-in amplifier was used as a detector. The µ-Raman scattering spectra were measured by a Jobin- Yvon T64000 system at room temperature using a He-Cd laser with a 325 nm in backscattering geometry. The incident and scattered light were parallel to the c axis, which in turn was normal to the growth surface.
3. Results and discussion
Figure 2 shows the PL spectra of the InGaN/GaN nanotips and as-grown InGaN/GaN MQWs at 13 K, which contain several distinct features. First, a large enhancement of the emission from the InGaN/GaN nanotips is observed. The integrated PL intensity of the InGaN/GaN nanotips is about 10 times stronger than that of as-grown InGaN/GaN MQWs. Second, the quality of the InGaN/GaN nanotips is further testified by the narrow bandwidth in the PL spectrum. The full width at half maximum (FWHM) of the spectra of InGaN/GaN nanotips and MQWs are 36 meV and 48 meV, respectively. Third, the PL peak of the InGaN/GaN nanotips shows a blueshift compared with that of the InGaN/GaN MQWs sample. Fourth, to further elucidate the unique behaviors of the InGaN/GaN nanotips, we have performed the power dependent PL spectra with optical excitation density ranging from 2.5x103 to 2x105 W/cm2 as shown in Fig. 3. We can clearly see that the main peak position of the PL spectra of the InGaN/GaN MQWs shows a blueshift as the optical excitation density increases, while that of the InGaN/GaN nanotips remains unchanged.
Let us now try to understand the above observed intriguing optical properties. Because the nanotip arrays were made from the MQWs sample, the only difference between both samples is their geometric structures. The underlying origin of the optical enhancement may therefore arise from this distinct feature. It is obvious that the enhanced PL intensity in InGaN/GaN nanotips can not be due to the reduction of nonradiative recombination centers, because the nanotip sample has a much larger surface-to-volume ratio than that of MQWs, which will generate a large number of surface states, and significantly quench the radiative recombination. Besides, the effect of quantum confinement can also be excluded in the interpretation of the above result. Because with an effective mass of about 0.2 me for InGaN layer, the effect of quantum confinement in a size of 15 nm, which is the diameter of the active InGaN/GaN layer, is negligible. In addition, it is very difficult to explain why the PL linewidth becomes narrower and the dependence of the PL spectra on excitation density is also changed.
Instead, the above results may be attributed to the strain relaxation in the nanotip through its inherent characteristic of a large surface-to-volume ratio. It means that the strain in the InGaN/GaN active region caused by the lattice mismatch between GaN and InGaN layers was greatly released after the formation of InGaN/GaN nanotip arrays. The band structure of the InGaN/GaN nanotips therefore becomes flatter, which will increase the wave function overlap of electrons and holes, and enhanced the radiative recombination rate. This provides the possibility to explain why the nanotip sample can have a higher PL intensity. It is worth noting here that the light extraction efficiency of nanotip arrays is better than that of MQWs.
According to previous reports [13,19], the light extraction efficiency through top surface rough process can only slightly increase the output emission intensity. Therefore, the large enhancement of the PL intensity can be mainly attributed to the strain relaxation.[19,20] Moreover, the reduction in the piezoelectric field caused by the strain relaxation can also induce the blueshift in the PL spectrum of the nanotip as shown in Fig. 1 due to the reduced tilting in the band alignment. Besides, the narrower linewidth of the nanotip PL spectrum can also be understood since a tilted energy band is known to broaden the energy distribution of energy states, and the range of indium composition fluctuations is reduced by the nanotips array . We are now ready to understand the difference in the power dependent PL spectra. In the typical case of the QCSE, such as in InGaN/GaN multiple quantum wells , the blueshift of PL spectra with increasing excitation density was frequently attributed to the screening of built-in PZ field. The underlying physics is based on the spatial separation of photoexcited electrons and holes due to the built-in piezoelectric (PZ) field. Because the induced field is opposite to the PZ field, the resultant field is reduced. Therefore, it is reasonable to infer that the blueshift in PL spectra with an increase in excitation density shown in Fig. 3(a) arising from the screening of polarization electric field in strained InGaN/GaN MQWs. As a consequence, the no apparent change of the PL peak position of the InGaN/GaN nanotips with different excitation densities can be attributed to strain relaxation. We therefore can see that based on the strain relaxation in InGaN/GaN nanotips all the distinct features in the PL spectra can be well understood. It is worth pointing out here that the interference pattern observed in the InGaN/GaN MQWs disappears in the nanotip arrays. This is due to the fact that the emission in the nanotip arrays now not only comes out from the top surface, but also it can come out through the side wall.
In order to further examine the possibility that the strain relief is indeed responsible for the enhanced optical properties, we have performed Raman scattering measurements as shown in Fig. 4. In Fig. 4(a), we show the room-temperature Raman spectra recorded in backscattering configuration of the InGaN/GaN nanotips and as-grown InGaN/GaN MQWs. There are two resolved phonon structures in each Raman spectrum. The 569 cm-1 peak is from E2 mode of the strained h-GaN [25-26], and the shoulder line located near 560 cm-1 is the InGaN E2 mode . The feature in the Raman scattering spectrum of the InGaN/GaN nanotips is much stronger than that of as-grown MQWs. Besides, the E2 phonon mode shows a clear low-frequency shift in the peak position of the InGaN/GaN nanotips compared with that of the InGaN/GaN MQWs. There are two possible mechanisms that can induce the phonon peak shift in our study: (i) size effects; (ii) strain relaxation. In the first mechanism, in general consideration for Raman scattering of one dimensional nano-sized system, the downshifts of the Raman peaks are mainly attributed to the size-confinement effect . When the crystalline size decreases, momentum conservation will be relaxed and Raman active modes will not be limited to be at the center of the Brillouin zone. The smaller the crystalline grain is, the bigger the frequency shifts and the more asymmetric the line shape. In the second mechanism, the shift to lower frequency in Raman spectra can be understood by the effect of the reduced compressive strain in the nanotip arrays. To obtain the exact mechanism, we have performed the Raman scattering spectra with different excitation densities as shown in Figs. 4(b) and 4(c). Interestingly, it is seen that the E2 phonon of the InGaN MQWs decreases in energy with increasing optical excitation density while that of the InGaN nanotips shows no apparent change. This redshift phenomenon in the E2 phonon of InGaN MQWs with increasing excitation density has been attributed to the effect of the reduced strain through converse PZ effect due to the screening of internal electric field by laser pumping.  The unchanged E2 phonon with increasing excitation density in the InGaN nanotips therefore evidences the strain relaxation.
Finally, according to the expression given below, one is able to estimate the magnitude of the strain relaxation based on the redshift of the E2 phonon mode. To estimate the magnitude of the strain under different pumping density, we use the following equation .
where a and b are phonon deformation potentials, and C13 and C33 are elastic constants, respectively. Figure 5 shows the strain as a function of excitation density according to the calculation of Eq. (1), where the deformation potentials a, b and the elastic constants C13, C33 were estimated by the extrapolation of the data for GaN and InN . The value of the calculated strain is in the same order of magnitude reported previously . Evidently, the strain of as-grown InGaN/GaN MQWs decreases with increasing excitation density as expected. We therefore can see that both of the zero shifts in PL spectra and Raman scattering of the InGaN/GaN nanotips spectra can be well explained in a consistent way based on the strain relaxation. In view of the change of the strain on the optical properties of InGaN/GaN nanotip arrays, our results shown here should be very useful for the application in optoelectronic based devices.
In conclusion, we have demonstrated a significant improvement of the emission from InGaN/GaN nanotip arrays compared with InGaN/GaN MQWs. The nanotip arrays were formed by a simple and low cost self-masked dry etching process, which is compatible with the current semiconductor technologies. Our unique approach is able to enhance the light output power by a factor of up to 10 times. Based on our study, we clearly demonstrate that the main underlying mechanism for the enhanced luminescence arises from the strain relaxation in the nanotip through its inherent characteristic of a large surface-to-volume ratio. In view of the existence of strain in most semiconductor heterostructures, our results should be very important for manufacturing a wide variety of highly efficient optoelectronic devices.
This work was supported by the Ministry of Education and the National Science Council of the Republic of China.
References and links
1. S. Nakamura and G. Fasol, The Blue Laser Diode (Springer, Berlin, 1997)
2. S. Nakamura, M. Senoh, N. Iwasa, and S. Nagahama, “High-Brightness InGaN Blue, Green and Yellow Light-Emitting Diodes with Quantum Well Structures,” Jpn. J. Appl. Phys. 34, L797–L799 (1995) [CrossRef]
3. S. Nakamura, M. Senoh, S. Nagahama, N. Iwasa, T. Yamada, T. Matsusshita, H. Kiyoko, and Y. Sugimoto, “InGaN-Based Multi-Quantum-Well-Structure Laser Diodes,” Jpn. J. Appl. Phys. 35, L74–L76 (1996) [CrossRef]
4. S. Nakamura, “The Roles of Structural Imperfections in InGaN-Based Blue Light-Emitting Diodes and Laser Diodes,” Science 281, 956–961 (1998) [CrossRef]
5. T. Takeuchi, S. Sota, H. Sakai, H. Amanoa, I. Akasaki, Y. Kaneko, S. Nakagawa, Y. Yamaoka, and N. Yamada, “High-power UV InGaN/AlGaN double-heterostructure LEDs,” J. Cryst. Growth 189/190, 778–781 (1998) [CrossRef]
6. P. Riblet, H. Hirayama, A. Kinoshita, A. Hirata, T. Sugano, and Y. Aoyagi, “Determination of photoluminescence mechanism in InGaN quantum wells,” Appl. Phys. Lett. 75, 2241–2243 (1999) [CrossRef]
7. T. Mukai, M. Yamada, and S. Nakamura, “Current and Temperature Dependences of Electroluminescence of InGaN-Based UV/Blue/Green Light-Emitting Diodes,” Jpn. J. Appl. Phys. 37, L1358–L1361 (1998) [CrossRef]
8. H. C. Yang, P. F. Kuo, T. Y. Lin, Y. F. Chen, K. H. Chen, L. C. Chen, and J. -I. Chyi, “Mechanism of luminescence in InGaN/GaN multiple quantum wells,” Appl. Phys. Lett. 76, 3712–3714 (2000) [CrossRef]
9. T. Takeuchi, S. Sota, M. Katsuragawa, M. Komori, H. Takeuchi, H. Amano, and I. Akasaki, “Quantum-Confined Stark Effect due to Piezoelectric Fields in GaInN Strained Quantum Wells,” Jpn. J. Appl. Phys., Part 2 , 36, L382–L385 (1997) [CrossRef]
10. T. Wang, D. Nakagawa, J. Wang, T. Sugahara, and S. Sakai, “Photoluminescence investigation of InGaN/GaN single quantum well and multiple quantum wells,” Appl. Phys. Lett. 73, 3571–3573 (1998) [CrossRef]
11. E. F. Schubert, Y.-H. Wang, A. Y. Cho, L.-W. Tu, and G. J. Zydzik, “Resonant cavity light-emitting diodes,” Appl. Phys. Lett. 60, 921–923 (1992) [CrossRef]
12. C. Huh, K. S. Lee, E. J. Kang, and S. J. Park, “Improved light-output and electrical performance of InGaN based light-emitting diode by microroughening of the p-GaN surface,” J. Appl. Phys. 93, 9383–9385 (2003) [CrossRef]
13. T. Fujii, Y. Gao, R. Sharma, E. L. Hu, S. P. DenBaars, and S. Nakamura, “Increase in the extraction efficiency of GaN-based light-emitting diodes via surface roughening,” Appl. Phys. Lett. 84, 855–857 (2004) [CrossRef]
14. E. Kuokstis, C. Q. Chen, J. W. Yang, M. Shatalov, M. E. Gaevski, V. Adivarahan, and M. A. Khan, “Room-temperature optically pumped laser emission from a-plane GaN with high optical gain characteristics,” Appl. Phys. Lett. 84, 2998–3000 (2004) [CrossRef]
15. T. N. Oder, H. S. Kim, J. Y. Lin, and H. X. Jiang, “III-nitride blue and ultraviolet photonic crystal light emitting diodes,” Appl. Phys. Lett. 84, 466–468 (2004) [CrossRef]
16. C. F. Lin, J. H. Zheng, Z. J. Yang, J. J. Dai, D. Y. Lin, C. Y. Chang, Z. X. Lai, and C. S. Hong, “High efficiency InGaN-based light-emitting diodes with nanoporous GaN:Mg structure,” Appl. Phys. Lett. 88, 083121 (2006) [CrossRef]
17. A. Kikuchi, M. Kawai, M. Tada, and K. Kishino, “InGaN/GaN multiple quantum disk nanocolumn light-emitting diodes grown on (111) Si substrate,” Jpn. J. Appl. Phys. 43L1524–L1526 (2004) [CrossRef]
18. Y. Sun, Y.-H. Cho, H.-M. Kim, and T. W. Kang, “High efficiency and brightness of blue light emission from dislocation-free InGaN/GaN quantum well nanorod arrays,” Appl. Phys. Lett. 87, 093115 (2005) [CrossRef]
19. S. Keller, C. Schaake, N. A. Fichtenbaum, C. J. Neufeld, Y. Wu, K. McGroddy, A. David, S. P. DenBaars, C. Weisbuch, J. S. Speck, and U. K. Mishra, “Optical and structural properties of GaN nanopillar and nanostripe arrays with embedded InGaN/GaN multi-quantum wells,” J. Appl. Phys. 100, 054314 (2006) [CrossRef]
20. H. S. Chen, D. M. Yeh, Y. C. Lu, C. Y. Chen, C. F. Huang, T. Y. Tang, C. C. Yang, C. S. Wu, and Chii-Dong Chen, “Strain Relaxation and Quantum Confinement in InGaN/GaN Nano-posts,” Nanotechnology , 17, 1454–1458 (2006) [CrossRef]
21. C. H. Hsu, H. C. Lo, C. F. Chen, C. T. Wu, J. S. Hwang, D. Das, J. Tsai, L. C. Chen, and K. H. Chen, “Generally Applicable Self-Masked Dry Etching Technique for Nanotip Array Fabrication,” Nano Letters, 4, 471–475 (2004) [CrossRef]
22. J. C. She, K. Zhao, S. Z. Deng, J. Chen, and N. S. Xu, “Field electron emission of Si nanotips with apexes of various compositions,” Appl. Phys. Lett. 87, 052105 (2005) [CrossRef]
23. S. C. Shi, C. F. Chen, Surojit Chattopadhyay, K. H. Chen, B. W. Ke, L. C. Chen, L. Trinkler, and B. Berzina, “Luminescence properties of wurtzite AlN nanotips,” Appl. Phys. Lett. 89, 163127 (2006) [CrossRef]
24. F. Della Sala, A. Di Carlo, P. Lugli, F. Bernardini, V. Fiorentini, R. Scholz, and J.-M. Jancu, “Free-carrier screening of polarization fields in wurtzite GaN/InGaN laser structures,” Appl. Phys. Lett. 74, 2002–2004 (1999) [CrossRef]
25. C. H. Chen, W. H. Chen, Y. F. Chen, and T. Y. Lin, “Piezoelectric, electro-optical, and photoelastic effects in InGaN/GaN multiple quantum wells,” Appl. Phys. Lett. 83, 1770–1772 (2003) [CrossRef]
26. A. G. Kontos, Y. S. Raptis, N. T. Pelekanos, A. Georgakilas, E. Bellet-Amalric, and D. Jalabert, “Micro-Raman characterization of InGaN/GaN/Al2O3 heterostructures,” Phys. Rev. B 72, 155336 (2005) [CrossRef]
27. C. H. Liang, L. C. Chen, J. S. Hwang, K. H. Chen, Y. T. Hung, and Y. F. Chen, “Selective-area growth of indium nitride nanowires on gold-patterned Si(100) substrates,” Appl. Phys. Lett. 81, 22–24 (2002) [CrossRef]
28. T. Y. Lin, “Converse piezoelectric effect and photoelastic effect in InGaN/GaN multiple quantum wells,” Appl. Phys. Lett. 82, 880–882 (2003) [CrossRef]
29. V. Darakchieva, P. P. Paskov, E. Valcheva, T. Paskova, B. Monemar, M. Schubert, H. Lu, and W. J. Schaff, “Deformation potentials of the E1(TO) and E2 modes of InN,” Appl. Phys. Lett. 84, 3636–3638 (2004) [CrossRef]