Letter pubs.acs.org/NanoLett
Efficient Spin-Light Emitting Diodes Based on InGaN/GaN Quantum Disks at Room Temperature: A New Self-Polarized Paradigm J. Y. Chen,† C. Y. Ho,† M. L. Lu,† L. J. Chu,† K. C. Chen,† S. W. Chu,† W. Chen,‡ C. Y. Mou,‡ and Y. F. Chen*,† †
Department of Physics, National Taiwan University, Taipei 106, Taiwan Department of Chemistry, National Taiwan University, Taipei 106, Taiwan
‡
S Supporting Information *
ABSTRACT: A well-behaved spin-light emitting diode (LED) composed of InGaN/GaN multiple quantum disks (MQDs), ferromagnetic contact, and Fe3O4 nanoparticles has been designed, fabricated, and characterized. The degree of circular polarization of electroluminescence (EL) can reach up to a high value of 10.9% at room temperature in a low magnetic field of 0.35 T, which overcomes a very low degree of spin polarization in nitride semiconductors due to the weak spin−orbit interaction. Several underlying mechanisms play significant roles simultaneously in this newly designed device for the achievement of such a high performance. Most of all, the vacancy between nanodisks can be filled by halfmetal nanoparticles with suitable energy band alignment, which enables selective transfer of spin polarized electrons and holes and leads to the enhanced output spin polarization of LED. Unlike previously reported mechanisms, this new process leads to a weak dependence of spin relaxation on temperature. Additionally, the internal strain in planar InGaN/GaN multiple quantum wells can be relaxed in the nanodisk formation process, which leads to the disappearance of Rashba Hamiltonian and enhances the spin relaxation time. Our approach therefore opens up a new route for the further research and development of semiconductor spintronics. KEYWORDS: Spin-light emitting diode, nitride semiconductor, half-metal, quantum disk, selective spin charge transfer, spin relaxation time
S
GaN (wurtzite structure) has a smaller optical polarization than GaAs (zinc blende structure) due to the weak spin−orbit interaction, which limits the optical polarization of GaN to about 3% even if electron spin is completely polarized.22 Moreover, optical experiments performed on GaN/InGaN quantum wells show that the electron spin relaxation time in these structures is shorter than that in bulk GaN and GaAs/ InGaAs QWs.23 The reason arises from the fact that, in InGaN/ GaN MQWs, there is a strong internal strain field caused by the large lattice mismatch of 11% between GaN and InN.24,25 This built-in piezofield breaks the reflection symmetry of confining potential, which leads to the presence of a large Rashba term in the conduction band Hamiltonian, and reduces the spin relaxation time. In stark contrast, the confinement potential in InGaAs/GaAs MQWs has reflection symmetry, and the Rashba spin−orbit interaction is therefore weak.26 In this letter, a new approach is proposed to overcome the above difficulty for nitride semiconductors by introducing nanodisk structures. We demonstrate that the internal strain in planar InGaN/GaN MQWs can be relaxed in the nanodisk formation process, which results in the disappearance of the
pintronics has been extensively investigated since the discovery of giant magnetoresistance (GMR) due to its academic and practical importances.1−3 Magnetoresistive devices based on harness electron spin are now commercially available, such as memory cells and sensors. In spintronics, several spin-related devices have been reported4−6 in which a spin-light emitting diode (LED) is one of the most intriguing devices. Spin-LED is basically the combination of a spin injector and an emissive spin absorber. It emits a circularly polarized light, which is very useful for applications such as the manipulation of spin state in quantum information,7 chiral synthesis in biology,8 circular dichroism spectroscopy,9 ultrafast magnetization control,10 and cryptography of optical communication.11 Even though a high degree of polarization of spinLED has been achieved in GaAs multiple quantum wells (MQWs),8,12−15 most of these devices need a high magnetic field and low working temperature to obtain an appreciable output spin polarization. Moreover, the band gap of GaAs is 1.35 eV, which can only provide infrared or red LEDs.16 Up to now, a blue spin-LED with high degree of spin polarization has not been achieved. In light of great brightness of InGaN/GaN MQW LEDs for commercial solid state lighting and backlight modules of flat panel and GaN-based field effect transistors, the extension of nitride based semiconductors to spintronic devices has triggered substantial efforts.17−21 However, it is known that © XXXX American Chemical Society
Received: January 28, 2014 Revised: May 1, 2014
A
dx.doi.org/10.1021/nl5003312 | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 1. (a) Schematic structure of InGaN/GaN MQD LED. Scanning electron microscope images of (b) the cross section and (c) the top view of InGaN/GaN MQD nanorods. (d) Morphology of the Fe3O4 nanoparticles was taken by transmission electron microscopy. (e) Photograph of light emission from InGaN/GaN MQD LED.
holes from nanodisks to nanoparticles much easier. Thus, it increases the population difference of heavy holes and light holes in MQDs, where electroluminescence (EL) is emitted. The other advantage for the structure of nanodisks is that it makes the separation between the ferromagnetic metal contact and the luminescent layer shorter than that of its planar counterpart, i.e., the bottom contact can be deposited just around the nanorod structures. As a result, the effect of spin injection arising from ferromagnetic electrodes can be greatly improved. Consequently, with the combination of spin injection from ferromagnetic electrodes, the effect of Fe3O4 magnetic nanoparticles, and the strain relaxation in MQDs, we are able to achieve a high value of EL spin polarization up to 10.9% at room temperature in a low magnetic field of 0.35 T based on nitride semiconductors. The schematic structure of the studied InGaN/GaN MQD LED is shown in Figure 1a. The sample was prepared by metal−organic chemical vapor deposition (MOCVD). A series of MQWs containing 10 periods of 2 nm In0.22Ga0.78N wells
quantum confined Stark effect (QCSE) and the Rashba Hamiltonian. In addition, the empty space between nanodisks can be filled with magnetic nanoparticles, such as half-metal Fe3O4 particles. These magnetic nanoparticles possess a different band alignment for electrons with spin-up and spindown, which enables the selective transfer of spin polarized electrons between nanodisks and nanoparticles. It is worth mentioning that half-metals have a novel property where they can act as a conductor for electrons with one spin orientation while acting as an insulator or semiconductor for electrons with the opposite orientation.27 We therefore can fully utilize the unique energy band structure for spin-up and spin-down electrons in Fe3O4 and its negative spin polarization characteristic to selectively increase the number of electrons with the preferred spin orientation in multiple quantum disks (MQDs).28 In addition, because the effective mass of heavy holes is much larger than that of light holes perpendicular to c axis in wurtzite structure, the magnetic nanoparticles with suitable energy band alignment can cause the transfer of light B
dx.doi.org/10.1021/nl5003312 | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
integrated intensities of the most dominant peak for σ+ and σ− circular polarization. In the beginning, we demonstrate the PL spectra measured by a nonpolarized 325 nm laser with 2.2 mW under a magnetic field of 0.35 T as shown in Figure 3a,b. Quite interestingly, it exhibits a large degree of circular polarization of 13.7% for the band gap emission, while the defect emission does not show any signature of spin polarization. The behavior of defect emission can be understood well by the spin flip scattering of the inherent nature of defect characteristics.31−33 Note that the PL spectra for the sample without Fe3O4 nanoparticles exhibit a negligible circular polarization for both band gap and defect emissions. This behavior provides a clear evidence showing that the measured circular polarization is not caused by the polarization effects of our measurement system such as the monochromator grating, mirrors, and the excitation source. In addition, the handedness of the circular polarization changes with magnetic field reversal ensures that the polarization arises from the population imbalance of the spin-up and spin-down electrons in the MQDs. In order to explain the above observation, let us consider the band alignment as shown in Figure 4a,b. First, we discuss the behavior of electrons. It is known that based on the unique band structure of Fe3O4, for spin-down electrons, Fe3O4 acts as a conductor, while for spin-up electrons, it acts as an insulator.28,34 Therefore, according to the band alignment between Fe3O4 nanoparticles and InGaN/GaN MQDs,28,34,35 when there are extra electrons in InGaN/GaN MQDs due to the injected current or light excitation source, spin-down electrons can pass through a monolayer of oleylamine and flow into Fe3O4 nanoparticles, while spin-up electrons remain in InGaN/GaN MQDs. As a result, Fe3O4 nanoparticles can be used to increase the percentage of spin-up electrons in MQDs and enhance the degree of output of the circular polarization of MQDs. Next, we discuss the behavior of holes. Because the effective mass of heavy holes is much larger than that of light holes perpendicular to the nanorods, light holes can transfer from nanodisks to nanoparticles much easier than heavy holes, which increase the population of heavy holes in nanodisks. The above unique features therefore break the theoretical limit of 3% for the circular polarization of light emission in nitride semiconductors.22 Moreover, we have performed the PL spectra excited by a circularly polarized laser as shown in Figure 4c, which reveals a large degree of circular polarization. This result demonstrates that there are more spin-up electrons than spin-down electrons remaining in the conduction band under the excitation of σ+ and σ− circularly polarized laser, respectively, and they can generate radiative recombination with defects and holes in the valence band. Because the PL intensities of pure InGaN/GaN MQDs excited by σ+ and σ− circularly polarized light source are the same, the above result implies that the spin-down electrons can transfer from InGaN/GaN MQDs to Fe3O4 nanoparticles, which is consistent with our proposed mechanism. To further support our interpretation, we have performed the time-resolved photoluminescence (TRPL) spectra of InGaN/GaN MQD nanorods filled with Fe3O4 nanoparticles at room temperature under a magnetic field of +0.35 T excited by a pulse laser with 260 nm. Generally, the spin lifetime can be analyzed by the excitation of a polarized light source (σ+ or σ−) and comparing σ+ and σ− circularly polarized TRPL spectra. The intensities of the TRPL spectra reveal a large degree of circular polarization. However, for the
and 9 nm GaN barriers were grown on a (0001) sapphire substrate. The MQWs were sandwiched between a 2 μm nGaN layer on the sapphire and a 0.5 μm p-GaN capping layer, which is a typical structure for nitride LEDs. The sample was then vertically etched into nanorods by the method of inductively coupled plasma reactive-ion etching (ICP-RIE).29 After the etching process, the InGaN/GaN MQWs will turn into MQD nanorods, in which the InGaN MQDs were embedded. The side view and top view of the sample were recorded by a scanning electron microscope (SEM) on a JSM6500F produced by JEOL as shown in Figure 1b,c, respectively. The nanorods were estimated to have an average diameter of 300 nm and an average length of 1.2 μm. The InGaN/GaN MQD nanorods were deposited with a 3.5 nm Ni film by evaporation in a base pressure lower than 5 × 10−7 Torr. The Ni film serves as the metal contact for hole and electron injection layers. The vacancies between nanorods were filled with Fe3O4 nanoparticles and were followed by spin coating of photoresist to avoid short circuit. A detailed synthesis process for the Fe3O4 nanoparticles can be found elsewhere.30 The morphology of the Fe3O4 nanoparticles was taken by transmission electron microscopy (TEM) on a JEOL JSM1200 EX II operating at 100 kV as shown in Figure 1d, and the diameter of the Fe3O4 nanoparticle is about 10 nm. The topmost photoresist was removed by acetone. Finally, the Ni/ Au (100 nm/100 nm) metal layers were deposited as electric contacts for p-type and n-type GaN. A bright emission arising from the fabricated InGaN/GaN MQD LED can be seen by the photograph as shown in Figure 1e. The magnetization hysteresis curve of Fe3O4 nanoparticles at room temperature is shown in Figure 2, which reveals a pronounced superparamagnetic characteristic with the saturation of the magnetization at around 0.35 T.
Figure 2. Hysteresis curve of Fe3O4 nanoparticles.
In this work, the EL and photoluminescence (PL) spectra were recorded by a SPEX 0.85 m monochromator and a cooled GaAs photomultiplier tube. A cw He−Cd laser with a wavelength of 325 nm was used as the excitation source for the measurement of PL spectra. The magnetic field was provided by a magnet around 0.35 T. We define the +z direction to be parallel to the direction of light propagation and the magnetic field, and perpendicular to the sample plane, i.e., we measure the EL and PL in the Faraday geometry. The degree of EL and PL polarization is defined by the following equation, P = (Iσ+ − Iσ−)/(Iσ+ + Iσ−), where Iσ+ (right circular polarization) and Iσ− (left circular polarization) are the C
dx.doi.org/10.1021/nl5003312 | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 3. (a,b) Photoluminescence spectra of InGaN/GaN MQD nanorods filled with Fe3O4 nanoparticles under magnetic field along positive and negative z direction, respectively.
Figure 4. Band diagram of (a) spin-down and (b) spin-up electrons in InGaN and Fe3O4. (c) Photoluminescence spectra of InGaN/GaN MQD nanorods filled with Fe3O4 nanoparticles excited by a circular polarized laser under the magnetic field along the positive z direction. (d) Timeresolved photoluminescence spectra of InGaN/GaN MQD nanorods filled with Fe3O4 nanoparticles under the magnetic field along the positive z direction.
normalized TRPL spectra under σ− laser excitation, the lifetime of σ+ and σ− luminescence are very similar as shown in Figure 4d. This result reveals the fact that the spin lifetime is much longer than the PL lifetime even at room temperature. Moreover, the lifetime of spin-down electrons excited by σ+ light source (120 ps) is much shorter than the lifetime of spinup electrons excited by σ− pulse laser (230 ps). This result implies that the spin-down electrons experience a very different
environment when InGaN/GaN MQD nanorods are filled with Fe3O4 nanoparticles, while the spin-up electrons are unaffected. Let us now examine the most important investigation of our current work for the EL characteristics of our newly designed LED. Figure 5a shows the I−V curve of InGaN/GaN MQD nanorods with Fe3O4 nanoparticles without an external magnetic field, which is almost identical with that of the sample without Fe3O4 nanoparticles. It shows a well behaved D
dx.doi.org/10.1021/nl5003312 | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
Figure 5. (a) I−V curve of InGaN/GaN MQD nanorods with Fe3O4 nanoparticles. Electroluminescence spectra of Ni/Au contact−InGaN/GaN MQD nanorods filled with Fe3O4 nanoparticles under magnetic field along (b) positive and (c) negative z directions and (d) zero magnetic field. (e) Degree of EL circular polarization under different magnetic fields. (f) Degree of EL circular polarization versus temperature of Ni/Au contact− InGaN/GaN MQD nanorods filled with Fe3O4 nanoparticles under a magnetic field of 0.35 T along the positive z direction.
the polarization arises from spin injection. Figure 5e shows the corresponding degrees of the EL circular polarizations under different magnetic fields. Two phenomena were found, one is that the degree of EL circular polarization increases when the external magnetic field increases, and another is that the rising trend saturates when the magnetic field exceeds 0.35 T. Quite surprisingly, as shown in Figure 5f, it is found that the difference between the degrees of EL circular polarization at 350 and 100 K under an applied current of 50 mA and a magnetic field of 0.35 T does not exceed 0.5%. It is well known that the relevant spin relaxation mechanisms, such as Bir− Aronov−Picus (BAP) mechanism due to electron−hole exchange interaction and Dyakonov−Perel (DP) mechanism due to the lack of inversion symmetry, do strongly depend on temperature.38 Therefore, this result reveals that a new mechanism caused by Fe3O4 nanoparticles is responsible for the spin relaxation and the EL circular polarization observed
diode characteristic and a negligible leakage current, and the deposition of Fe3O4 nanoparticles does not affect the I−V characteristics of the LED. The I−V curves of InGaN/GaN MQD nanorods with Fe3O4 nanoparticles under zero and 0.35 T of magnetic field in logarithmic scale can be found in Figure S1 in the Supporting Information. Figure 5b−d shows the EL spectra of Ni/Au contact−InGaN/GaN MQD filled with Fe3O4 nanoparticles under positive and negative z direction, zero magnetic fields, and an applied current of 50 mA, and the corresponding degrees of the EL circular polarizations are −10.9%, +10.9%, and 0%, respectively. The degree of EL circular polarization observed here at room temperature in a low magnetic field of 0.35 T far exceeds all of the previous reports based on nitride semiconductors.23,36,37 The disappearance of circular polarization without an external magnetic field can also be used to rule out the systematic errors of our experimental apparatus. Again, the handedness of the circular polarization changes with magnetic field reversal ensures that E
dx.doi.org/10.1021/nl5003312 | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
To figure out the influence of Fe3O4 nanoparticles on the degree of EL circular polarization of InGaN/GaN MQDs, we first consider the magnetization hysteresis curve of Fe3O4 nanoparticles as shown in Figure 2. The saturation of the magnetization at around 0.35 T is approximately the same as that of the degree of EL circular polarization shown in Figure 5e, indicating the important role played by the Fe 3 O4 nanoparticles. Second, the I−V characteristic of InGaN/GaN MQD nanorods filled with Fe3O4 nanoparticles reveals that the current flow in MQDs does not pass through Fe 3 O 4 nanoparticles because the deposition of Fe3O4 nanoparticle does not affect the I−V curve as shown in Figure 5a. Therefore, the observed spin polarization of EL does not result from the influence of the spin injection through the characteristic of the negative spin polarization of Fe3O4 nanoparticles. With the result of EL shown in Figure S2c in the Supporting Information, it is certain that the Fe3O4 nanoparticles can increase the percentage of spin-up electrons in MQDs due to the unique band alignment as shown in Figure 4. Additionally, it is also supported by the fact that the large circular polarization of PL spectra shown in Figure 3 arises from the transfer of spin-down electrons from MQDs to Fe3 O 4 nanoparticles. It is stressed here that the spin injection from Fe3O4 nanoparticles is not the underlying reason responsible for the output spin polarization of PL spectra since there is no injection current. In addition, the above mechanism can also be used to interpret the intriguing fact that the degree of EL circular polarization is not very sensitive to the change of temperature as shown in Figure 5f. It is because the increase of spin-down electrons due to thermal scattering can easily flow into the Fe3O4 nanoparticles. This unique process can suppress the effect of the relaxation of the electrons with spin-up and retain the high degree of EL circular polarization. Furthermore, we have also measured the dependence of the polarization of EL and PL spectra on applied current and excitation power (see Figure S4 in the Supporting Information). It shows that when the applied current or excitation power increases, the degree of circular polarization of EL or PL increases, which is in stark contrast with the previous report.12 The obtained result does support our proposed mechanism again because the increased spin-down electrons due to the applied current or excitation power can easily flow into Fe3O4 nanoparticles. It is worth noting that even if the Fe3O4 grain boundaries may reverse the spin orientation of the polarized electrons, it will not alter the population of spin-up and spin-down electrons in InGaN/GaN MQDs, which is responsible for the light emission. For instance, before spin-down electrons transfer from InGaN/ GaN MQDs into Fe3O4 nanoparticles, the interface may disturb and reverse the spin orientation. If this process occurs, it will enhance the population of spin-up electrons in the luminescent layer. However, if the interface reverses the spin orientation of spin-up electrons, it will increase the possibility of the transfer of spin-down electrons into the Fe3O4 nanoparticles. Therefore, the interface between InGaN/GaN MQDs and Fe3O4 nanoparticles does not play a decisive role in the degree of EL circular polarization. This phenomenon is also very different from the conventional devices based on ferromagnetic electrodes in which the interface can greatly degrade the circular polarization of spin-LED. In conclusion, we have demonstrated that efficient spinpolarized LEDs can be achieved at room temperature in a low magnetic field based on the composites consisting of InGaN/
here, which dissimilates the usual mechanisms arising from ferromagnetic spin injectors.12 Furthermore, to explore the underlying mechanism, the EL spectra under an applied current of 50 mA in a magnetic field of 0.35 T at room temperature for the sample with different structures and composites have been performed (see Figure S2 in the Supporting Information). Figure S2a in the Supporting Information displays that the degree of EL circular polarization for planar InGaN/GaN MQWs is negligible. It indicates that the effect of spin injection through the Ni contact is not efficient for planar InGaN/GaN MQWs, which is similar to the result obtained previously.23 Figure S2b in the Supporting Information shows that the degree of EL circular polarization is around −1.3% for pure InGaN/GaN MQDs with ferromagnetic Ni spin injectors, which is much better than that of the planar InGaN/GaN MQWs. Figure S2c in the Supporting Information demonstrates that the degree of EL circular polarization is around −6.5% for InGaN/GaN MQDs without ferromagnetic Ni spin injectors, but the empty space between nanorods was filled with Fe3O4 nanoparticles. Figure S2d in the Supporting Information indicates that the degree of EL circular polarization is around −10.9% for InGaN/GaN MQDs with Ni spin injectors and Fe3O4 nanoparticles. The above results show that both ferromagnetic Ni spin injectors and Fe3O4 nanoparticles do play a significant role for the detected circular polarization arising from InGaN/GaN MQDs. To clarify the exact underlying mechanism of our observation, we first have to exclude the possibility that the magnetic circular dichroism both in the Ni electrodes and Fe3O4 nanoparticles are responsible for the measured polarized EL and PL signal. As shown in Figure S2a in the Supporting Information, the degree of EL circular polarization for planar InGaN/GaN MQDs with Ni contact is negligible, which implies that the magnetic circular dichroism of the Ni contact is not responsible for the detected EL circular polarization. In addition, as shown in Figure 3a, there is no circular polarization in the defect emission for the InGaN/GaN MQD nanorods filled with Fe3O4 nanoparticles. It provides an excellent evidence showing that the possibility of the magnetic circular dichroism arising from Fe3O4 nanoparticles can be ruled out in the interpretation of the output circular polarization. We now try to understand the origin behind the different degrees of EL circular polarization between InGaN/GaN MQWs and MQDs. According to the previous reports, this behavior may be attributed to the internal strain relaxation in MQDs, which can eliminate the reduction of spin relaxation time caused by the Rashba Hamiltonian.23,39 The relaxation of the internal strain in planar InGaN/GaN MQW structure in MQDs is demonstrated by the comparison between the EL spectra of InGaN/GaN MQWs and MQDs under different applied currents (see Figure S3 in the Supporting Information). Referring to Figure S3a in the Supporting Information, we note that a blue shift of the normalized EL spectra from 453 to 448 nm takes place on the planar LED, as the injection current increases from 40 to 60 mA. This blue shift can be mainly ascribed to the screening of QCSE caused by the internal strain as the injection current increases.40 For the case of InGaN/ GaN MQDs, the blue shift of the normalized EL spectra is negligible as shown in Figure S3b in the Supporting Information. Therefore, compared with the significant and negligible blue shift in MQWs and MQDs, it provides a strong evidence for the strain relaxation in MQDs structure. F
dx.doi.org/10.1021/nl5003312 | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
(7) Jonker, B. IEEE Proc. 2003, 91, 727. (8) Adelmann, C.; Lou, X.; Strand, J.; Palmstrøm, C. J.; Crowell, P. A. Phys. Rev. B 2005, 71, 121301. (9) Barron, L. Molecular Light Scattering and Optical Activity; Cambridge University Press: Cambridge, U.K., 2004. (10) Awschalom, D. D.; Warnock, J.; von Molńar, S. Phys. Rev. Lett. 1987, 58, 812. (11) Holub, M.; Bhattacharya, P. J. Phys. D: Appl. Phys. 2007, 40, 179. (12) Hanbicki, A. T.; Jonker, B. T.; Itskos, G.; Kioseoglou, G.; Petrou, A. Appl. Phys. Lett. 2002, 80, 1240. (13) Ohno, Y.; Young, D. K.; Beschoten, B.; Matsukura, F.; Ohno, H.; Awschalom, D. D. Nature 1999, 402, 790. (14) Hetterich, M.; Asshoff, P.; Wst, G.; Merz, A.; Kalt, H. Phys. Status Solidi C 2011, 8, 1157. (15) Fiederling, R.; Keim, M.; Reuscher, G.; Ossau, W.; Schmidt, G.; Waag, A.; Molenkamp, L. W. Nature 1999, 402, 787. (16) Svensson, C. P. T.; Martensson, T.; Tragardh, J.; Larsson, C.; Rask, M.; Hessman, D.; Samuelson, L.; Ohlsson, J. Nanotechnology 2008, 19, 305201. (17) Horng, R.-H.; Han, P.; Wuu, D.-S. IEEE Photonics Technol. Lett. 2008, 20, 1139. (18) Lee, W.; Kim, M.-H.; Zhu, D.; Noemaun, A. N.; Kim, J. K.; Schubert, E. J. Appl. Phys. 2010, 107, 063102. (19) Chang, H. J.; Chen, T. W.; Chen, J. W.; Hong, W. C.; Tsai, W. C.; Chen, Y. F.; Guo, G. Y. Phys. Rev. Lett. 2007, 98, 136403. (20) Lee, J.-H.; Choi, I.-H.; Shin, S.; Lee, S.; Lee, J.; Whang, C.; Lee, S.-C.; Lee, K.-R.; Baek, J.-H.; Chae, K. H.; Song, J. Appl. Phys. Lett. 2007, 90, 032504. (21) Nakamura, S.; Senoh, M.; Ichi Nagahama, S.; Iwasa, N.; Yamada, T.; Matsushita, T.; Kiyoku, H.; Sugimoto, Y. Jpn. J. Appl. Phys. 1996, 35, L217. (22) Chen, W.; Buyanova, I.; Nishibayashi, K.; Kayanuma, K.; Seo, K.; Murayama, A.; Oka, Y.; Thaler, G.; Frazier, R.; Abernathy, C.; et al. Appl. Phys. Lett. 2005, 87, 192107. (23) Buyanova, I. A.; Izadifard, M.; Chen, W. M.; Kim, J.; Ren, F.; Thaler, G.; Abernathy, C. R.; Pearton, S. J.; Pan, C.-C.; Chen, G.-T.; Chyi, J.-I.; Zavada, J. M. Appl. Phys. Lett. 2004, 84, 2599. (24) Chen, H.-S.; Yeh, D.-M.; Lu, Y.-C.; Chen, C.-Y.; Huang, C.-F.; Tang, T.-Y.; Yang, C. C.; Wu, C.-S.; Chen, C.-D. Nanotechnology 2006, 17, 1454. (25) Moustakas, T.; Xu, T.; Thomidis, C.; Nikiforov, A. Y.; Zhou, L.; Smith, D. J. Phys. Status Solidi A 2008, 205, 2560. (26) Pearton, S.; Norton, D.; Frazier, R.; Han, S.; Abernathy, C.; Zavada, J. IEEE Proc.: Circuits Devices Syst. 2005, 152, 312. (27) Hu, G.; Suzuki, Y. Phys. Rev. Lett. 2002, 89, 276601. (28) Yanase, A.; Hamada, N. J. Phys. Soc. Jpn. 1999, 68, 1607. (29) Chiu, C.; Lu, T.-C.; Huang, H.; Lai, C.; Kao, C.; Chu, J.; Yu, C.; Kuo, H.-C.; Wang, S.; Lin, C.; et al. Nanotechnology 2007, 18, 445201. (30) Wu, S.-H.; Lin, C.-Y.; Hung, Y.; Chen, W.; Chang, C.; Mou, C.Y. J. Biomed. Mater. Res. 2011, 99B, 81. (31) Stroud, R. M.; Hanbicki, A. T.; Park, Y. D.; Kioseoglou, G.; Petukhov, A. G.; Jonker, B. T.; Itskos, G.; Petrou, A. Phys. Rev. Lett. 2002, 89, 166602. (32) Gohda, Y.; Oshiyama, A. Phys. Rev. B 2008, 78, 161201. (33) Li, G.; Chua, S. J.; Xu, S. J.; Wang, W.; Li, P.; Beaumont, B.; Gibart, P. Appl. Phys. Lett. 1999, 74, 2821. (34) Tang, J.; Myers, M.; Bosnick, K. A.; Brus, L. E. J. Phys. Chem. B 2003, 107, 7501. (35) Fonin, M.; Pentcheva, R.; Dedkov, Y. S.; Sperlich, M.; Vyalikh, D. V.; Scheffler, M.; R̈ udiger, U.; G̈ untherodt, G. Phys. Rev. B 2005, 72, 104436. (36) Ham, M.-H.; Yoon, S.; Park, Y.; Bian, L.; Ramsteiner, M.; Myoung, J.-M. J. Phys.: Condens. Matter 2006, 18, 7703. (37) Buyanova, I.; Bergman, J.; Chen, W.; Thaler, G.; Frazier, R.; Abernathy, C.; Pearton, S.; Kim, J.; Ren, F.; Kyrychenko, F.; et al. J. Vac. Sci. Technol., B: Microelectron. Nanometer Struct. 2004, 22, 2668. (38) Han, L.; Zhu, Y.; Zhang, X.; Tan, P.; Ni, H.; Niu, Z.; et al. Nanoscale Res. Lett. 2011, 6, 84.
GaN MQDs and Fe3O4 nanoparticles. The underlying mechanisms arise from the combination of several factors. First, the internal strain in InGaN/GaN MQDs is relaxed, which eliminates the reduction of the spin coherence time due to the Rashba Hamitonian. Second, the space between nanorods can be filled with half-metal particles, such as Fe3O4 nanoparticles, which is able to enhance the number of heavy holes and electrons with the preferred spin orientation in MQDs. In addition, the structure of the nanodisk makes the distance between the ferromagnetic metal contact and the luminescent layer shorter than that of its planar counterpart. As a result, the degree of EL circular polarization of InGaN/GaN MQDs can reach up to 10.9% at room temperature in a magnetic field of 0.35 T. The demonstration of output spin polarization in semiconductor devices, particularly at room temperature in a low magnetic field, provides a great deal of new potential applications, such as integration of nonvolatile storage and logic devices, development of spin-laser for quantum information, control of magneto-electric effects, and cryptography of optical communication. In addition, the working principle shown here is quite general, which can be applied to many other material systems. Our approach therefore can open up a new avenue for the further research and development of spintronics based on semiconductor devices. In view of the mature nitride LED industry at present, our study should be very useful and timely.
■
ASSOCIATED CONTENT
S Supporting Information *
I−V curves of InGaN/GaN MQD nanorods decorated by Fe3O4 nanoparticles with and without an applied magnetic field, electroluminescence spectra for the samples with different structures and composites, electroluminescence spectra of InGaN/GaN MQW LED and InGaN/GaN MQD LED under different applied currents, and electroluminescence and photoluminescence spectra under different applied current and excitation power. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*(Y.F.C.) E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS This work was supported by the National Science Council and the Ministry of Education of the Republic of China. REFERENCES
(1) Baibich, M. N.; Broto, J. M.; Fert, A.; Van Dau, F. N.; Petroff, F.; Etienne, P.; Creuzet, G.; Friederich, A.; Chazelas, J. Phys. Rev. Lett. 2008, 61, 2472. (2) Binasch, G.; Gr̈unberg, P.; Saurenbach, F.; Zinn, W. Phys. Rev. B 1989, 39, 4828. (3) Prinz, G. A. Science 1998, 282, 1660. (4) Awschalom, D.; Loss, D.; Samarth, N. Semiconductor Spintronics and Quantum Computation; Springer Verlag: Berlin, Germany, 2002. (5) Wolf, S. A.; Awschalom, D. D.; Buhrman, R. A.; Daughton, J. M.; von Molnr, S.; Roukes, M. L.; Chtchelkanova, A. Y.; Treger, D. M. Science 2001, 294, 1488. (6) Bao, J.; Bragas, A. V.; Furdyna, J. K.; Merlin, R. Nat. Mater. 2003, 2, 175. G
dx.doi.org/10.1021/nl5003312 | Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters
Letter
(39) Wu, Y.-R.; Chiu, C.; Chang, C.-Y.; Yu, P.; Kuo, H.-C. IEEE J. Sel. Top. Quantum Electron. 2009, 15, 1226. (40) Wang, C.-Y.; Chen, L.-Y.; Chen, C.-P.; Cheng, Y.-W.; Ke, M.-Y.; Hsieh, M.-Y.; Wu, H.-M.; Peng, L.-H.; Huang, J. Opt. Express 2008, 16, 10549.
H
dx.doi.org/10.1021/nl5003312 | Nano Lett. XXXX, XXX, XXX−XXX