- Biomedical Optics & Medical Imaging
- Defense & Security
- Electronic Imaging & Signal Processing
- Illumination & Displays
- Lasers & Sources
- Micro/Nano Lithography
- Optical Design & Engineering
- Optoelectronics & Communications
- Remote Sensing
- Sensing & Measurement
- Solar & Alternative Energy
- Sign up for Newsroom E-Alerts
- Information for:
Illumination & Displays
Growing thin films that contain embedded voids
Uniform voiding defects in thin films offer unique designs for brighter and more efficient LEDs.
23 June 2011, SPIE Newsroom. DOI: 10.1117/2.1201105.003750
Over the last decade gallium nitride (GaN) has become the semiconductor material of choice in several optical and electronic devices. In particular, GaN can help realize the potential of solid-state lighting (SSL), a multi-billion dollar emerging technology using LEDs that promises to fundamentally alter lighting and contribute to energy savings. Unfortunately, GaN and other III-nitride materials—such as aluminum and indium nitrides or their alloys—suffer from a high density of dislocations and other defects (∼108–1010cm−2) because of the lack of lattice-matched substrates. In general, these defects act as non-radiative recombination and scattering centers that impact the diffusion length1 and minority carrier lifetime, reduce thermal conductivity,2 and form easy pathways for impurity diffusion. Thus, they limit the performance, reliability, breakdown voltage, and lifetime3 of both optoelectronic and power devices. Here, we discuss our approach to reduce GaN defects by controlling the void density.
Dislocations generated at the GaN/sapphire interface run through the crystalline structure of the films and terminate at free surfaces, affecting the function of active, multi-quantum well (MQW) layers. To combat this, our embedded voids approach (EVA)4 intentionally introduces a high density of micro-voids—a few micrometers in length and less than a micrometer in diameter—into the GaN layer near its interface with the substrate. By doing so, we create an efficient ‘trapping zone’ near the GaN/sapphire interface, where the voids act as both sinks for defects as well as expansion joints for lattice mismatches. The active layers of III-nitride epitaxial films are then grown on the void-embedded layer, free from dislocation disturbance.
EVA is a three-step process (see Figure 1). First, we grow bulk GaN by metal organic chemical vapor deposition on the sapphire substrate. Then, we form GaN nanowires (NWs) from the bulk-grown film using maskless inductively coupled plasma-reactive ion etching (ICP-RIE, see Figure 2). Maskless ICP-RIE is based on the assumption that dislocations and other defects represent sites of high elastic energy where localized elevated etching rates proceed. Thus, we can form good quality GaN NWs that exhibit superior optical, electrical, and x-ray quality compared to bulk material. Additionally, because GaN NWs are a strain-free, low defect semiconductor material with high surface ratio, they can be used in a variety of applications.5–7 In the final step of EVA, we overgrow epitaxial GaN on the GaN NW template.
3D schematics of (a) bulk-grown gallium nitride (GaN) film on a sapphire substrate, (b) GaN nanowires (NWs) formed using the maskless inductively coupled plasma-reactive ion etching (ICP-RIE), and (c) NW overgrowth and void formation.5
(a) Schematic of maskless ICP-RIE for GaN NW formation, where the etching gas is composed of boron trichloride (BCl3
) and chlorine (Cl2
) in a 1:5 ratio and (b) high resolution scanning electron micrograph of GaN NWs produced.5
We revealed with transmission electron microscopy (TEM) the void formation between NWs (see Figure 3). Enhancement of dislocation-dislocation interactions and the trapping mechanism of dislocations after coalescence—where voids act as sinks—were also made apparent. Atomic force microscopy (AFM) performed on the overgrown films showed a planarized surface with lower roughness than the original.4, 5 We confirmed with both TEM and AFM that our EVA reduced dislocation density uniformly by 2 to 3 orders of magnitude. The conventionally grown GaN film on sapphire substrate, free of voids, has a dislocation density of ∼1010cm−2 in contrast to ∼107cm−2 achieved in our GaN films with embedded voids.4, 5
Transmission electron micrographs of dislocation networks in (a) GaN film grown conventionally on a sapphire substrate, and (b) GaN film regrown with embedded voids.4
Having fabricated and characterized our GaN films, we then examined this material for use in LEDs (see Figure 4). We first grew multiple emitting layers conformally on facets of n-GaN nanowires, which were either perpendicular (i.e., nonpolar planes) or tilted (i.e., semipolar planes) to the substrate. By doing so, we improved the quantum efficiency because of the reduction of polarization fields induced by biaxial strain in the emitting layers. We then deposited fully coalesced p-GaN on these NWs. Overgrowth on the nanowire tips resulted in the inclusion of a high density of voids—approximately 1μm in height—in the LED device structure. As a result, the light output intensity of our NW LEDs was more than three times that of corresponding conventional LEDs grown simultaneously (see Figure 5). Our NW LEDs have a reduced defect density, increased effective area of conformally grown emitting layers, absence of polar plane orientation, and improved light extraction.8
Figure 4. Schematics of (a) conventional and (b) our NW LEDs.
(a) Electrical probe inspection of new LED devices on a wafer, (b) integrated light output intensity plotted against applied current density of conventional and NW LEDs.8
In summary, we have developed a method for embedding voids in GaN thin films near the interface with the substrate as a means for diverting dislocations away from the free surface. Our EVA technique excels over many existing defect reduction techniques because we eliminate the intermediate masking/patterning step in device fabrication. Additionally, EVA provides a feasible method to produce large areas of GaN with a low density of dislocations. This will be crucial in cutting manufacturing costs and enabling market penetration of GaN devices. As we look to make LED lighting an economically viable option for the future, low dislocation densities ensure a longer lifetime and better thermal conductivity. Additionally, our EVA could be applied to other material systems to achieve LEDs with MQWs of different materials on foreign substrates, to expand the color range of the devices and solve some of the economical demands of LED technology. Improving the performance LEDs based on void-embedded materials can be used in an array of applications, ranging from sterilizing water and treating skin diseases to aiding forensic investigations and combating forgery. The possibility of growing conformal MQWs, or multiple emitting layers, on GaN NW facets as part of our EVA technique provides the basis for many novel device structures that could be explored. These possibilities and prospects constitute our ongoing work.
Pavel Frajtag, Nadia El-Masry
Department of Material Science and Engineering
North Carolina State University (NCSU)
Pavel Frajtag is a graduate student whose current interests include sidewall epitaxy, III-nitrides nanowires, and their applications.
Nadia El-Masry is a graduate professor at NCSU and a program director in the Division of Materials Research at the National Science Foundation. She has published in the area of III-V oxides and ferromagnetic thin film materials on semiconductor substrates. She has experience in thin film growth and characterization.
Salah Bedair, Aadhithya Hosalli, Geoffrey Bradshaw
Department of Electrical and Computer Engineering
North Carolina State University
Salah Bedair is a distinguished professor who has published in the area of semiconductor materials and devices, including atomic layer epitaxy growth, laser-assisted deposition, and molecular beam epitaxy. His research is in the field of semiconductor devices such as detectors, wave guides, solar cells, and LEDs.
Aadhithya Hosalli is a graduate student whose current interests include III-nitride LED devices and materials, nanowires, and their properties.
Geoffrey Bradshaw is a graduate student whose current interests include multi-junction solar cells, strained layer superlattices, and fabrication of III-nitride LED devices.
1. K. Kumakura, T. Makimoto, N. Kobayashi, T. Hashizume, T. and H. Hasegawa, Minority carrier diffusion length in GaN: Dislocation density and doping concentration dependence, Appl. Phys. Lett
. 86, pp. 052105, 2005. doi:10.1063/1.1861116
2. C. Mion, J. F. Muth, E. A. Preble, D. Hanser, Accurate dependence of gallium nitride thermal conductivity on dislocation density, Appl. Phys. Lett
. 89, pp. 092123, 2006. doi:10.1063/1.2335972
3. S. Nagahama, High-power and long-lifetime InGaN multi-quantum-well laser diodes grown on low-dislocation-density GaN substrates, Jpn. J. Appl. Phys
. 39, pp. L647-650, 2000. doi:10.1143/JJAP.39.L647
4. P. Frajtag, N. A. El-Masry, N. Nepal, S. M. Bedair, Embedded voids approach for low defect density in epitaxial GaN films, Appl. Phys. Lett.
98, pp. 023115, 2011. doi:10.1063/1.3540680
5. P. Frajtag, J. P. Samberg, N. A. El-Masry, N. Nepal, S. M. Bedair, Embedded voids formation by overgrowth on GaN nanowires for high-quality GaN films, J. Cryst. Growth
322, pp. 27-32, 2011. doi:10.1016/j.jcrysgro.2011.02.032
6. H. J. Chang, Y. P. Hsieh, T. T. Chen, Y. F. Chen, C.-T. Liang, T. Y. Lin, S. C. Tseng, L. C. Chen, Strong luminescence from strain relaxed InGaN/GaN nanotips for highly efficient light emitters, Opt. Express
15, pp. 9357-9365, 2007. doi:10.1364/OE.15.009357
7. H.-S. Chen, D.-M. Yeh, Y.-C. Lu, C.-Y. Chen, C.-F. Huang, T.-Y. Tang, C. C. Yang, C.-S. Wu, C.-D. Chen, Strain relaxation and quantum confinement in InGaN/GaN nanoposts, Nanotechnol.
17, pp. 1454-1458, 2006. doi:10.1088/0957-4484/17/5/048
8. P. Frajtag, A. M. Hosalli, G. K. Bradshaw, N. Nepal, N. A. El-Masry, S. M. Bedair, Improved light-emitting diode performance by conformal overgrowth of multiple quantum wells and fully coalesced p-type GaN on GaN nanowires, Appl. Phys. Lett
. 98, pp. 143104, 2011. doi:10.1063/1.3572032