Chapter 2 Mechanism and Properties…
2.2 Key issues for realizing high efficiency LEDs
2.2.2 Light extraction of GaN LEDs
Limitations in light extraction come from total internal reflection at interfaces and light absorption within the device or in the packaging. The generation of light in active region of an LED is most captured with GaN and sapphire by the guided modes.
It is due to the high contrast refractive index at the GaN(n=2.45)/air(n=1) and GaN/sapphire(n=1.78) interfaces, resulting in total internal reflection that traps light in the high refractive index and in sapphire substrate. To improve the light extraction efficiency, there are several methods reported, such as patterned sapphire substrates, surface texturing, and air-void formation by nano-patterning.
2.3 Wet etching
2.3.1 Defect properties on GaN surface
Successful fabrication of GaN-based devices depends on the ability to grow epilayer on substrates such as sapphire or silicon carbide, with a low density of defects.[3,4] A high density (108~1010 cm-2) of threading dislocations results from the lattice constant and thermal expansion coefficient mismatch in the nitride film.[5-7]
We knew that these defects have influence on both the electrical and optical properties of the material.[8,9] Therefore, the availability of reliable and quick methods to investigate the defects and dislocations in GaN is of great interest.
21
Wet-chemical etching is a commonly used technique for surface defect investigation due to its advantage of low cost and simple experimental procedure. Hot phosphoric acid (H3PO4) and molten potassium hydroxide (KOH) have been shown to etch pits at defect sites on the c-plane of GaN.[10-13] The following segments was presented by P. Visconti and co-workers. Kozawa et al.[10] found etch pits tentatively ascribed to dislocations using molten KOH to etch metalorganic chemical-vapor deposition (MOCVD) GaN samples. However, the etch-pit density (EPD) was 2×107 cm-2, while the dislocation density found by transmission electron microscopy (TEM) was 2×108 cm-2. Hong and co-workers[11,12] related the hexagonal-shaped etch pits formed by H3PO4 etching on MOCVD GaN samples to nanopipes (open-core screw dislocations). EPD is hundreds or thousands times lower than the dislocation density evaluated by TEM. Lu[13] investigated etch pits formed on MOCVD GaN samples by molten KOH etching. By atomic-force microscopy (AFM) and TEM analyses, they attributed the origin of etch pits. Besides, the origin of etch pits is still controversial and the obtained EPD (in the range 4×105~1×108 cm-2) is lower than the dislocation density (108~1010 cm-2) found by TEM. The etch pits size varies even in the same sample. The shapes of etch pits are well correlated with types of defects, and the etch pits density (EPD) may correspond to the density of defects. However, for GaN, the density, types, and distribution of defects vary significantly due to growth-related
22
conditions, which makes it difficult to reach an agreement about the origin of the etch pits, and it can be even more difficult for test techniques.
2.3.2 Etching process in molten KOH
The discrepancy of etching characteristics in Ga-face (+c GaN, Ga-polarity) and N-face (-c GaN, N-polarity) has been specifically investigated as illustrated in Fig.
2.3.1. Some reports showed that gallium nitride could be etched in the aqueous sodium hydroxide (NaOH) solution but etching ceased when the formation of an insoluble coating of presumably gallium hydroxide (Ga(OH)3) [14,15]. For further etching, it would need removing of the coating by continual jet action. Various aqueous acid and base solutions have been tested for etching of GaN were list in Table 2.3.1 [16-18]. The undetermined etch rate (nm/min) was because it various from sample to sample and differences in the defect density. According to the research reports in recent years; the common cognition related to gallium nitride etching process was that the most of gallium nitride could be etched rapidly in N-face. The reason for the face-dependent gallium nitride etching process has been studied by Li et al., who utilized the X-ray photoelectron spectroscopy (XPS) to examine the
surface chemistries before and after etching process in aqueous KOH solutions for both Ga- and N-face gallium nitride. The conclusion is that the different etching
23
results in Ga- and N-face gallium nitride crystals are due to the different states of surface bonding. Besides, the most important is the etching process only dependent on the polarities, not on the surface morphology, growth condition and which atoms form the surface termination layer. The GaN chemical etching reaction with KOH could be described as the following formula [19]:
2GaN+3H2OKOH → Ga2O3+2NH3 (2-11) Here, the molten KOH act as a catalyst and a solvent for the resulting Ga2O3 (Fig.
2.3.2 (d)) as well. The mechanism about etching N-face gallium nitride substrate was illustrated in Fig. 2.3.2. The hydroxide ions (OH-) were first absorbed on the gallium nitride surface (Fig. 2.3.2 (b)) and finally react with Ga atoms once the OH- ions with sufficient kinetic energy as shown in the Fig. 2.3.2 (c). The etching could be started at step (c) if the surface was Ga-terminated. The inertness of Ga-face GaN was ascribed to the hydroxide ions would be repelled by the negatively-charged triple dangling bonds of nitrogen near the surface. Thus, if the Ga-face GaN was Ga-terminated, the etching process stops after the first gallium atom layer was removed. In contrast, for the N-face GaN, every nitrogen atom bears a single dangling bond to prevent the hydroxide ions attacking from Ga atoms.
24
Table 2.3 Various chemicals etch GaN.[20]
25
Fig. 2.3.1 Illustration of different polarity, (a) Ga-face (+c GaN, GaN polarity ), (b) N-face (-c GaN, N-polarity). [22]
26
Fig. 2.3.2 Schematic diagrams of the cross section GaN film viewed along [-1-120]
direction for N-polar GaN to explain the mechanism of the polarity selective etching.
(a) Nitrogen terminated layer with one negatively charged dangling bond on each nitrogen atom; (b) absorption of hydroxide ions; formation of oxides; (d) dissolving the oxides. [23]
27
2.4 Reference
[1] A. Usui, H. Sunakawa, A. Sakai and A. A. Yamaguchi, “Thick GaN epitaxial growth with low dislocation density by hydride vapor phase epitaxy,” Jpn. J. Appl.
Phys. 36, L889 (1997).
[2] M. Yamada, T. Mitani, Y, Narukawa, S. Shioji, I. Niki, S. Sonobe, K. Deguchi, M.
Sano and T. Mukai, “InGaN-based near-ultraviolet and blue-light-emitting diodes with high external quantum efficiency using a patterned sapphire substrate and a mesh electrode,” Jpn. J. Appl. Phys. 41, L1431 (2002).
[3] H. Morkoc¸, S. Strite, G. B. Gao, M. E. Lin, B. Sverdlov, and M. Burns,
“Large‐band‐gap SiC, III‐V nitride, and II‐VI ZnSe‐based semiconductor device technologies,” J. Appl. Phys. 76, 1363 (1994).
[4] S. Nakamura, T. Mukai, and M. Senoh, “Candela‐class high‐brightness
InGaN/AlGaN double‐heterostructure blue‐light‐emitting diodes,” Appl. Phys. Lett., 64, 1687 (1994).
[5] S. D. Lester, F. A. Ponce, M. G. Craford, and D. A. Steigerwald, “High
dislocation densities in high efficiency GaN‐based light‐emitting diodes,” Appl. Phys.
Lett. 66, 1249 (1995).
[6] W. Qian, M. Skowronski, M. DeGraef, K. Doverspike, L. B. Rowland, and D. K.
Gaskill, “Microstructural characterization of α‐GaN films grown on sapphire by
28
organometallic vapor phase epitaxy,” Appl. Phys. Lett. 66, 1252 (1995).
[7] X. H. Wu, L. M. Brown, D. Kapolnek, S. Keller, B. Keller, S. P. Den-Baars, and J.
S. Speck, “Defect structure of metal‐organic chemical vapor deposition‐grown epitaxial (0001) GaN/Al2O3,” J. Appl. Phys. 80, 3228 (1996).
[8] B. Garni, J. Ma, N. Perkins, J. Liu, T. F. Kuech, and M. G. Lagally, “Scanning tunneling microscopy and tunneling luminescence of the surface of GaN films grown by vapor phase epitaxy,” Appl. Phys. Lett. 68, 1380 (1996).
[9] S. J. Rosner, E. C. Carr, M. J. Ludowise, G. Girolami, and H. I. Erikson,
“Correlation of cathodoluminescence inhomogeneity with microstructural defects in epitaxial GaN grown by metalorganic chemical-vapor deposition,” Appl. Phys. Lett.
70, 420 (1997).
[10] T. Kozawa, T. Kachi, T. Ohwaki, Y. Taga, N. Koide, and M. Koike, “Dislocation Etch Pits in GaN Epitaxial Layers Grown on Sapphire Substrates,” J. Electrochem.
Soc. 143, L17 (1996).
[11] S. K. Hong, T. Yao, B. J. Kim, S. Y. Yoon, and T. I. Kim, “Origin of
hexagonal-shaped etch pits formed in (0001) GaN films,” Appl. Phys. Lett. 77, 82 (2000).
[12] S. K. Hong, B. J. Kim, H. S. Park, Y. Park, S. Y. Yoon, and T. I. Kim,
“Evaluation of nanopipes in MOCVD grown (0 0 0 1) GaN/Al2O3 by wet chemical
29
etching,” J.Cryst. Growth 191, 275 (1998).
[13] L. Lu, Z. Y. Gao, B. Shen, F. J. Xu, S. Huang, Z. L. Miao, Y. Hao, Z. J. Yang, G.
Y. Zhang, X. P. Zhang, J. Xu, D. P. Yu, “Microstructure and Origin of dislocation etch pits in GaN epilayers grown by metal organic chemical vapor deposition,” J.
Appl. Phys., 104, 123525 (2008)
[14] T.L. Chu, “Gallium Nitride Films,” J. Electrochem. Soc. 118, 1200 (1971).
[15] J.I. Pankove,” Electrolytic Etching of GaN,” J. Electrochem. Soc. 119, 1118 (1972).
[16] H. Cho, D.C. Hays, C.B. Vartuli, S.J. Pearton, C.R. Abernathy, J.D. MacKenzie, F. Ren, J.C. Zolper, “Wet chemical etching survey of III-nitrides,” Mater. Res. Soc.
Symp. Proc. 483, 265 (1998).
[17] C.B. Vartuli, S.J. Pearton, C.R. Abernathy, J.D. MacKenzie, F. Ren, J.C. Zolper, R.J. Shul, “Wet chemical etching survey of III-nitrides,” Solid-State Electron. 41 (12), 1947 (1998).
[18] S.J. Pearton, R.J. Shul, Gallium nitride I, in: J. Pankove, T.D. Moustakas (Eds.),
“The Properties of Hydrogen in GaN and Related Alloys,” Semiconductor and Semimetals Series, vol. 50, Academic Press, New York, NY, p. 103 (1998).
[19] D. Li, M. Sumiya, S. Fuke, D. Yang, D. Que, Y. Suzuki, Y. Fukuda,
30
“Selective etching of GaN polar surface in potassium hydroxide solution studied by x-ray photoelectron spectroscopy,” J. Appl. Phys. 90, 4219 (2001).
[20] D. A. Stocker, E. F. Schubert and J. M. Redwing, “Crystallographic wet chemical etching of GaN,” Appl. Phys. Lett., Vol. 73, No. 18, 2 November (1998).
[21] M. H. Lo, P. M. Tu, C. H. Wang, C. W. Hung, S. C. Hsu, Y. J. Cheng, H. C. Kuo, H. W. Zan, S. C. Wang, C. Y. Chang, and S. C. Huang, “High efficiency light emitting diode with anisotropically etched GaN-sapphire interface,” Appl. Phys. Lett,95 041109 (2009)
[22] O Ambacher, “REVIEW ARTICLE Growth and applications of Group III-nitrides,” J. Phys. D: Appl. Phys. 31 2653–2710 (1998)
[23] D. S. Li, H. Chen, H. B. Yu, H. Q. Jia, Q. Huang, and J. M. Zhou, “Dependence of leakage current on dislocations in GaN-based light-emitting diodes,” J. Appl. Phys., Vol. 96, No. 2, pp. 1111-1114, Jul. (2004)
31
Chapter 3 Measurement System
3.1 Scanning electron microscopy (SEM)
The scanning electron microscope is built of the following parts:
(i) The electron gun
(ii) The system of three-stage electromagnetic lens is used to demagnify (focus, condense) the electron beam diameter to 5~10 nm at the specimen.
(iii) Detectors may detect electrons, X-ray or cathodo-luminescent (CL) light.
(iv) The microscope column is evacuated to 10-5 torr.
Fig. 3.1.1 shows that schematic diagram of a scanning electron microscope (SEM).
Two pairs of deflection coils are shown in the SEM column. This double deflection allows the scanning beam to pass through the final aperture. Four pairs are actually used, for double deflection in x and y directions.
Fig. 3.1.1 Schematic diagram of a scanning electron microscope (SEM).[1]
32
SEM is a technique which forms an image of microscopic region of the specimen surface. An electron beam from 5~10 nm in diameter is scanned across the specimen.
The interaction of the electron beam with the specimen produces a series of phenomena such as:
(i) backscattering of electrons of high energy (ii) secondary electrons of low energy (iii) absorption of electrons
(iv) X - ray
(v) visible light (cathodoluminescence, CL)
Fig. 3.1.2 indicates that any of these signals can be continuously monitored by detectors.
Fig. 3.1.2 Information that can be generated in the SEM by an electron beam striking
33
the sample.[2]
3.2 Cathodoluminescent spectroscopy (CL)
Cathodoluminescence (CL) is a SEM-based technique that can be used for analyzing the characteristic of semiconductor materials and devices. CL is the emission of light as the result of electron or “cathode-ray” bombardment. SEM-based and CL can provide information on the concentration and distribution of luminescent centers, distribution and density of electrically active defects, and electrical properties including minority carrier diffusion lengths and lifetimes.
Fig. 3.2 JSM-7000F SEM and CL System.
34
3.3 Atomic Force Microscopy (AFM)
Atomic force microscopy (AFM) or scanning force microscopy (SFM) is a very high-resolution type of scanning probe microscopy (SPM) instead of optical imaging one. In 1986, the AFM was invented by Gerd Binnig, Christoph Gerber, and Calvin F.
Quate. The AFM is one of the foremost tools for imaging, measuring, and manipulating matter at the nanoscale. A very tiny, pyramidal probe is attached on the cantilever. The tip must be very tiny (single atom size) with sharp angle for large-area scan.
The AFM utilizes a sharp probe moving over the surface of a sample in a raster scan. When the probe is approaching sample surface, attractive (van der Waals force) or repulsive force (Coulomb repulsion) between tip and sample is formed and detected. Forces between the tip and the sample lead to a deflection of the cantilever according to Hooke's law. The interaction force causes cantilever to shift along z-axis and thus the topology of sample is obtained. The small probe-sample separation (on the order of the instrument’s resolution) makes it possible to take measurements over a small area. To acquire an image the microscope-scans the probe over the sample while measuring the local property in question. The resulting image resembles an image on a screen in that both consist of many rows or lines of information placed on above the other. Unlike the traditional microscopes, scanned probe system do not use
35
lenses, so the size of the probe rather than diffraction effect generally limits their resolution.
Followings are the operating mode of AFM, shown as the Figs. 3.3[(a)-(c)]:
1. Contact mode: The Interaction mainly comes from repulsive force between tip and sample. It is easy to obtain atomic-scale resolution, but easy to damage surface of sample.
2. Non-contact mode: The Interaction mainly comes from van der Waals force between tip and sample. The tip never touches sample surface; resolution is lower (~50 nm). The surface of samples is preserved.
3. Tapping mode: The tip touches the surface of samples periodically. The resolution could be as high as contact-mode. The surface of samples could be damaged sometimes.
36
(a) (b)
(c)
Fig. 3.3 Operating mode of AFM (a) contact mode, (b) non-contact mode, (c) tapping mode.[2]
3.4 Micro photoluminescence spectroscopy (µ-PL)
Photoluminescence (PL) spectroscopy has been used as a measurement method to detect the optical properties of the materials because of its nondestructive characteristics. PL is the emission of light from the material under optical excitation.
Reducing the laser beam spot size to micrometer by beam expanders and objective lens is the so-called µ-PL. Fig 3.4.1 illustrates the photoluminescence process. The laser light source used to excite carriers should have large energy band gap than the semiconductors. When the laser light absorbed within the semiconductors, it should excite the carriers from the valence band to the conduction band. Then, is produces
37
the electrons in the conduction band and the holes in the valance band. When the electron in an excited state returns to the initial state, it will emit a photon whose energy is equal to energy difference between the excited state return and the initial state, therefore, we can observe the emission wavelength peak from PL spectrum.
Fig. 3.4.1 Inter-band transitions in photoluminescence system.[3]
38
3.5 Reference
[1] http://en.wikipedia.org/wiki/File:Schema_MEB_(en).svg [2] Class of Materials analysis, S. H. Yang, NCTU in Tainan [3] http://ned.ipac.caltech.edu/level5/Sept03/Li/Li4.html
39
Chapter 4 Experiment Process
4.1 Experiment Process Flow
First, we reveal the bulk GaN pits of threading dislocation by chemical wet etching.
Second, filling the defect pits by coating silica nanospheres. Finally, regrow UV LED structure.
In this thesis, we prepare two process samples and a bulk GaN sample as reference.
One of the process sample use molten KOH as etching solution, while the other one use phosphoric acid. We observe the surface morphology and use optical measurement to analyze the epitaxial quality.
4.2 GaN surface etching by phosphoric acid and molten KOH
We verified that there are three types of etched pits：screw, edge, mixed（α, β, and γ）type [1]. Each etched pits correspond to different threading dislocations and
threading dislocations having a screw component act as strong nonradiative centers [2], and the different etching liquid may forms different etched pit types. For this reason, a test round of the two GaN wafer was immersed in molten KOH and phosphoric acid (H3PO4), respectively.
The etching solution temperature and etching time dominate the pits size and
40
defect density. If the solution temperature or etching time is insufficient, it may lead to small pit size that is unable to confine the silica nanospheres in the pits. On the other hand, too much etching time or high solution temperature leads to larger pits size, which need to re-grow a thicker layer increasing the difficulty to seal the pits.
Moreover, if the GaN is over-etched, the flat area of surface would be too fragile to provide the platform for re-growth. Therefore, it is important to find an appropriate recipe for etching process. As shown in fig. 4.2.1, the SEM image of GaN surface after etching by KOH and H3PO4. In the etching time versus temperature scheme, we chose 300°C 3 minutes for KOH and 240°C and 4 minutes for H3PO4 etching process.
Both KOH and H3PO4 etching pits size are average about 1.2 um to 1.6 um, and the pit density is about 107 cm-2 and 106 cm-2 , respectively.
As shown in fig. 4.2.2, three etched pits type are observed. It is known that a screw type threading dislocation creates a step when it terminates at the GaN surface. In KOH etching process, these steps are easily attacked by OH-, and further etching finally stop at the Ga terminate due to the chemical stabilization of Ga face as showed in fig. 4.2.3 [1]. Finally, the screw type threading dislocations can be etched to an inverse trapezoid. On the other hand, the edge type etched pits correspond to edge type threading dislocations. Since every atom in this line has dangling bond, the atom in this line was easily attacked, and finally formed a inversed pyramid. Mixed type
41
threading dislocations has both screw and edge type morphology.
Fig. 4.2.4 shows the top view SEM image of bulk GaN after etching test. The GaN etched by KOH revealed three types of etched pits. After calculating the etching pits density, we found that the edge type pits dominated KOH etching pits, the ratio of edge type pits is around 85%. On the other hand, We found only screw and mixed type etched pits in H3PO4 etching sample. The ratio of screw type and mixed type pits are around 95% and 5%, respectively.
In summary, we could infer that molten KOH prefer to etch edge type dislocation, while H3PO4 solution prefer to attack screw type dislocation, which is treat as non-radiative center[2]. In other words H3PO4 solution would much easier reveal non-radiative center than molten KOH.
42
(a)
(b)
Fig. 4.2.1 Bulk GaN surface etching test by (a)KOH and (b)H3PO4 solution
43
Fig. 4.2.2 The SEM image of GaN wet etching results, three etched pits types are observed.
44
Fig. 4.2.3 [1] (a) Step formed at the beginning of etching screw type threading dislocation. (b) A Ga face to prevent further vertical etching. (c) (d) Edge type threading dislocation was easily etching along the vertical dangling bond line.
(b) (a)
(c) (d)
45
(a)
(b)
Fig. 4.2.4 GaN wafer etched by (a)KOH and (b) H3PO4
46
4.3 Coating silica nanospheres on GaN etching surface
We coated diameters 100nm silica nanospheres on GaN surface after revealing the pits of threading dislocation. The SEM image is shown in fig. 4.3.1. Next step we removed the silica on the flat surface area and leaves silica nanspheres in the etching pits. The surface cleaning is the key issue. We wipe off the nanospheres on wafer surface by dust-free cloth, then clean by ultrasonic vibration in DI water for 5 minutes.
Fig. 4.3.2 shows the difference between our clean process and without using dust-free cloth. The additional wiping process could totally remove the residual surface nanospheres.
Fig. 4.3.3 shows the GaN wafer after all cleaning process. We confined the silica nanospheres in etching pits successfully, and no nanospheres remained on surface.
Fig. 4.3.1 GaN wafer coating with 100nm silica nanopheres after etching process.
47
(a) (b)
Fig. 4.3.2 Nanosphere cleaning process with (a) and without (b) dust-free cloth wiping off the surface.
(a) (b)
Fig.4.3.3 GaN wafer etched by KOHand H3PO4, then spin coating silica nanospheres with diameter 100nm. After cleaning process, the KOH sample (a) and H3PO4
confined the nanospheres successfully.(b)
48
4.4 Regrowth InAlGaN LED structure
The InAlGaN based LED structure were regrown on all samples by using MOCVD. The LED structure are consisted of a 0.6um u-AlGaN layer, 2.04um n-InAlGaN layer. Ten pairs of InGaN/InAlGaN multiple quantum well active layer, 29.7nm AlGaN electron blocking layer, and a 55.6 nm p-GaN layer. The LED structure is showed in Fig. 4.4
The detailed parameters of all samples are shown in table 4.4.
Fig. 4.4 Scheme of LED structure
49
Table 4.4 The detail recipe and etching pits density of all process sample
Sample Ref KOH H3PO4
Etching solution - KOH H3PO4
Etching time - 3min 4min
Etching Pits
Density(cm-2)
- 1.6x107 4.25x106
Coat Silica Size - 100nm 100nm
50
4.5 Reference
[1] L. Lu, Z. Y. Gao, B. Shen, F.J.Xu, S.Huang, Z. L. Miao, Y.Hao, Z. J. Yang,
[1] L. Lu, Z. Y. Gao, B. Shen, F.J.Xu, S.Huang, Z. L. Miao, Y.Hao, Z. J. Yang,