SELECTIVE UV-LASER PROCESSING FOR LIFT ... - Semantic Scholar

SELECTIVE UV-LASER PROCESSING FOR LIFT-OFF OF GaN THIN FILMS FROM SAPPHIRE SUBSTRATES W.S. Wong†, J. Krüger†, Y. Cho†, B.P. Linder‡, E.R. Weber†, N.W. Cheung‡, and T. Sands† †

Department of Materials Science and Mineral Engineering, University of California, Berkeley, CA 94720 ‡ Department of Electrical Engineering and Computer Science, University of California, Berkeley, CA 94720

ABSTRACT

Gallium nitride (GaN) thin films on sapphire substrates were successfully separated and transferred onto Si substrates by pulsed UVlaser processing. A single 600 mJ/cm2, 38 ns KrF excimer laser pulse was directed through the transparent substrate to induce a rapid thermal decomposition of the GaN at the GaN/sapphire interface. The decomposition yields metallic Ga and N2 gas that allows separation of the GaN film from the substrate. Three-micron-thick free-standing GaN membranes were also fabricated using the laser lift-off technique. Surface roughness of the exposed interfacial layer was measured to be ~24 nm (rms) by atomic force microscopy. Photoluminescence measurements of the GaN membranes showed no optical degradation of the GaN after liftoff from the sapphire. Based on a 10 meV red-shift of the donor-bound exciton peak, an estimated biaxial compressive stress of ~0.4 GPa in the GaN film was relieved by separation from the sapphire growth substrate.

INTRODUCTION The recent advancements in the growth and processing of GaN thin films have allowed rapid development of GaN based light-emitting diodes (LED) and lasers.1,2 GaN based blue and blue-green light emitting devices have the potential to enable a wide range of technological advances including full-color LED displays, efficient and reliable green LED traffic lights, and low-cost blue lasers for ultra-high density optical storage. In spite of this swift progress, the symmetry of the GaN crystal structure, combined with the high GaN growth temperatures prevents deposition of high-quality material directly onto common semiconductor substrates such as GaAs, InP or Si. This restriction impedes the direct integration of GaN with existing electronic and optoelectronic semiconductor technologies. Consequently, GaN thin films are deposited using heteroepitaxy on available but dissimilar substrates such as sapphire or SiC.3-6 The most commonly used

Proceedings of the Symposium on LED for Optoelectronic Applications and the 28th State of the Art Programs on Compound Semiconductors 98-2, 377 (1998).

growth substrate, sapphire, still imposes constraints on GaN based devices due to its relatively low thermal conductivity for high power electronic applications and difficulty in cleaving smooth mirror facets to form laser cavities. A major impediment to GaN technology has been the thermal decomposition of this compound at relatively low temperatures (~800-900°C) to produce metallic Ga and N2 gas.7,8 Although this decomposition prevents the growth of large bulk crystals and complicates efforts to grow thin films, the decomposition reaction may be exploited to create a thin sacrificial interface layer for lift-off of GaN on sapphire substrates by selective laser processing. Epitaxial lift-off, in conjunction with wafer bonding, may be used as a direct approach for integration of GaN with other dissimilar semiconductors. Wet-chemical etch lift-off processes have been demonstrated by other groups to separate GaAs thin films grown on GaAs substrates.9-11 In addition, a sacrificial ZnO buffer layer has also been used in the lift-off of very thick (200 - 400 µm) GaN on sapphire.12 UV-laser processing using a Q-switched Nd:YAG laser has been reported to separate GaN thin films from sapphire substrates. 13 Recently, we have demonstrated the efficacy of this laser-selective process at shorter wavelengths (248 nm) for the separation and transfer of GaN thin films from sapphire onto Si substrates.14 As further demonstration of this bandgap-selective laser process, we will present the effectiveness of this technique to create free-standing GaN membranes and show that the optical quality of these membranes do not degrade after separation from its sapphire growth substrate.

EXPERIMENTAL Commercially purchased GaN films of thickness 3 µm on a single-sided polished sapphire substrate were used as the starting material. A boron doped, p-type Si (001) wafer was bonded to the surface of the GaN thin film using a low-melting temperature wax forming a sapphire/GaN/wax/Si structure. Prior to the wafer bonding, the backside of the sapphire substrate was polished using diamond paste. All laser processing of the sapphire/GaN/wax/Si structures was performed in air using a Lambda Physik Lextra 200 KrF pulsed excimer laser (38 ns pulse width) with the incident beam directed through the sapphire substrate. The energy density of the laser light was varied between 100 and 600 mJ/cm2 by defocusing the laser beam with a fused silica plano convex lens having a 350 mm focal length. Determination of the temperature rise of the irradiated GaN thin films was done using the one-dimensional heat equation, ∂T α 1 ∂  ∂T  = I ( z, t ) + κ . ∂t ρC p ρC p ∂z  ∂t 

[] 1


I(z,t) is the power density of the incident laser light at a depth z and time t. The variables T, ρ, Cp, κ, and α are the temperature, density, specific heat, thermal conductivity and optical absorption coefficient, respectively, of GaN. Assuming the thin film is a homogeneous absorbing medium, the incident power density can be written as: I(z,t)=I0(t)(1-R)e-αz,

[2]

where I0(t) is the output laser power density and R is the reflectivity. The heat equation can then be solved analytically assuming the thermal and optical parameters are invariant with temperature, the GaN film is of a semi-infinite thickness, and no phase change occurs during the laser pulse. The solution is given as: 15, 16

2 I Dt   z  T ( z, t ) =  0 ierfc  (1 − R )  κ   2 Dt  

[] 3

where D = κ/ρCp. For a single 38 ns pulse from a KrF laser, equation 3 reveals that a laser fluence of approximately 400 mJ/cm2 is needed to raise the surface of the GaN film to the decomposition temperature of 900°C. The values for ρ, Cp, κ, R and α used in the calculations were 6.11 g/cm3, 9.745 cal/mol-K, 1.3 W/cm-K, 0.3 and 4.4×105 cm-1, respectively.17-20 A simulated temperature profile for a single 38 ns, 400 mJ/cm2 pulse from a KrF laser as a function of time and depth is shown in Figure 1. The figure shows the temperature rise, coupled with a large temperature gradient across the thickness of the GaN film, occurs in a highly localized heated area that is within 100 nm below the irradiated GaN/sapphire interface. This selective localized heating allows for the formation of a thin interfacial layer to yield separation of the GaN film from the sapphire substrate.

RESULTS AND DISCUSSION Successful lift-off and transfer of the GaN film from sapphire onto a Si substrate was accomplished after a single 600 mJ/cm2 laser pulse through the transparent sapphire substrate. The actual laser fluence needed for separation agrees well with the calculations since attenuation of the laser beam through the 0.5 mm thick sapphire substrate is approximately 20-30% at 248nm.21 Warming the sample on a hot plate above the Ga Proceedings of the Symposium on LED for Optoelectronic Applications and the 28th State of the Art Programs on Compound Semiconductors 98-2, 377 (1998).

100

time (ns)

80 60

900°C

800°C

700°C

40

600°C

500°C

400°C

300°C 200°C

20

100°C

0

0.5

1.0

1.5

depth (µm) Figure 1: A simulated temperature profile for a single 38 ns, 400 mJ/cm2 pulse from a KrF laser. The temperature rise, coupled with a large temperature gradient across the thickness of the GaN film, occurs in a highly localized heated area that is within 100 nm below the irradiated GaN/sapphire interface (referenced at 0 µm depth).

melting point (Tm = 30°C) completed the separation process. GaN films up to 5 mm × 5 mm were successfully transferred onto 10 mm × 10 mm Si substrates. A thin Ga rich layer on the surface of the now exposed interface was easily removed with a 1:1 solution of HCl and de-ionized water. The films were then characterized by x-ray diffraction and atomic force microscopy (AFM) to verify the structural integrity of the GaN film after lift-off. Free-standing GaN membranes were then fabricated by dissolving the wax bonder in acetone allowing the GaN to float off the supporting Si substrate. The membranes were then transferred onto a supporting frame to allow for characterization by photoluminescence (PL) on either surface of the free-standing GaN. Analysis of the GaN film by x-ray rocking curve measurement was reported in an earlier work.14 The measured full-width at half maximum (FWHM) of the GaN 0002 reflection showed no broadening after separation from the sapphire substrate indicating the film had not suffered mechanical damage during the laser processing. The surface morphology of the exposed GaN interface was characterized by AFM. Figure 2 shows an AFM scan of a 20 µm × 20 µm area of the former GaN interface. The scan shows a Proceedings of the Symposium on LED for Optoelectronic Applications and the 28th State of the Art Programs on Compound Semiconductors 98-2, 377 (1998).

relatively smooth surface morphology with a measured surface roughness of approximately 24 nm (rms).

0

Figure 2: AFM scan of a 20 µm × 20 µm area of the former GaN interface. The scan shows a relatively smooth surface morphology with a measured surface roughness of approximately 24 nm (rms).

Characterization of the optical quality for the separated GaN films was performed using low-temperature (4 K) PL. For these measurements, 3 µm thick GaN membranes were measured and compared to GaN on sapphire. Figure 3 shows the donor-bound exciton (DX) peak of the GaN films before and after separation from the sapphire substrate. The measured FWHM of the DX peaks after separation did not show appreciable broadening indicating no detectable optical degradation of the GaN films. A DX peak red-shift was observed for the GaN membrane compared to the GaN on sapphire. This shift can be interpreted as a release of the compressive biaxial stress on the GaN film, introduced during the growth process,22 when the substrate constraint is removed. The measured red-shift of ~10 meV for the GaN membrane after separation corresponds to a biaxial compressive stress of approximately 0.4 GPa. The resulting peak position of the free-standing membrane matches well with that of stress-free GaN thin films grown on bulk GaN substrates.22


GaN membrane GaN on Sapphire

1.0

4K

3.4660 eV

3.4766 eV

FWHM =

FWHM =

4.8 meV

4.7 meV

Intensity (a.u.)

0.8 0.6 0.4 0.2 0.0 3.45

3.46

3.47

3.48

3.49

Energy (eV) Figure 3: Low-temperature (4 K) PL spectra for GaN/sapphire and GaN membranes. The measured FWHM of the DX peaks after separation did not show appreciable broadening, indicating no detectable optical degradation of the GaN films. The measured red-shift of ~10 meV for the GaN membrane corresponds to a biaxial compressive stress relief of approximately 0.4 GPa after separation of the GaN from sapphire.

CONCLUSION

Lift-off of GaN thin films on sapphire substrates using pulsed excimer laser processing was demonstrated. The process allows for the separation of GaN thin films from transparent sapphire substrates using a KrF laser, directed through the sapphire, to selectively decompose a thin region of the GaN/sapphire interface. In this way, GaN thin films were separated from sapphire substrates and transferred onto Si substrates. The surface morphology of the former GaN interface, characterized by AFM, had a measured roughness of ~24 nm (rms). Free-standing GaN membranes were also created using the laser lift-off process. PL measurements of GaN membranes fabricated using the laser liftoff technique showed no detectable optical degradation of the GaN after lift-off. A 10 meV red-shift in the DX peak of the free-standing film relative to GaN/sapphire suggest that a compressive biaxial stress relief of ~0.4 GPa occurred after the GaN separated from


sapphire. The laser lift-off process demonstrated allows for rapid, simple and relatively inexpensive separation of GaN thin films from sapphire substrates.

ACKNOWLEDGEMENTS

This work is supported in part by the Air Force Office of Scientific Research (AFOSR/JSEP) under contract # FDF49620-97-1-0431-05/00 (442427-23117) and benefited from the use of the UC Berkeley Integrated Materials Laboratory which is sponsored by the National Science Foundation (Grant # DMR-9214370). REFERENCES [1] S. Nakamura, J. Crystal Growth, 170, 11 (1997). [2] S. Nakamura, M. Senoh, N. Iwasa, and S.-I.Nagahama, Appl. Phys. Lett. 67, 1868 (1995). [3] S. Nakamura, M. Senoh, and T. Mukai, Appl. Phys. Lett. 62, 2390 (1993). [4] H. Gotoh, T. Suga, H. Suzuki, and M. Kimata, Jpn. J. Appl. Phys. 20, L545 (1981). [5] R.F. Davis, S. Tanaka, L.B. Rowland, R.S. Kern, Z. Sitar, S.K. Ailey, and C. Wang, J. Crystal Growth, 164 132 (1996). [6] H. Amano, M. Kito, K. Hiramatsu, and I. Akasaki, Jpn. J. Appl. Phys. 28, L21 (1989). [7] Z.A. Munir, and A.W. Searcy, J. Chem. Phys. 42, 4233 (1965). [8] N. Newman, J. Ross, and M. Rubin, Appl. Phys. Lett. 62, 1242 (1993). [9] E. Yablonovitch, T. Gmitter, J.P. Harbison, and R. Bhat, Appl. Phys. Lett. 51, 2222 (1987). [10] E. Yablonovitch, T. Sands, D.M. Hwang, I. Schnitzer, T.J. Gmitter, S.K. Shastry, D.S. Hill, and J.C.C. Fan, Appl. Phys. Lett. 59, 3159 (1991). [11] E. Yablonovitch, K. Kash, T.J. Gmitter, L.T. Florez, J.P. Harbison, and E. Colas, Electronics Letters 25, 171 (1989). [12] T. Detchprohm, H. Amano, K. Hiramatsu, and I. Akasaki, J. Crystal Growth 128, 384 (1993). [13] M.K. Kelly, O. Ambacher, R. Dimitrov, R. Handschuh, and M. Stutzmann, Phys. Stat. Sol., (A) 159, R3 (1997). [14] W.S. Wong, T. Sands, and N.W. Cheung, Appl. Phys. Lett. 72, 599 (1998). [15] P. Baeri, G. Foti, J.M. Poate, and A.G. Cullis, Phys. Rev. Lett. 45, 2036 (1980). [16] H.S. Carslaw, and J.C. Jaeger “Conduction of Heat in Solids” Oxford Univ. Press, London and New York (1959). [17] I. Barin, O. Knacke, and O. Kubaschewski, “Thermophysical Porperties of Inorganic Substances” Springer, Berlin (1977). [18] G.A. Slack, J. Phys. Chem. Solids 38, 330 (1977).


[19] [20] [21] [22]

S. Bloom, G. Harbeke, E. Meier, I.B. Ortenburger, Phys. Status Solidi B 66, 161 (1974). K. Osamua, K. Nakajima, Y. Murakami, Solid State Commun. 11, 617 (1972). I.H. Malitson, JOSA 52, 1377 (1962). C. Kisielowski, J. Krüger, S. Ruvimov, T. Suski, J.W. Ager III, E. Jones, Z. Liliental-Weber, M. Rubin, E.R. Weber, M.D. Bremser, and R.F. Davis, Phys. Rev. BII 54,17745 (1996).
