A Hydrogen-Free Doping Process by Mg and N
2Diffusion
for Fabrication of P-Type GaN
C. Y. Fan (), C. F. Lin. (), M. S. Feng ( )
Institue of Material Science and Engineering, National Chiao Tung University,
Hsinchu 30049, Taiwan, R.O.C.
J. D. Guo ( ), J.S.Tsang( ), W. C. Lai()
National Nano Device Laboratory, Hsinchu 30049, Taiwan, R.O.C.
Abstract:
A hydrogen-free doping process is adapted in this experiment which is revealed by put Mg with the un-doped WZ-GaN[0001] film into the diffusion ampoule. Diffusion process is taken place at 900and 1000for three hours. The 900diffused GaN film has a strong A-exciton line at 355 nm (~3.493 eV) which is the highest value by now, and the 1000 diffused GaN film has a strong A-exciton line at 356 nm (~3.483 eV) which is as high as ever being measured.1. Both
exist free exciton line, I1(exciton bound to the neutral acceptor) line and D-A pair with its LO-phonon replicas . PL
measurement are taken under different temperature and excitation power with a 325.5 nm He-Cd laser. The ex-situ and hydrogen-free doping technique is proven to be valid in our work. And we have the highest A-exciton ( free exciton ) peak frequency than ever being reported. The emission mechanism seem to be different at different temperature. The deep-level-related yellow band emission dominate at 20 K. The emission intensity decay rapidly when the temperature exceeding 160 K.
Introduction:
T he III-V nitride wide band gap semiconductors have become a promising material for the fabrication of
t h e green-blue to the UV region1 However, a classical
MOCVD grows GaN film with NH3 as a nitrogen source,
w h i c h w i l l d i s s o c i ate hydrogen atoms and produce a c c e p t o r -hydrogen complex. So it needs an annealing process to dissociate the acceptor-hydrogen complex and activate the dopant as real acceptor. The annealing process c a n b e r e v e a l e d e i ther by low-energy electron-beam irradiation (LEEBI) treatment or simple thermal annealing
a t a t e m p e r a t u r e a b o v e 5 0 0 .2 , 3 , 4 , 5
To avoid the formation of Mg-H complex, we don’t
dope the GaN film by furnishing Cp2Mg into reaction
chamber. Instead, we dope Mg into GaN film by diffusion
method. The GaN film and Mg and N2 gas are put into a sealed diffusion ampoule, and proceed a constant-temperature diffusion process. The reason for using Mg with N2 atmosphere as a Mg source is to avoid
too much VN produced. The N2 in the diffusion ampoule
is then as the atmosphere inhibiting the dissociation of the GaN and avoid the formation of Mg-H complex.
Experiment :
The diffusion ampoule is made of a quartz tube which was cleaned by 1% HF solution for 2 hour and rinsed bye DI water for 4 hours. Then the tube was dried
by pure N2 purge The un-doped GaN film and the
diffusion source are put into the quartz tube as a whole to form a diffusion ampoule. In this experiment the diffusion
process are taken place in a tube furnace at 900and 1000 for 3 hours. All of the diffusion ampoules suffer an uniform thermal profile, i.e, no thermal gradient occurred. After diffusion, PL spectrum is measured at variant temperature and power. The exciting source is 325.5 nm KIMMON IK series He-Cd CW Laser with a maximum power output of 60mW and 2 mm beam size in diameter. After focusing , the power density can reach up to about
20 W/cm2. The monochrometer is ARC Spectrapro-750.
Thorton EMI PMT and SRS830 Lock-in Amplifier are used to amplify the signal and reduce the noise. CTI-Cryogenic compressor and Model 22 Refrigerator are used to supply a low-temperature environment. Laser power is adjusted by attenuation filter.
Results and Discussion:
3500 3600 3700 3800 3900 4000 4100 4200 4300 4400 4500 0. 0 2.0x10-4 4.0x10-4 6.0x10-4 8.0x10-4 1.0x10-3 3560 3640 3770 3880 3990 (c3) (c2) (c1) (b) (a) (~3.108 eV) (~3. 196 eV) (~3. 289 eV) (~3. 407 eV ) (~3. 483 eV) PL Intens ity (A rb . Un it) Wavelengt h (A )
Fig.1. A PL spectrum measured at 10 K for Mg-diffused GaN film at 1000. (a).The main peak at 356 nm(~3.483 eV) responses to the A-exciton ( free-exciton ) which is as the same as observed by K. Naniwae, e.a. .(b) The 364 nm line represents the exciton bound to the neutral acceptor. (c1~3) . The D-A pair with its three LO-phonon replicas.
Figure 1 is a typical PL spectrum measured at 10 K for this experiment which is obtained . It shows that a sharp peak at 355.0nm (~3.492 eV) which corresponds to the A-exiton which attributed to the decay of the free
excitons6 and 27 meV high than observed than in-situ
doped.7 The 365 nm (~3.397 eV) is attributed by the LO
phonon replica of the A-excition line and a strong no-phonon D-A pair band at 377.0 nm(~3.289 eV) accompanied by a series of phonon replicas.
Figure 2 represents the temperature dependence of the intensity for the A-excition (free-exciton) and exciton bound to neutral acceptor (Ex-A). In the case of diffusing
at 1000,as can be seen in figure 2, when the measuring
temperature increases , the peak intensity tends to decrease except for that at 20 K. We suppose that the luminescence mechanism at 20 K is different with other temperatures, maybe relating to the deep level emission, we can not make sure. We can not read out the yellow band emission signal to a significant value due to the limitation of lens and read-out instrument. But we suppose that the yellow emission is stronger at 20 K.
F ig. 2. Temperature dependence of emission intensity for each peak. Curve (a): The free exciton (355 nm) for sample diffused at 900. Curve (b): The free exciton (356 nm ) for sample diffused at 1000. Curve (C): The exciton bound to neutral acceptor (Ex-A) for sample diffused at 1000. Curve (d): The exciton bound to neutral acceptor (Ex-A) for sample diffused at 900. Note that this is plotted with logarithm scale.
Another interesting phenomenon is its free exciton line decays in a faster rate than the Ex-A. The temperature dependence of the A-exciton is more sensitive than the Ex-A. As the temperature get higher and higher , both the free exciton and the Ex-A diminish.
In the case of diffusing at 900, there exists
different trend for the variation of peaks intensity. It seems that the Ex-A emission is getting to dominate and the free-exciton emission is getting extinguished as the temperature is increasing.
Figure 3 shows the power density dependence of peak position for the free exciton and Ex-A line. As can be seem one the plot, the free exciton keeps at 3.483 eV when the power varies from 0.2 to 0.8. This is because the temperature variation is not very large. If we measured the PL until room temperature, it is believed that the
peak will shift toward the direction of long wavelength9 at
higher temperature.. Eg=3.503+(5.08x10-4T2)/(T-996) (1) (a) (b) ! "#
Fig 3. (a)The excitation power dependence of peaks position for free exciton and Ex-A lines. This plot is obtained by measured the 1000diffused GaN film at 10 K. The power density is relative units. (b) The same but diffusion at 900
and measured at 80 K.
However the Ex-A mode doesn’t show the same
result. It tends to decay as power is increasing and has a minimum value at 0.5 unit in this experiment. This phenomenon is very similar to the Cd-doped GaN made
by VPE8, but the photon energy of free-exciton in Fig3-(b)
is higher than ever reported. The binding energy of D-A pair can be calculated to be ~22.7 meV by the following equation:
EB = E g- ED + EC- Epeak (2)
Figure 4 represents the excitation power dependence of the PL intensity. The free exciton line is increasing as the excitation is increasing. But the Ex-A line has a minimum intensity at 0.4 power unit. It means that the different excitation power may result in different emission mechanism. ! " !#
Fig.4 The power density dependence of PL intensity for the 1000diffused GaN film which is measured at 20 K. Note that this is plotted with logarithm scale.
Conclusions:
An ex-situ doping technique of GaN is adapted in this experiment which is proven to be valid. PL measurement at variant temperatures and excitation power density are made systematically. The results show that the free-exciton emission in our works is higher (3.493 eV at 10 K) than any being reported . The Ex-A emission has lower intensity when the excitation power is at 0.4 power unit. And the deep level related yellow band emission is the main mechanism at 20 K. The diffusion
occured at 900and 1000 both have a D-A pair binding energy of 22.8 meV.
Acknowledgement:
This work is supported by the Naitional Science Council of Republic of China, under Contract number NSC-88-2218-E-009-051.
References
1 S. Nakamura, S. Senoh, N. Iwasa, and S. Nagahama, Jpn. J. Appl. Phys. 34,L797 (1995).
2 S. Nakamura, N. Iwasa, M. Senoh, and T.Mukai, Jpn. J. Appl. Phys., Part1 31, 1258 (1992).
3 H. Amano, M. Kito, K. Hiramatsu, and I. Akasaki, Jpn. J. Appl. Phys., Part 2 28, L2112 (1989).
4 S. Nakamura, T. Mukai, M. Senoh, and N. Iwasa, Jpn. J. Appl. Phys., Part 2 31, L139 (1992).
5 Y. Li, Y. Lu, H. Shen, M. Wraback, C.-Y. Hwang, M. Schurman, W. Mayo, T. Salagaj, and R. A. Stall, Mater. Res. Soc. Symp. Proc. 395, 369(1996).
6. K. Naniwae, S. Itoh, H. Amano, K. Itoh, K. Hiramatsu, I. Akasaki, J. Cryst. Growth (Netherlands) 99 381-4 (1990) 7 M. Ilegems and R. Dingle, J. Appl. Phys. 44, 4234 (1973). 8 O. Lagerstedt and B. Monemar, J. Appl. Phys. 45, 2266
(1973).
9.B. Monemar [ Phys. Status Solidi A (Germany) vol.6 (1975) p.445-8]
