APPLIED PHYSICS LETTERS
VOLUME 78, NUMBER 13
26 MARCH 2001
Effect of alcohol-based sulfur treatment on Pt Ohmic contacts
to p-type GaN
Chul Huh, Sang-Woo Kim, Hyun-Min Kim, Dong-Joon Kim,
and Seong-Ju Parka)
Department of Materials Science and Engineering and Center for Optoelectronic Materials Research,
Kwangju Institute of Science and Technology (K-JIST), Kwangju 500-712, Korea
共Received 7 December 2000; accepted for publication 29 January 2001兲
The effects of an alcohol-based (NH4 ) 2 S solution 关 t-C4H9OH⫹共NH4兲2S兴 treatment on Pt Ohmic
contacts to p-type GaN are presented. The specific contact resistance decreased by three orders of
magnitude from 2.56⫻10⫺2 to 4.71⫻10⫺5 ⍀ cm2 as a result of surface treatment using an
alcohol-based (NH4 ) 2 S solution compared to that of the untreated sample. The O 1s and Pt 4 f
core-level peaks in the x-ray photoemission spectra showed that the alcohol-based (NH4 ) 2 S
treatment was effective in removing of the surface oxide layer. Compared to the untreated sample,
the alcohol-based (NH4 ) 2 S-treated sample showed a Ga 2p core-level peak which was shifted
toward the valence-band edge by 0.25 eV, indicating that the surface Fermi level was shifted toward
the valence-band edge. These results suggest that the surface barrier height for hole injection from
Pt metal to p-type GaN can be lowered by the surface treatment, thus resulting in a drastic reduction
in specific contact resistance. © 2001 American Institute of Physics. 关DOI: 10.1063/1.1358356兴
Rapid progress has recently been achieved in obtaining
high-brightness GaN-based light-emitting diodes 共LEDs兲 in
the visible and ultraviolet wavelength range.1 For the case of
the highly efficient LED, several critical issues must be addressed, such as high-quality crystal growth, low-resistance
Ohmic contact, and etching–sawing–cleaving for device
processing. Low-resistance Ohmic contacts to p-type GaN
are particularly crucial in improving the performance of optoelectronic devices as well as electronic devices.2–4
The growth of heavily doped p-GaN (⬎1018 cm⫺3) and
the absence of contact metals, which have a work function
larger than that of p-GaN 共the sum of the electron affinity of
4.1 eV and the band gap of 3.4 eV兲,5 have been realized as
major obstacles in the achievement of low-resistance Ohmic
contacts to p-GaN. To obtain a low-resistance Ohmic contact to p-type GaN, a variety of surface treatments of p-GaN
using solutions such as KOH,6 aqua-regia solution,7 and
(NH4 ) 2 Sx , 8,9 have been proposed. It has been reported that
the surface treatment with a (NH4 ) 2 Sx solution is particularly efficient in removing the native oxide on the p-type
GaN surface, resulting in a low-resistance Ohmic contact to
p-type GaN.9
In the area of GaAs devices, several studies have reported that the alcohol-based (NH4 ) 2 Sx solution removes the
native oxide on the GaAs surface more efficiently than the
normal (NH4 ) 2 Sx solution due to the low dielectric constant
of the alcohol-based (NH4 ) 2 Sx solution, resulting in a larger
improvement in device characteristics.10,11 Zhilyaev et al.12
have reported that the photoluminescence intensity of n-type
GaN is considerably enhanced as a result of the surface treatment with an alcoholic sulfide solution. Our previous
results13 also showed that, for the case of n-type GaN, the
surface treatment with an alcohol-based (NH4 ) 2 S solution
removes the insulating layer on the n-type GaN surface more
effectively than the normal, aqueous (NH4 ) 2 S solution, leada兲
Electronic mail: [email protected]
ing to more enhanced electrical properties. In this letter, we
report on an investigation of the effect of an alcohol-based
(NH4 ) 2 S solution 关 t-C4H9OH⫹共NH4兲2S兴 for use in the treatment of Ohmic contacts to p-type GaN by observing the
oxide layer and the shift of the surface Fermi level, as evidenced by the x-ray photoelectron spectroscopy 共XPS兲 spectra. In addition, the current–voltage (I – V) result showed
that the specific contact resistivity for a sample treated with
an alcohol-based (NH4 ) 2 S solution is drastically decreased
by three orders of magnitude over that of an untreated
sample.
The p-GaN samples were grown on a c-face 共0001兲 sapphire substrate by using a metalorganic chemical-vapor
deposition system 共Emcore DG125™兲. A 30-nm-thick GaN
buffer layer was grown on the substrate at 500 °C, followed
by the growth of a 1-␮m-thick Mg-doped p-type GaN epitaxial film at a temperature of 1020 °C. Hall measurements
showed that the bulk hole concentration was about 2
⫻1017 cm⫺3. The GaN layer was degreased by treatment
with trichloroethylene, acetone, and methanol for 5 min at
each step, and then rinsed in deionized 共DI兲 water for 10
min. Prior to the fabrication of the transfer length method
共TLM兲 patterns, mesa structures were patterned by inductively coupled plasma etching 共Oxford Plasma 100兲 using
Cl2 /CH4 /H2 /Ar and the TLM patterns were defined by standard photolithographic techniques. The size of the pads was
100⫻200 ␮ m2 and the spacings between the pads were 5,
10, 15, 20, 25, and 35 ␮m. The samples were first etched in
HCl:H2O(1:1) solution for 1 min and were then dipped into
a 60 °C alcoholic solution for 10 min, which was composed
of 90 vol % tert-butanol (t-C4H9OH) and 10 vol %
(NH4 ) 2 S. The residual solution was removed from the
sample by rinsing the samples in DI water for 1 min. After
the sulfur treatment, the samples were immediately placed in
a vacuum chamber for metal deposition. The Pt 共8 nm兲 films
were then deposited on p-type GaN by electron-beam evaporation 共PLS 500 model兲. I – V data were measured at room
0003-6951/2001/78(13)/1942/3/$18.00
1942
© 2001 American Institute of Physics
Downloaded 22 Nov 2010 to 115.145.199.16. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
Appl. Phys. Lett., Vol. 78, No. 13, 26 March 2001
Huh et al.
1943
FIG. 1. I – V characteristics of untreated and treated samples with an
alcohol-based (NH4 ) 2 S solution.
temperature using a parameter analyzer 共HP4155A兲. To characterize the chemical bonding at the interface between the Pt
layer and sulfur-treated p-type GaN, XPS was performed using the Mg K ␣ line 共1253.6 eV兲 as an excitation source in a
ultra-high-vacuum system with a base pressure of ⬃1
⫻10⫺10 Torr.
Figure 1 shows the I – V characteristics of the samples
which had been untreated and treated with alcohol-based
(NH4 ) 2 S solution. While the I – V curves of the untreated
sample exhibits a nonlinear I – V behavior, the I – V curve of
the alcohol-based (NH4 ) 2 S solution treated sample shows a
linear behavior. Specific contact resistances for the samples
were determined from a plot of the measured resistance versus the spacings between the TLM pads. The specific contact
resistance was determined to be 2.56⫻10⫺2 ⍀ cm2 for the
untreated samples and 4.71⫻10⫺5 ⍀ cm2 for the alcoholbased (NH4 ) 2 S treated samples, respectively. Compared to
the untreated sample, the specific contact resistance decreased by three orders of magnitude as a result of the surface treatment using the alcohol-based (NH4 ) 2 S solution.
To investigate the mechanisms for the improvement of
the Ohmic characteristics due to this surface treatment, XPS
spectra were obtained from the Pt contacts 共2.5 nm兲 on the
sulfur-treated p-type GaN. Figure 2 shows the O 1s core
level of the Pt/p-GaN interface for untreated and alcoholbased (NH4 ) 2 S-treated samples, respectively. As shown in
Fig. 2, the O 1s core-level peak for samples treated with
alcohol-based (NH4 ) 2 S solution were significantly decreased
compared to that of the untreated sample. This result shows
that the oxide layer was efficiently removed from the p-GaN
surface by the surface treatment with alcohol-based (NH4 ) 2 S
solution. It has been reported that an interfacial oxide layer
increases the barrier height for hole injection from metal to
p-type GaN, leading to a detrimental effect on Ohmic contacts to p-type GaN.14 Hence, the improved Ohmic characteristics for samples treated with the alcohol-based 共NH4兲2S
solution can be attributed to the removal of the interfacial
oxide layer.
Figure 3 shows the Pt 4 f core level of Pt deposited on
the p-GaN for two samples. As shown in Fig. 3, the Pt 4 f
core level of the alcohol-based 共NH4兲2S-treated sample 关Fig.
3共b兲兴 shifts toward the lower-energy side compared to that of
untreated sample 关Fig. 3共a兲兴. The binding energy of the
FIG. 2. O 1s core-level spectra of the Pt/p-GaN interface for 共a兲 an untreated sample and 共b兲 an alcohol-based (NH4 ) 2 S-treated sample.
Pt 4 f 7/2 core-level peak for Pt, which is known to be around
71 eV, is smaller than that of a core-level peak for the Pt
oxide.15 Therefore, the shift of the Pt 4 f core level to the
lower-binding-energy side is due to the changes in the
chemical bonding state of Pt at the interface, as a result of
the alcohol-based 共NH4兲2S treatment. These results are in
agreement with the O1s core level behavior, as shown in
Fig. 2.
The O 1s and Pt 4 f core-level spectra show that an
alcohol-based 共NH4兲2S treatment is very effective in removing the surface oxide layer. However, it has been reported
that the removal of the surface oxide layer itself does not
guarantee a drastic reduction in the metal contact
resistance.16 An alcohol-based 共NH4兲2S treatment can also
alter the surface electronic properties, and hence, the surface
barrier height, which is closely related to the subsequent
FIG. 3. Pt 4 f core-level spectra of the Pt/p-GaN interface for 共a兲 an untreated sample and 共b兲 an alcohol-based (NH4 ) 2 S-treated sample.
Downloaded 22 Nov 2010 to 115.145.199.16. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
1944
Huh et al.
Appl. Phys. Lett., Vol. 78, No. 13, 26 March 2001
FIG. 4. Ga 2p core-level spectra of the Pt/p-GaN interface for 共a兲 an
untreated sample and 共b兲 an alcohol-based (NH4 ) 2 S-treated sample.
metal contact behavior. It is known that the energy position
of the Ga-based core-level peaks is one of the most accurate
means of determining the shifts in the GaN surface Fermilevel position.6 Therefore, the surface Fermi-level position
was determined by observing the Ga 2p core-level of an
alcohol-based 共NH4兲2S-treated sample. The Ga 2p core-level
spectra of the Pt/p-GaN interface of the untreated and an
alcohol-based 共NH4兲2S-treated sample are shown in Fig. 4.
Figure 4 shows that the Ga 2p core-level peak of the untreated sample 关Fig. 4共a兲兴 shifts toward the lower-energy side
by 0.25 eV compared to that of an alcohol-based
共NH4兲2S-treated sample 关Fig. 4共b兲兴. This indicates that a
small surface barrier height for hole injection from metal to
p-type GaN can further reduce the specific contact resistance,
as shown in Fig. 1.
It is well known that for GaAs, the sulfur passivation in
a S-containing solution is essentially an etching process, during which a chemical reaction and dissolution alternately
take place.17 A similar mechanism can be proposed for the
alcohol-based sulfur passivation of p-type GaN. The native
oxide layer on the GaN surface can be initially removed
through the formation of gallium hydroxide, which is produced via the OH⫺ groups present in the solution when the
GaN sample is dipped in the alcohol-based sulfur solution.
The selective removal of surface Ga via gallium hydroxide in
a sulfur solution can lead to the formation of Ga vacancies. It
is well known that the Ga vacancies are able to act as acceptors in p-type GaN.9 Hence, the surface reactions of GaN
with the sulfur solution can result in the formation of
acceptor-like Ga vacancies, leading to a shift of the surface
Fermi level toward the valence-band edge, as shown in the
Ga 2p core-level spectra Fig. 4. A previous study18 reported
that sulfur does not bond with nitrogen but bonds with gallium or occupies nitrogen-related vacancies. The S ions in
the solution would then bond to the Ga atoms on the GaN
surface, producing Ga sulfides. For the case of passivation of
a GaAs surface by using 共NH4兲2S solution, it has been reported that the sulfides, which are formed on the GaAs surface, are soluble in the sulfur solution.19 Hence, the simulta-
neous process of reaction and dissolution can leave a very
thin sulfide layer on the sulfur-treated GaN surfaces. This
very thin sulfide layer can play an important role in preventing the formation of oxide on the GaN surface on exposure
to air prior to the metal deposition. These processes can lead
to the effective removal of surface oxide, formation of Ga
vacancies, and protection from the formation of the native
oxide on exposure to air prior to metal deposition, thus resulting in enhanced Ohmic characteristics for the samples
treated by an alcohol-based 共NH4兲2S solution.
In summary, we report on an investigation of the effect
of an alcohol-based 共NH4兲2S solution treatment on Ohmic
contacts to p-type GaN. Compared to the untreated sample,
the specific contact resistance was drastically decreased by
three orders of magnitude as a result of the surface treatment.
The O 1s and Pt 4 f core-level spectra showed that the
alcohol-based 共NH4兲2S treatment is very effective in the removal of the surface oxide layer. The Ga 2p core-level peak
for the alcohol-based 共NH4兲2S-treated sample showed a shift
by 0.25 eV toward the valence-band edge compared to that
for the untreated sample. The drastic improvement in the
Ohmic characteristics of the alcohol-based 共NH4兲2S-treated
sample can be attributed to the effective removal of the surface oxide and the shift of the surface Fermi level toward the
valence-band edge.
This work was supported by grants from the Ministry of
Science and Technology in Korea through the Critical
Technology-21 Program 共No. 98-N5-01-01-A-08兲 and by the
Brain Korea 21 Project.
1
S. Nakamura, M. Senoh, N. Iwasa, S. Nagahama, T. Yamada, and T.
Mukai, Jpn. J. Appl. Phys., Part 2 34, L1332 共1995兲.
2
J.-K. Ho, C.-S. Jong, C. C. Chiu, C.-N. Huang, K.-K. Shih, L.-C. Chen,
F.-R. Chen, and J.-J. Kai, J. Appl. Phys. 86, 4491 共1999兲.
3
L. Zhou, W. Lanford, A. T. Ping, I. Adesida, J. W. Yang, and A. Khan,
Appl. Phys. Lett. 76, 3451 共2000兲.
4
J.-S. Jang, S.-J. Park, and T.-Y. Seong, Appl. Phys. Lett. 76, 2898 共2000兲.
5
V. M. Bermudez, J. Appl. Phys. 80, 1190 共1996兲.
6
J. Sun, K. A. Rickert, J. M. Redwing, A. B. Ellis, F. J. Himpsel, and T. F.
Kuech, Appl. Phys. Lett. 76, 415 共2000兲.
7
J. K. Kim, J.-L. Lee, J. W. Lee, H. E. Shin, Y. J. Park, and T. Kim, Appl.
Phys. Lett. 73, 2953 共1998兲.
8
J.-S. Jang, S.-J. Park, and T.-Y. Seong, J. Vac. Sci. Technol. B 17, 2667
共1999兲.
9
J.-L. Lee, J. K. Kim, J. W. Lee, Y. J. Park, and T. Kim, Electrochem.
Solid-State Lett. 3, 53 共2000兲.
10
V. N. Bessolov, M. V. Lebedev, and D. R. T. Zahn, J. Appl. Phys. 82,
2640 共1997兲.
11
V. N. Bessolov, A. F. Ivankov, E. V. Konenkova, and M. V. Lebedev,
Tech. Phys. Lett. 21, 20 共1995兲.
12
Y. V. Zhilyaev, M. E. Kompan, E. V. Konenkova, and S. D. Raevskii,
MRS Internet J. Nitride Semicond. Res. 4S1, G6.14 共1999兲.
13
C. Huh, S.-W. Kim, H.-S. Kim, I.-H. Lee, and S.-J. Park, J. Appl. Phys.
87, 4591 共2000兲.
14
X. A. Cao, S. J. Pearton, G. Dang, A. P. Zhang, F. Ren, and J. M. V.
Hove, Appl. Phys. Lett. 75, 4130 共1999兲.
15
J. Chastain and R. C. King, Jr., Handbook of X-Ray Photoelectron Spectroscopy 共Physical Electronics, Eden Prairie, MN, 1995兲.
16
H. Ishikawa, S. Kobayashi, Y. Koide, S. Yamasaki, S. Nagai, J. Umezaki,
M. Koike, and M. Murakami, J. Appl. Phys. 81, 1315 共1997兲.
17
Z. L. Yuan, X. M. Ding, H. T. Hu, Z. S. Li, J. S. Yang, X. Y. Miao, X. Y.
Chen, X. A. Cao, X. Y. Hou, E. D. Lu, S. H. Xu, P. S. Xu, and X. Y.
Zhang, Appl. Phys. Lett. 71, 3081 共1997兲.
18
Y.-J. Lin, C.-D. Tsai, Y.-T. Lyu, and C.-T. Lee, Appl. Phys. Lett. 77, 687
共2000兲.
19
Handbook of Chemistry and Physics, 64th ed., edited by R. C. Weast,
M. J. Astle, and W. H. Beyer 共CRC, New York, 1985兲.
Downloaded 22 Nov 2010 to 115.145.199.16. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions