Half-cycle atomic layer deposition reaction studies of Al2O 3 on (N H4)2S passivated GaAs(100) surfaces

Article abstract of DOI:10.1063/1.3033404

Half-cycle atomic layer deposition reactions of trimethyl aluminum (TMA) and water on GaAs exposed to wet chemical sulfur treatments are studied for the formation of Al₂O₃. Trivalent oxides of gallium and arsenic are completely reduced following the first TMA pulse. The same processing step also removes As-S bonding below the level of detection, while the relative concentration of gallium suboxides as well as Ga-S bonds is not affected. A concomitant decrease in the S 2p peak intensity is observed, indicating that sulfur is lost through a volatile reaction product. Further precursor exposures do not measurably affect substrate surface chemistry.

Full text of DOI:10.1063/1.3033404

Half-cycle atomic layer deposition reaction studies of
surfaces
on
(100)
M. Milojevic, F. S. Aguirre-Tostado, C. L. Hinkle, H. C. Kim, E. M. Vogel, J. Kim, and R. M. Wallace
Citation: Appl. Phys. Lett. 93, 202902 (2008); doi: 10.1063/1.3033404
View online: http://dx.doi.org/10.1063/1.3033404
View Table of Contents: http://aip.scitation.org/toc/apl/93/20
Published by the American Institute of Physics

APPLIED PHYSICS LETTERS 93, 202902 ͑2008͒
Half-cycle atomic layer deposition reaction studies of Al O3
2
on In Ga As „100… surfaces
0
.2
0.8
a͒
M. Milojevic, F. S. Aguirre-Tostado, C. L. Hinkle, H. C. Kim, E. M. Vogel, J. Kim, and
b͒
R. M. Wallace
Department of Materials Science and Engineering, University of Texas at Dallas,
Richardson, Texas 75083, USA
͑
Received 16 June 2008; accepted 29 October 2008; published online 19 November 2008͒
The reduction in III–V interfacial oxides by atomic layer deposition of Al O on InGaAs is studied
2
3
by interrupting the deposition following individual trimethyl aluminum ͑TMA͒ and water steps ͑half
cycles͒ and interrogation of the resultant surface reactions using in situ monochromatic x-ray
photoelectron spectroscopy ͑XPS͒. TMA is found to reduce the interfacial oxides during the initial
exposure. Concentrations of Ga oxide on the surface processed at 300 °C are reduced to a
concentration on the order of a monolayer, while AsO species are below the level of detection of
x
XPS. © 2008 American Institute of Physics. ͓DOI: 10.1063/1.3033404͔
1
7
Engineering the chemistry at the substrate/dielectric
interface is critical to the manufacturing of a high perfor-
mance metal-oxide-semiconductor ͑MOS͒ device on III–V
and a DIW rinse. Samples were then mounted and intro-
duced into an ultrahigh vacuum ͑UHV͒ system in less than
10 min. A dual chamber ALD reactor integrated to an UHV
multitechnique deposition/characterization system was used
for this in situ study. TMA was used as the Al precursor
for subsequent Al O formation using water deposition
chemistry. Prior to the first ALD deposition, the unloaded
reactor was baked at 400 °C and exposed to 300 cycles ͓1
cycleϭ0.1 s ͑TMA/N ͒+4 s N +0.1 s ͑H O/N ͒+4 s N ͔
in order to encourage reproducible reaction behavior and re-
duce any spurious wall deposition/reaction effects due to re-
sidual precursor components from previous depositions.
High purity ͑99.999%͒ N is used as the purging gas. Two
1
8
1
–3
substrates. Controlling the oxidation state and the relative
amount of Ga–O and As–O have been shown recently to be
directly related to the accumulation capacitance frequency
1
9
2
3
4
dispersion phenomenon. A Si-interfacial passivation layer
5
–8
͑
IPL͒ can be used to control surface oxidation, although it
is currently unclear if this compromises some of the perfor-
2
2
2
2
2
9
,10
mance potential of the MOS device.
Ideally the dielectric
deposition process itself, combined with appropriate surface
preparation, might remove the need for an IPL. In a previous
atomic layer deposition ͑ALD͒ study, it has been reported
that both precursor and oxidation state dependent reactions
2
0
2
separate sets of film depositions were performed at sample/
reactor temperatures of 200 and 300 °C. The ALD chamber
pressure was ϳ7.6 Torr during deposition, and was con-
nected to the sample analysis module through a UHV trans-
port tube by a buffer chamber. The pressure of the UHV
transport tube during wafer transfer between half-cycles was
Ͻ2ϫ10− mbar, enabling surface characterization without
spurious ex situ atmospheric exposure or contamination from
the vacuum as determined by control experiments ͑not
shown͒. Analysis of the deposited films was done using an
in situ monochromatic XPS using an Al K␣ x-ray source
͑1486.7eV͒ ͑Ref. 21͒ with a linewidth of ϳ0.25 eV and pass
11
effect interfacial oxide formation on GaAs. This “clean-up”
effect indicates that the interfacial bonding type can be con-
trolled by varying the deposition conditions, as seen in pre-
1
2–14
vious ex situ studies on GaAs.
of the InGaAs interface has reported similar clean-up effects
Previous ex situ analysis
1
5
10
for HfO deposited by ALD. In this work, we examine the
2
ALD dielectric deposition process in detail by interrupting
the growth cycle ͑referred to here as “half cycles” of the
ALD process͒ with in situ monochromatic x-ray photoelec-
tron spectroscopy ͑XPS͒ analysis to characterize the surface
reactions associated with the ALD of trimethyl aluminum
2
2
͑
TMA͒ on In Ga As.
energy of 15 eV. XPS data were taken following each half
0
.2
0.8
The samples used in this study were n-type ͑Si,
cycle ͑TMA/N +XPS, then H O/N +XPS͒. The take-off
angle from the substrate surface was 45°, with an analyzer
acceptance angle of 16°. Binding energy calibration is per-
2
2
2
1
8
3
16
1
ϫ10 /cm ͒ epitaxial
In Ga As. The In Ga As
0
.2
0.8
0.2
0.8
structure was grown by starting with a semi-insulating GaAs
2
3
substrate on top of which a 150 nm buffer layer of n-GaAs
formed regularly to accepted procedures.
1
7
3
͑
Si, 4ϫ10 /cm ͒ followed by growth of a 130 nm layer of
Figure 1͑a͒ shows the Al 2p feature, which indicates
fully oxidized Al after each TMA pulse, and is consistent
with the O 1s peak in Fig. 1͑b͒, which compares the O 1s
peak before and after the first TMA pulse at 300 °C. The
primary components prior to TMA exposure are attributed to
1
7
3
n-GaAs ͑Si, 1ϫ10 /cm ͒. Finally a 13.5 nm thick active
layer of In Ga As was grown with a 1:4 In to Ga ratio
0
.2
0.8
confirmed by XPS.
Samples were initially degreased for 1 min in each
of acetone, methanol, and isopropyl alcohol. The native
oxides were etched using HCl:de-ionized water ͑DIW͒ ͑1:1͒
for 10 min followed by a rinse in flowing DIW for Ͻ10 s.
The final treatment involved a 10 min dip in ͑NH ͒ S ͑22%͒
2
4,25
As–O, Ga–O, and In–O bonding.
Importantly, the total O
1s peak intensity does not change following the first TMA
pulse, indicating that the Al is oxidized at the expense of
the interfacial substrate oxides. Both remaining Ga–O and
In–O bonding features are reduced in intensity and appear
at a lower binding energy ͑ϳ530.8 eV͒ than Al–O bonds
4
2
a͒
Electronic mail: jiyoung.kim@utdallas.edu.
Electronic mail: rmwallace@utdallas.edu.
b͒
26
͑ϳ531.3 eV͒. Therefore as these surface oxides are re-
0003-6951/2008/93͑20͒/202902/3/$23.00
93, 202902-1
© 2008 American Institute of Physics

202902-2
Milojevic et al.
Appl. Phys. Lett. 93, 202902 ͑2008͒
FIG. 1. ͑Color online͒ ͑a͒ Al 2p region following the first ͑TMA1, DIW1͒
and second ͑TMA2, DIW2͒ half-cycle exposures as well as after ten full
cycles ͑1 nm͒ at 300 °C on In0.2Ga0.8As. ͑b͒ O 1s peak before and after the
first TMA pulse showing a shift toward a higher binding energy but no
change in total area.
duced and Al–O bond formation ensues, the O 1s peak ap-
pears to shift to higher binding energy, with possible Al–OH
2
7
bond formation as well. The C 1s intensity ͑not shown͒ is
near the limit of detection and appears unaffected by the
subsequent ALD cycles, suggesting that the C incorporated
originates with the TMA precursor.
Figure 2 shows the Ga 2p3/2 peak following surface
preparation, after the first TMA pulse, and further after a
FIG. 3. ͑Color online͒ As 2p3/2 spectra showing the removal of arsenic
oxides following the first four half cycles and the subsequent spectra upon
completion of the 1 nm ALD Al O film at 200 and 300 °C.
1
nm thick Al O film ͑ten full cycles͒ is grown. We note
2 3
2
3
that the surfaces prepared for the subsequent 200 °C ALD
experiments have slightly higher initial oxide content at the
surface and are attributed to surface preparation variations.
The higher binding energy peak ͑ϳ1118.4 eV͒ corresponds
to Ga3+ bonding while the bulk Ga–As bonding is contained
in the primary peak at ͑ϳ1117.3 eV͒. The position of the
Ga–O suboxide peak ͑ϳ1117.9 eV͒ has been derived by
2
8
29–31
Hinkle et al. from detailed photoelectron studies.
A
peak corresponding to a Ga–S bond is difficult to resolve
3
2
from this suboxide peak. After the first TMA exposure, the
3
+
Ga oxidation state is dramatically reduced for both depo-
sition temperatures. This indicates that the initial metal pre-
cursor TMA pulse is primarily responsible for the clean up of
interfacial oxides. The area ratio of residual gallium oxides
to the Ga–As is found to be 0.15, which is calculated by
taking into account the escape depth of a 360 eV electron to
correspond to signal on the order of a monolayer.
We also compare the As 2p3/2 features ͑Fig. 3͒, where
the surface-sensitive high binding energy peak is attributed
3
+
33
to an As oxidation state ͑ϳ1326 eV͒, As–S bonding
͑
2
4,32
ϳ1324.5 eV͒, and As–As bonding ͑ϳ1323.5 eV͒.
The
first TMA pulse is effective at reducing arsenic oxide and
sulfide bonding below the level of detection of XPS at
3
00 °C but not at 200 °C. This behavior likely stems from
3
4
the relatively weak bonding and as well as the relatively
high vapor pressure of elemental As. Since the native ox-
ides of III–V compounds may be thought of as a glassy net-
work connected by oxygen atoms, increasing the tempera-
ture from 200 to 300 °C would aid in destabilizing the entire
structure, even though As–O bond scission dominates at such
deposition temperatures.
3
5
3
0
The tendency of reducing agents such TMA to preferen-
tially attack trivalent oxides is confirmed by the dramatic
reduction in trivalent gallium and arsenic oxides and the con-
stant Ga–O/Ga–As intensity ratio with growth. The In 3d
peaks ͑not shown͒ show much lower levels of initial oxida-
FIG. 2. ͑Color online͒ Ga 2p3/2 spectra showing the reduction in gallium
oxides following the first four half cycles and the subsequent spectra upon
completion of the 1 nm ALD Al O film at 200 and 300 °C.
2
3

202902-3
Milojevic et al.
Appl. Phys. Lett. 93, 202902 ͑2008͒
1
1
C. L. Hinkle, A. M. Sonnet, E. M. Vogel, S. McDonnell, G. J. Hughes, M.
Milojevic, B. Lee, F. S. Aguirre-Tostado, K. J. Choi, H. C. Kim, J. Kim,
and R. M. Wallace, Appl. Phys. Lett. 92, 071901 ͑2008͒.
P. D. Ye, G. D. Wilk, B. Yang, J. Kwo, S. N. G. Chu, S. Nakahara, H.-J.
L. Gossmann, J. P. Mannaerts, M. Hong, K. K. Ng, and J. Bude, Appl.
Phys. Lett. 83, 180 ͑2003͒.
M. L. Huang, Y. C. Chang, C. H. Chang, Y. J. Lee, P. Chang, J. Kwo, T.
B. Wu, and M. Hong, Appl. Phys. Lett. 87, 252104 ͑2005͒.
M. M. Frank, G. D. Wilk, D. Starodub, T. Gustafsson, E. Garfunkel,
Y. J. Chabal, J. Grazul, and D. A. Muller, Appl. Phys. Lett. 86, 152904
tion than the Ga 2p spectra. It is reasonable to expect that the
extent of interfacial oxide reduction depends on temperature
as well as the surface concentration of reducing species ͑i.e.,
TMA͒. It has been shown that the surface concentration of
adsorbed TMA is inversely proportional to the substrate
temperature, which would cause a diminished clean-up ef-
fect at higher substrate temperatures due to a lower surface
concentration of TMA. We observe the opposite, however,
which suggests that the kinetics of the reaction depend pri-
marily on temperature.
In conclusion, the results show that for TMA/water
based alumina ALD, the first TMA half cycle pulse is re-
sponsible for the interfacial oxide reduction. At 300 °C,
As–O bonding is reduced below the XPS level of detection
and Ga–O bonding is reduced to the order of monolayer
concentration at the interface demonstrating a sharp
substrate/oxide interface. Similar measurements of the ALD
reactions performed at 200 °C indicate that the clean-up ef-
fect was less effective.
12
13
3
6
1
4
15͑
2005͒.
C. H. Chang, Y. K. Chiou, Y. C. Chang, K. Y. Lee, T. D. Lin, T. B. Wu, M.
Hong, and J. Kwo, Appl. Phys. Lett. 89, 242911 ͑2006͒.
Intelligent Epitaxy Technologies, http://www.intelliepi.com.
M. H. Zhang, M. Oye, B. Cobb, F. Zhu, H. S. Kim, I. J. Ok, J. Hurst, S.
Lewis, A. Holmes, J. C. Lee, S. Koveshnikov, W. Tsai, M. Yakimov, V.
Torkanov, and S. Oktyabrsky, J. Appl. Phys. 101, 034103 ͑2007͒.
Picosun Oy ͑http://www.picosun.com͒.
P. Sivasubramani, J. Kim, M. J. Kim, B. E. Gnade, and R. M. Wallace,
J. Appl. Phys. 101, 114108 ͑2007͒.
The TMA pulse partial pressure is 300 mbar.
J. Schweppe, R. D. Deslattes, T. Mooney, and C. J. Powell, J. Electron
Spectrosc. Relat. Phenom. 67, 463 ͑1994͒.
1
1
6
7
18
1
9
2
2
0
1
2
2
2
3
Omicron Nanotechnology, Gmbh ͑http://www.omicron.de͒.
The authors acknowledge the useful discussions with
Dr. M. Passlack. This work is supported by the FCRP on
Materials, Structures, and Devices, the National Institute of
Standards and Technology, Semiconductor Electronics Divi-
sion, System IC 2010 ͑funded by MKE in Korea͒, and the
Texas Enterprise Fund.
“
Standard practice for calibration of the electron binding-energy scale
of an x-ray photoelectron spectrometer,” ASTM Standard Report No.
E2108-05, ASTM International, West Conshohocken, PA, www.astm.org;
C. J. Powell, Surf. Interface Anal. 23, 121 ͑1995͒; M. P. Seah, I. S.
Gilmore, and G. Beamson, ibid. 26, 642 ͑1998͒.
M. V. Lebedev, D. Ensling, R. Hunger, T. Mayer, and W. Jaegermann,
Appl. Surf. Sci. 229, 226 ͑2004͒.
F. S. Aguirre-Tostado, M. Milojevic, C. L. Hinkle, E. M. Vogel, R. M.
Wallace, S. McDonnell, and G. J. Hughes, Appl. Phys. Lett. 92, 171906
2008͒.
J. F. Moulder, W. F. Stickle, P. E. Sobol, and K. D. Bomben, Handbook of
X-ray Photoelectron Spectroscopy ͑Physical Electronics, Eden Prairie,
MN, 1995͒, pp. 230–232.
B. R. Strohmeier, J. Vac. Sci. Technol. A 7, 3238 ͑1989͒.
C. L. Hinkle, M. Milojevic, B. Brennan, A. M. Sonnet, F. S. Aguirre-
Tostado, G. J. Hughes, E. M. Vogel, and R. M. Wallace, “Detection of Ga
suboxides and their impact on III-V passivation and Fermi-level pinning”
͑unpublished͒.
2
2
4
5
1
T. Ito and Y. Sakai, Solid-State Electron. 17, 751 ͑1974͒.
2
M. Passlack, in Materials Fundamentals of Gate Dielectrics, edited by
26͑
A. A. Demkov and A. Navrotsky ͑Springer, Berlin, 2005͒, pp. 403–467.
M. Hong, J. R. Kwo, P. Tsai, Y. Chang, M. Huang, C. Chen, and T. Lin,
3
Jpn. J. Appl. Phys., Part 1 46, 3167 ͑2007͒.
C. L. Hinkle, A. M. Sonnet, E. M. Vogel, S. McDonnell, G. J. Hughes, M.
4
2
2
7
8
Milojevic, B. Lee, F. S. Aguirre-Tostado, K. J. Choi, J. Kim, and R. M.
Wallace, Appl. Phys. Lett. 91, 163512 ͑2007͒.
S. Tiwari, S. L. Wright, and J. Batey, IEEE Electron Device Lett. 9, 488
5
6͑
1988͒.
J. L. Freeouf, D. A. Buchanan, S. L. Wright, T. N. Jackson, J. Batey, B.
Robinson, A. Callegari, A. Paccagnella, and J. M. Woodall, J. Vac. Sci.
Technol. B 8, 860 ͑1990͒; J. L. Freeouf, D. A. Buchanan, S. L. Wright,
T. N. Jackson, and B. Robinson, Appl. Phys. Lett. 57, 1919 ͑1990͒.
J. Ivanco, T. Kubota, and H. Kobayashi, J. Appl. Phys. 97, 073712 ͑2005͒.
S. Koveshnikov, W. Tsai, I. Ok, J. C. Lee, V. Tokranov, M. Yakimov, and
S. Oktyabrsky, Appl. Phys. Lett. 88, 022106 ͑2006͒.
2
3
9
0
G. Landgren, R. Ludeke, J. F. Morar, Y. Jugnet, and F. J. Himspel, Phys.
Rev. B 30, 4839 ͑1984͒.
G. Hollinger, R. Skheyta-Kabbani, and M. Gendry, Phys. Rev. B 49,
7
8
11159 ͑1994͒.
3
3
1
2
J. Ivanco, T. Kubota, and H. Kobayashi, J. Appl. Phys. 97, 073712 ͑2005͒.
S. Arabasz, E. Bergignat, G. Hollinger, and J. Szuber, Appl. Surf. Sci.
252, 7659 ͑2006͒.
Z. H. Lu, M. J. Graham, X. H. Feng, and B. X. Yang, Appl. Phys. Lett.
62, 2932 ͑1993͒.
9
D. J. Webb, J. Fompeyrine, S. Nakagawa, A. Dimoulas, C. Rossel,
M. Sousa, R. Germann, S. F. Alvarado, J. P. Locquet, C. Marchiori, H.
Siegwart, A. Callegari, E. Kiewra, Y. Sun, J. De Souza, and N. Hoffmann,
Microelectron. Eng. 84, 2142 ͑2007͒.
3
3
3
4
For example, As2O3 melting point ϳ313 °C cf. Ga2O3 melting point
Ͼ1725 °C; see http://www.webelements.com/.
1
0
J. P. de Souza, E. Kiewra, Y. Sun, A. Callegari, D. K. Sadana, G. Shahidi,
D. J. Webb, J. Fompeyrine, R. Germann, C. Rossel, and C. Marchiori,
Appl. Phys. Lett. 92, 153508 ͑2008͒.
3
3
5
6
C. Sasaoka, Y. Kato, and A. Usui, Appl. Phys. Lett. 62, 2338 ͑1993͒.
R. L. Puurunen, J. Appl. Phys. 97, 121301 ͑2005͒.