Structural and electrical properties of ternary Ru-AlN thin films prepared by plasma-enhanced atomic layer deposition

Article abstract of DOI:10.1016/j.materresbull.2011.12.004

Ruthenium-aluminum-nitride (Ru-AlN) thin films were grown by plasma-enhanced atomic layer deposition (PEALD) at 300 °C. The Ru intermixing ratio of Ru-AlN thin films was controlled by the number of Ru unit cycles, while the number of AlN unit cycles was fixed to one cycle. The electrical resistivity of Ru-AlN thin film decreased with increasing the Ru intermixing ratio, but a drastic decrease in electrical resistivity was observed when the Ru intermixing ratio was around 0.58-0.78. Bright-field scanning transmission electron microscope (BF-STEM) and energy-dispersive X-ray spectroscopy (EDX) element mapping analysis revealed that the electrical resistivity of Ru-AlN thin film was strongly dependent on the microstructures as well as on the Ru intermixing ratio. Although the electrical resistivity of Ru-AlN thin films decreased with increasing the Ru intermixing ratio, a drastic decrease in electrical resistivity occurred where the electrical paths formed as a result of the coalescence of Ru nanocrystals.

Full text of DOI:10.1016/j.materresbull.2011.12.004

Materials Research Bulletin 47 (2012) 790–793
Contents lists available at SciVerse ScienceDirect
Materials Research Bulletin
journal homepage: www.elsevier.com/locate/matresbu
Structural and electrical properties of ternary Ru–AlN thin films prepared by
plasma-enhanced atomic layer deposition
a
a
b
c
d
Yu-Ri Shin , Won-Sub Kwack , Yun Chang Park , Jin-Hyock Kim , Seung-Yong Shin ,
d
a
a,
Kyoung Il Moon , Hyung-Woo Lee , Se-Hun Kwon *
a
National Core Research Center for Hybrid Materials Solutions, Pusan National University, 30 Jangjeon-Dong Geumjeong-Gu, Busan 609-735, Republic of Korea
b
Measurement & Analysis Team, National NanoFab Center, 335 Gwahakno, Yuseong-Gu, Daejeon 305-806, Republic of Korea
c
Hynix Semiconductor Incorporated, San 136-1, Ami-ri, Bubal-eub, Icheon-si, Kyoungki-do 467-701, Republic of Korea
d
Eco Materials & Processing Department, Korea Institute of Industrial Technology, 7-47 Songdo-Dong, Yeonsu-Gu, Incheon 406-840, Republic of Korea
A R T I C L E I N F O
A B S T R A C T
Article history:
Ruthenium–aluminum-nitride (Ru–AlN) thin films were grown by plasma-enhanced atomic layer
deposition (PEALD) at 300 8C. The Ru intermixing ratio of Ru–AlN thin films was controlled by the number
of Ru unit cycles, while the number of AlN unit cycles was fixed to one cycle. The electrical resistivity of
Ru–AlN thin film decreased with increasing the Ru intermixing ratio, but a drastic decrease in electrical
resistivity was observed when the Ru intermixing ratio was around 0.58–0.78. Bright-field scanning
transmission electron microscope (BF-STEM) and energy-dispersive X-ray spectroscopy (EDX) element
mapping analysis revealed that the electrical resistivity of Ru–AlN thin film was strongly dependent on
the microstructures as well as on the Ru intermixing ratio. Although the electrical resistivity of Ru–AlN
thin films decreased with increasing the Ru intermixing ratio, a drastic decrease in electrical resistivity
occurred where the electrical paths formed as a result of the coalescence of Ru nanocrystals.
ß 2011 Elsevier Ltd. All rights reserved.
Received 27 September 2011
Accepted 3 December 2011
Available online 9 December 2011
Keywords:
A. Thin films
B. Plasma deposition
D. Electrical properties
D. Microstructure
1
. Introduction
2 3
TaC [5], Ru–WCN [6], Ru–SiN [7], Ru–Al O [8] films can be utilized
as single-layered Cu diffusion barrier layers that do not require
additional glue or Cu seed layers. In these Ru-based ternary thin
films, low resistivity Ru was intentionally intermixed with
In the recent years, ruthenium-based ternary thin films that are
prepared by atomic layer deposition (ALD) or plasma-enhanced
ALD (PEALD) have received extensive attention. This is particularly
because their extraordinary properties such as a relatively low
electrical resistivity with a nanocrystal-embedded amorphous
structure, superior chemical and oxidation resistance, high
thermal stability, and good adhesion to substrates such as Si,
immiscible binary nitrides (TiN, TaN, WCN, and SiN
x
), carbides
(TaC) or oxides (Al ) in order to prevent the columnar growth of
2 3
O
Ru and improve the diffusion barrier properties. Moreover, these
single-layer Ru-based ternary diffusion barrier thin films for Cu are
more desirable than the traditional trilayer structures for further
scaling down of the devices. In addition, other potential applica-
tions of these Ru-based ternary thin films are now being explored,
e.g., as the oxygen diffusion barrier, whose required properties are
similar to those of Cu diffusion barrier films, in capacitors for future
dynamic random access memory (DRAM) and ferroeletric random
access memory (FeRAM) [9]. However, despite the increased
interest in the use of Ru-based ternary thin films and the fact that
many researchers have studied the performance of these films, the
relationship between microstructural evolution and electrical
resistivity has not yet been systematically discussed.
Herein, we report the synthesis of one of the Ru-based ternary
thin films, Ru–AlN thin films, by PEALD, where AlN was intermixed
with Ru to prevent columnar growth of Ru. The microstructural
evolution depending on the Ru intermixing ratio in Ru–AlN thin
films was characterized using an X-ray diffractometer (XRD) and a
bright-field scanning transmission electron microscope (BF-STEM)
2
SiO , low-k dielectrics, and Cu [1–9] make them suitable as Cu
diffusion barriers in ultra-large-scale integration (ULSI) metalliza-
tion and/or oxygen diffusion barriers in capacitors for future
dynamic random access memory (DRAM) and ferroelectric random
access memory (FeRAM).
One of important advantage of these Ru-based ternary thin
films is that they can potentially substitute the conventional
complex interconnection process with a simple one; i.e., although a
traditional Cu interconnection structure has employed a complex
and thick trilayer configuration (Cu seed/Glue/Barrier layers) prior
to Cu electrochemical deposition, recent studies revealed that Ru-
based ternary thin films such as Ru–TiN [1,2], Ru–TaN [3,4], Ru–
*
Corresponding author. Tel.: +82 51 510 3775; fax: +82 51 518 3360.
E-mail address: sehun@pusan.ac.kr (S.-H. Kwon).
0025-5408/$ – see front matter ß 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2011.12.004

Y.-R. Shin et al. / Materials Research Bulletin 47 (2012) 790–793
791
with energy-dispersive X-ray spectroscopy (EDX) element map-
ping. Further, the relationship between microstructural evolution
and electrical resistivity was systematically investigated.
TMA pulse time (sec)
2
1
3
a
0.050
0.045
PEALD-Ru
PEALD-AlN
2
. Experimental details
2
Ru–AlN films were deposited on 260-nm-thick SiO /Si sub-
0
0
.040
.035
strates by PEALD at a deposition temperature of 300 8C and
pressure of 3 Torr. Trimethylaluminum (TMA) and bis(ethylcyclo-
2
pentadienyl)ruthenium [Ru(EtCp) ] were used as the Al and Ru
layer precursors, respectively. The canisters containing Ru(EtCp)
2
and TMA was maintained at a temperature of 80 8C and 25 8C,
respectively. During the deposition, 100 sccm Ar gas was
continuously supplied into the reactor. Ru–AlN films were
deposited by repeating a supercycle consisting of Ru and AlN
subcycles. Each subcycle consisted of several unit cycles. In each
unit cycle, four consecutive pulses were supplied. A unit cycle for
AlN was comprised of TMA precursor injection pulse with
0.030
0
0
.025
.020
0
1
2
3
4
5
6
7
8
Ru(EtCp) pulse time (sec)
1
00 sccm Ar carrier gas, a purge pulse with 50 sccm Ar, a pulse
for an exposure to NH plasma with 100 sccm NH gas, and another
0 sccm Ar purge pulse. Also, a unit cycle for Ru was comprised of
Ru(EtCp)2 injection pulse with 50 sccm Ar carrier gas, a purge
pulse with 50 sccm Ar, a pulse for an exposure to NH plasma with
00 sccm NH gas, and another 50 sccm Ar purge pulse. Radio
2
3
3
1
0
0
.0
5
b
x in Ru -(AlN)
x
1-x
3
.8
1
3
frequency (RF) plasma was used at an electrical power of 180 W.
The intermixing ratios of Ru in the deposited films could be
controlled by changing the number of unit cycles in a Ru deposition
subcycle, whereas the number of unit cycles allocated for AlN
deposition subcycle was fixed as one cycle.
.6
0.4
0.2
The film thickness was analyzed by field-emission scanning
electron microscopy (FESEM), and the film composition was
2+
analyzed by 9.0 MeV He Rutherford backscattering spectroscopy
RBS) and Auger electron spectroscopy (AES). The sheet resistance
(
was measured using a four-point probe, and the film resistivity
was calculated from the sheet resistance and the film thickness.
0
.0
0
5
10
15
Ru
The microstructures of the films were studied using an XRD with
˚
a radiation at 1.5405 A and a BF-STEM. Further, TEM-EDX
Number of Ru unitcycle in Ru-AlN supercycle
Cu-K
element mapping was performed to investigate the relationship
between microstructural evolution and electrical resistivity of
films.
Fig. 1. (a) Dependence of the thickness per cycle (nm/cycle) on the Ru(EtCp)
2
pulse
time in case of Ru PEALD and on the TMA pulse time in the case of AlN PEALD at a
growth temperature of 300 8C. (b) Dependence of the Ru intermixing ratio in the
Ru–AlN thin films on the number of Ru unit cycles in a super cycle for Ru–AlN thin
films, while the number of AlN unit cycle was fixed to one cycle.
3
. Results and discussion
Fig. 1(a) shows the growth characteristics of Ru and AlN films
between Ru and AlN can be easily controlled by the number of Ru
unit cycles. Since the thickness of each Ru and AlN layer grown in
prepared by PEALD depending on the pulse time of their
precursors. For Ru PEALD, the pulse time of Ru(EtCp) was varied
from 1 to 8 s. And, for AlN PEALD, the pulse time of TMA was varied
from 1 to 3 s. For both Ru and AlN PEALD, NH plasma pulse time
˚
2
one cycle is fairly below the monolayer thickness of Ru (ꢀ3 A) [10]
˚
and AlN (ꢀ2 A) [11], the Ru–AlN thin films prepared in this study
3
are expected to have a single-layer mixed-phase structure rather
than a multilayered structure.
was fixed at 15 s. The given plasma pulse time, 15 s for Ru and AlN
PEALD, was confirmed to be sufficient for completing the chemical
reaction of the adsorbed precursor. As shown in Fig. 1, the growth
Fig. 2 shows the electrical resistivity of the Ru–AlN thin films
depending on the Ru intermixing ratios in the Ru–AlN super cycle.
To carefully compare the electrical resistivity of Ru–AlN thin films
with different Ru intermixing ratios, the thickness of the thin films
was controlled to have a similar thickness of about 50 nm. When
the Ru intermixing ratios in Ru–AlN thin films was less than 0.47,
the electrical resistivity of the films are very high because AlN is a
well-known electrical insulating material having a high electrical
2
rate of Ru is saturated at 0.039 nm/cycle at the Ru(EtCp) pulse
time of 5 s, and that of AlN is saturated at 0.17 nm/cycle at the TMA
pulse time of 2 s. On the basis of the growth characteristics of Ru
and AlN, a super cycle for Ru–AlN PEALD was designed to obtain
purposed intermixing ratios between Ru and AlN. In each subcycle
for Ru and AlN, the precursor pulse time was set in a saturation
1
7
20
region, 7 s for Ru(EtCp)
2
and 2 s for TMA, and the plasma pulse time
resistivity (rbulk AlN = 10 –10 m V cm) [12]. However, a drastic
was 15 s for Ru and AlN. The Ru intermixing ratios in the deposited
Ru–AlN thin film was controlled by changing the number of unit
cycles in the Ru subcycle from 1 to 15 times, while the number of
unit cycles in the AlN subcycle was fixed to one cycle. In Fig. 1(b),
the dependence of the Ru intermixing ratio in the Ru–AlN thin
films on the number of Ru unit cycles executed during a super cycle
is shown. The Ru intermixing ratio gradually increased with the
number of Ru unit cycles. It shows that the intermixing ratios
decrease in the electrical resistivity of Ru–AlN thin film was
observed when the Ru intermixing ratio was around 0.58–0.78.
Then, the electrical resistivity gradually decreased when the Ru
intermixing ratio increased beyond 0.78. It is reasonable that the
electrical resistivity of Ru–AlN film decreased with increasing Ru
intermixing ratio because Ru has a low electrical resistivity of
13 m V cm. However, such
a drastic decrease in electrical
resistivity when the Ru intermixing ratios was around 0.58–0.78

7
92
Y.-R. Shin et al. / Materials Research Bulletin 47 (2012) 790–793
1
1
E13
E11
AlN
Ru0.1-(AlN)0.9
Electrical Resistivity of Ru-AlN thin films
1E9
Ru0.47-(AlN)0.53
1
E7
Ru0.58-(AlN)0.42
6
5
4
3
2
1
00
00
00
00
00
00
0
Ru0.78-(AlN)0.22
Ru0.86-(AlN)0.14
Ru
Ru
Ru0.86-(AlN)0.14
Ru0.78-(AlN)0.22
Ru0.58-(AlN)0.42
Ru0.47-(AlN)0.53
Ru0.1-(AlN)0.9
AlN
3
4
36
38
40
42
44
46
2θ
0
.0
0.1
0.2 0.3
0.4
0.5
0.6
0.7 0.8
0.9
1.0
x in Ru -(AlN)
x
1-x
Fig. 3. XRD patterns of PEALD–Ru, PEALD–AlN and PEALD–Ru–AlN thin films with
various Ru intermixing ratios.
Fig. 2. Variation in the electrical resistivity of Ru–AlN films as a function of the Ru
intermixing ratio.
is hard to be explained on the basis of only the Ru intermixing
ratios, and therefore, the effect of microstructural variation in Ru–
AlN thin films on their electrical resistivity should be considered.
Fig. 3 shows the crystallinity of Ru–AlN thin films investigated
by XRD analysis. The crystallinity of the Ru and AlN thin films
prepared by PEALD were considered as a reference. To compare
the microstructures of the films under the same condition,
the thicknesses of all the films were fixed at about 50 nm. From
the JCPDS No. 06-0663, the crystallinity of Ru thin films is
identified as hexagonal, and a clear Ru(1 0 0) and Ru(1 0 1)
Fig. 4. BF-STEM images with EDX element mapping of (a) Ru0.47–(AlN)0.53, (b) Ru0.58–(AlN)0.42, and (c) Ru0.78–(AlN)0.53 samples.

Y.-R. Shin et al. / Materials Research Bulletin 47 (2012) 790–793
793
diffraction peaks were observed. On the other hand, no clear peaks
were found for the AlN thin film, indicating that the AlN film has an
amorphous structure. For Ru–AlN thin films, it was revealed as an
amorphous structure when the Ru intermixing ratio of the film was
below 0.58. Further, weak and broad Ru(1 0 0) and Ru(1 0 1) peaks
start to be appeared when the Ru intermixing ratios in the Ru–AlN
thin film were further increased from 0.58 to 0.78. However, when
the Ru intermixing ratio in the Ru–AlN thin film was higher than
continuous electrical paths rather than discrete Ru nanocrystals at
the same Ru intermixing ratio, the lower electrical resistivty of Ru-
based ternary thin films can be achieved.
4. Conclusions
In summary, Ru–AlN thin films were prepared by PEALD, and
the relationship between electrical and microstructural properties
was systematically investigated. From the BF-STEM and EDX
element mapping analysis, it was revealed that the electrical
resistivity of Ru–AlN thin film was strongly dependent on the
microstructures of Ru–AlN films as well as Ru intermixing ratio.
Although the electrical resistivity of Ru–AlN thin films decreased
with an increase in the low resistive Ru intermixing ratio, a drastic
decrease in electrical resistivity occurred when the electrical paths
formed as a result of the coalescence of Ru nanocrystals. This
observation suggests that the control of microstructure as well as
of Ru intermixing ratio should be considered in preparation of Ru-
based ternary thin films to obtain a sufficiently low electrical
resistivity for the electronic device applications.
0
.87, the crystal structure of Ru–AlN thin film was almost similar to
that of Ru thin films formed by PEALD although the peak intensities
of Ru–AlN thin film was still weaker than those of the Ru thin films.
From the XRD and electrical resistivity observations, it was
noteworthy that microstructural change of Ru–AlN thin film from
amorphous to polycrystalline was occurred when the Ru inter-
mixing ratio was around 0.58 and 0.78, where electrical resistivity
of the film changed suddenly. This means that the electrical
resistivity of Ru–AlN thin films may be strongly dependent on their
microstructural changes. To investigate more detailed microstruc-
ture of Ru–AlN films and the relationship between the microstruc-
tural evolution and electrical properties, BF-STEM measurements
with EDX element mapping was performed on (a) Ru0.47–(AlN)0.53
,
(b) Ru0.58–(AlN)0.42, and (c) Ru0.78–(AlN)0.53 samples. Fig. 4 shows a
Acknowledgments
plan-view BF-STEM image, EDX element mapping of Ru, EDX
element mapping of Al, and EDX element mapping of both Ru and
Al together. The BF-STEM results show that the microstructure of
Ru–AlN films in which about 3-nm-sized Ru crystallites are
embedded in AlN amorphous matrix is very similar to well-known
nanocomposite structures. However, the EDX element mapping
analysis revealed that the microstructures of Ru–AlN thin films
vary depending on the Ru intermixing ratios. When the Ru
intermixing ratio was 0.47, Ru atoms formed discrete Ru
nanocrystals, which are not connected to each other and are
separated by the AlN amorphous matrix. However, as shown in
Fig. 4(b) and (c), the BF-STEM image and EDX element mapping
indicate that Ru nanocrystals started to connect with each other
and form electrical paths in Ru–AlN thin films. Moreover, as the
intermixing ration of Ru increased up to 0.78, the number and
width of electrical paths are further increased, as indicated from
the variation in the EDX element mapping of Ru atoms. From the
above observation, it can be concluded that the drastic decrease in
the electrical resistivity of Ru–AlN film is responsible for the
microstructural variation in Ru–AlN thin films, i.e., from the
structure with discrete Ru nanocrystals embedded in the
amorphous AlN matrix to the formation of continuous electrical
paths as a results of the coalescence of Ru nanocrystals. These
observations provide insight into the engineering required to
realize suitable electrical resistivity in Ru-based ternary thin films
for their application in electronic devices. First, the electrical
resistivity of Ru-based ternary thin films can be controlled on the
basis of the Ru intermixing ratios. Moreover, if the microstructure
of Ru-based ternary thin films is carefully controlled to form
This research was supported by Basic Science Research Program
through the National Research Foundation of Korea (NRF) funded
by the Ministry of Education, Science and Technology (2010-
0006643), by a grant from the Fundamental R&D Program for Core
Technology of Materials funded by the Ministry of Knowledge
Economy, Republic of Korea, and by NCRC (National Core Research
Center) program through the National Research Foundation of
Korea funded by the Ministry of Education, Science and Technology
(
2010-0001-226).
References
[1] S.H. Kwon, O.K. Kwon, J.S. Min, S.W. Kang, J. Electrochem. Soc. 153 (2006) G578–
G581.
[2] J. Li, H.S. Lu, Y.W.X. Wang, P. Qu, Microelectron. Eng. 88 (2011) 635–640.
[3] S.W. Kim, S.H. Kwon, S.J. Jeong, S.W. Kang, J. Electrochem. Soc. 155 (2008) H885–
H888.
[4] S. Kumar, D. Greenslit, T. Chakraborty, E.T. Eisenbraun, J. Vac. Sci. Technol. A 27
(2009) 572–576.
[5] J.S. Fang, J.H. Lin, B.Y. Chen, T.S. Chin, J. Electrochem. Soc. 158 (2011) H97–H102.
6] D. Greenslit, S. Kumar, T. Chakraborty, E.T. Eisenbraun, ECS Trans. 13 (2008) 63–
0.
[
7
[7] T.K. Eom, S.H. Kim, K.S. Park, S. Kim, H. Kim, Electrochem. Solid-State Lett. 14
(2011) D10–D12.
[8] T.H. Cheon, S.H. Choi, S.H. Kim, D.H. Kang, Electrochem. Solid-State Lett. 14 (2011)
D57–D61.
[9] S.J. Jeong, D.I.S. Kim, O. .Kim, T.H. Han, J.D. Kwon, J.S. Park, S.H. Kwon, J. Nanosci.
Nanotechnol. 11 (2011) 671–674.
[10] K.J. Park, Ph.D Thesis, North Carolina State University, 2005.
[11] H. Lewalter, B.O. Kolbesen, Z. Anorg. Allg. Chem. 629 (2003) 1760–1768.
[12] K.C. Winco, J.W. Yung, T.M. Yuea, B.L. Zhua, C.P. Lee, Composites: Part A 38 (2007)
2055.