Characteristics of Ti -capped Co films deposited by a remote plasma ALD method using cyclopentadienylcobalt dicarbonyl

Article abstract of DOI:10.1149/1.2769327

Co films were deposited by a remote plasma atomic layer deposition (RPALD) method using cyclopentadienylcobalt dicarbonyl [CpCo(CO) 2] as the Co precursor with H2 plasma. The impurity contents in the Co films were minimized under the optimized process conditions with H2 plasma using a process pressure range between 0.1 and 2 Torr and a plasma power of 300 W. The ALD process window of the Co films showed a saturated temperature range between 125 and 175°C. The carbon and oxygen contents of as-deposited Co films were about 8 and 1 atom %, respectively. However, the carbon content in the Co films decreased from 8 to 4 atom % after in situ annealing at 400°C. For in situ annealed Co films deposited on Si substrates, a polycrystalline CoSi2 phase was observed. The surface and interface morphologies of CoSi2 Si were rough compared to Ti -capped CoSi2 Si after ex situ annealing at 600°C. In addition, CoSi was completely transformed to CoSi2 at 600°C. However, in the in situ annealed Co films with Ti -capped layer, the diffraction peak of CoSi2 (200) began to appear at 700°C. The formation temperature of the Ti -capped CoSi2 phase was retarded by about 100°C compared to the Co film without the Ti -capped layer. In addition, the surface and interface morphologies of the Ti -capped CoSi2 layer were smooth.

Full text of DOI:10.1149/1.2769327

Characteristics of Ti -Capped Co Films Deposited by a Remote
Plasma ALD Method Using Cyclopentadienylcobalt Dicarbonyl
Keunwoo Lee, Keunjun Kim, Taeyong Park, Hyeongtag Jeon, Youngjin Lee, Jeongtae Kim and
Seungjin Yeom
J. Electrochem. Soc. 2007, Volume 154, Issue 10, Pages H899-H903.
doi: 10.1149/1.2769327
Receive free email alerts when new articles cite this article - sign up in the box at the
top right corner of the article or click here
Email alerting
service
To subscribe to Journal of The Electrochemical Society go to:
http://jes.ecsdl.org/subscriptions
© 2007 ECS - The Electrochemical Society
Downloaded on 2013-02-26 to IP 132.174.255.3 address. Redistribution subject to ECS license or copyright; see www.esltbd.org

Journal of The Electrochemical Society, 154 ͑10͒ H899-H903 ͑2007͒
H899
0013-4651/2007/154͑10͒/H899/5/$20.00 © The Electrochemical Society
Characteristics of Ti-Capped Co Films Deposited by a Remote
Plasma ALD Method Using Cyclopentadienylcobalt
Dicarbonyl
Keunwoo Lee,a Keunjun Kim,a Taeyong Park,a Hyeongtag Jeon,a,z Youngjin Lee,b
Jeongtae Kim,b and Seungjin Yeomb
aDivision of Materials Science and Engineering, Hanyang University, Seoul 133-791, Korea
bHynix Semiconductor Incorporated, Research and Development Division, Bubal-eup, Icheon-si,
Kyoungi-do 467-701, Korea
Co films were deposited by a remote plasma atomic layer deposition ͑RPALD͒ method using cyclopentadienylcobalt dicarbonyl
͓CpCo͑CO͒ ͔ as the Co precursor with H2 plasma. The impurity contents in the Co films were minimized under the optimized
2
process conditions with H2 plasma using a process pressure range between 0.1 and 2 Torr and a plasma power of 300 W. The
ALD process window of the Co films showed a saturated temperature range between 125 and 175°C. The carbon and oxygen
contents of as-deposited Co films were about 8 and 1 atom %, respectively. However, the carbon content in the Co films decreased
from 8 to 4 atom % after in situ annealing at 400°C. For in situ annealed Co films deposited on Si substrates, a polycrystalline
CoSi2 phase was observed. The surface and interface morphologies of CoSi2/Si were rough compared to Ti-capped CoSi2/Si after
ex situ annealing at 600°C. In addition, CoSi was completely transformed to CoSi2 at 600°C. However, in the in situ annealed Co
films with Ti-capped layer, the diffraction peak of CoSi2͑200͒ began to appear at 700°C. The formation temperature of the
Ti-capped CoSi2 phase was retarded by about 100°C compared to the Co film without the Ti-capped layer. In addition, the surface
and interface morphologies of the Ti-capped CoSi2 layer were smooth.
© 2007 The Electrochemical Society. ͓DOI: 10.1149/1.2769327͔ All rights reserved.
Manuscript submitted April 2, 2007; revised manuscript received June 17, 2007. Available electronically August 20, 2007.
Metal silicides have been widely applied in advanced semicon-
ductor devices such as ohmic contacts, gate electrodes, Schottky
barriers, and interconnects.1,2 Among the metal silicides, Co-silicide
is considered as an excellent choice due to its good thermal and
chemical stability, low resistivity, immunity to shrinkage of the line
width, and close lattice mismatch with the Si substrate.3,4 Because of
these advantages, Co-silicide has been extensively investigated for
application in ultralarge-scale integration ͑ULSI͒ devices. However,
the use of Co-silicide can result in a rough interface formation and
the incorporation of impurities such as oxygen.5 A possible solution
for these problems is to use a capping layer. In particular, the use of
a Ti capping layer improves the interface roughness of Co-silicide.
During annealing, the reactive Ti capping layer captures oxygen
which is present in the interface and protects the silicide from oxy-
gen contamination.6 Many researchers have already reported that the
uniform interface smoothness and thermal stability of the CoSi2
layer were improved by using a Ti capping layer.7,8
quality. This method was designed such that the substrate was
placed out of the plasma region with the expectation that impurities
and substrate damage would be reduced.16 We previously reported
that RPALD minimizes the problems caused by the use of a direct
plasma.17
Employing the RPALD method, we deposited Co films using
cyclopentadienylcobalt dicarbonyl ͓CpCo͑CO͒ ͔ as the Co precur-
2
sor. After Co deposition, Co films were characterized with transmis-
sion electron microscope ͑TEM͒, Auger electron spectroscopy
͑AES͒, and X-ray diffraction ͑XRD͒ to investigate their physical and
chemical properties. We carried out the in situ annealing of Co films
and the ex situ annealing of Ti-capped Co films at various tempera-
tures. Then, we also investigated the variation of impurity content
and interface morphology of the Co films as a function of the an-
nealing temperature.
Experimental
To investigate the formation of Co-silicide, we applied an atomic
layer deposition ͑ALD͒ method. Conventionally, a sputtering
method has been the predominant deposition technique for CoSi2
formation. However, the poor step coverage of the sputtering
method becomes increasingly problematic in deep contact holes
with high aspect ratio.9 Moreover, the chemical vapor deposition
͑CVD͒ method offers advantages such as good conformal coverage
and high growth rates but it generally introduces a large amount of
Using the RPALD method with CpCo͑CO͒ as the Co precursor
2
and H2 as a plasma reactant gas, Co films were deposited on boron-
doped p-type Si͑100͒ substrates with a resistivity of 3–9 ⍀ cm. The
Si substrates were cleaned by dipping in piranha solution
͑H2SO4/H2O2 = 4:1͒ for 10 min and then in a dilute hydrofluoric
acid ͑HF͒ solution ͑HF/H2O = 1:50͒ for 1 min to remove organics
and native oxide, respectively, and then immediately loaded into the
reaction chamber. A downstream-type RPALD reactor with a
13.56 MHz radio frequency ͑rf͒ power source was used in this
impurities by gas phase reaction and
a high deposition
temperature.10,11 Consequently, to overcome the shortcomings of
conventional sputtering and CVD methods, we used an ͑ALD͒
method for this study.12 Generally, using the ALD method, the pre-
cursor and reactant gas are introduced separately by maintaining a
purge gas between the source and the reactants to prevent gas-phase
reactions. Then, film growth by the ALD method is controlled by
surface reactions through surface and self-limited mechanisms.13 It
is well known that the ALD method has many advantages such as
ultrathin-film-growth capability, control of thickness, and elimina-
tion of particle generation due to gas-phase reactions.14,15 Recently,
reactive species such as ozone and plasma have been applied to
improve the film quality in the ALD method. In particular, we ap-
plied the remote plasma ALD ͑RPALD͒ method to improve the film
study. One deposition cycle consisted of exposure to CpCo͑CO͒
2
cobalt precursor, a purge with argon, an exposure to H2 plasma, and
another argon purge. Argon purge gas was introduced to ensure the
complete separation of the Co precursor and H2 reactant gas. The H2
gas flow rates were fixed at 20 sccm. The source injection time was
fixed at 5 s for the CpCo͑CO͒ precursor. The processing times for
2
the initial Ar purge, H2 reactant plasma, and second Ar purge were
10, 20, and 10 s, respectively. The Co precursor was vaporized at
20°C for CpCo͑CO͒ and introduced into the reaction chamber us-
2
ing a bubbler with Ar as a carrier gas. In the previous study, we
optimized the process condition. The growth rate of the Co films
was about 1.1 Å per cycle in the ALD process window with a satu-
rated temperature range between 125 and 175°C. Accordingly, Co
films were deposited by the deposition temperature of 150°C with
H2 plasma using a process pressure range between 0.1 and 2 Torr at
z E-mail: hjeon@hanyang.ac.kr
Downloaded on 2013-02-26 to IP 132.174.255.3 address. Redistribution subject to ECS license or copyright; see www.esltbd.org

H900
Journal of The Electrochemical Society, 154 ͑10͒ H899-H903 ͑2007͒
Figure 1. Schematic of cluster tool in-
cluding RPALD system and E-beam
evaporator.
the plasma power of 300 W. After Co deposition, the Co film was
annealed in the reaction chamber without exposure to air. All
samples were in situ annealed at 400°C. Then, a Ti capping layer
with a thickness of 10 nm was deposited on the in situ annealed Co
film by an e-beam evaporator without breaking the vacuum, as
shown in Fig. 1. Subsequently, ex situ rapid thermal annealing
͑RTA͒ was carried out in N2 ambient at temperatures between 500
and 800°C for 1 min. The chemical composition and impurity con-
tent were analyzed by AES. The phase identifications were carried
out by XRD. The film thickness and microstructure of Co films were
examined with cross-sectional transmission electron microscope
͑XTEM͒.
content, the as-deposited Co film was treated with in situ annealing
after Co deposition. In situ annealing was carried out in the deposi-
tion chamber without breaking the vacuum at a substrate tempera-
ture of 400°C. As a result, the carbon content in the Co film after in
situ annealing decreased from about 8 to about 4 atom %, as shown
in Fig. 2. However, the oxygen content in the Co film remained at
about 1 atom %.
To investigate the phase identification of Co-silicide at various
temperatures, in situ annealed 110 Å Co films without a Ti capping
layer were taken out of the chamber, and ex situ annealing by RTA
was performed at temperatures ranging between 500 and 800°C for
1 min in N2 ambient. The phase identification was done with XRD,
and the XRD patterns of Co-silicide formed on the Si substrate are
shown in Fig. 3. XRD data showed that the ͑111͒ and ͑200͒ diffrac-
tion peaks of the ␤ − Co phase were observed with the in situ an-
nealed Co film. After ex situ annealing at 500°C, the ͑210͒ and
͑211͒ peaks of the CoSi phase were observed by XRD analysis. It is
usually reported that the silicide formation sequence of PVD-Co
film, depending on the annealing temperature, is Co2Si → CoSi
→ CoSi2.20 The phase transition fromCo2Si to CoSi occurs at
350°C and the transition from CoSi to CoSi2 occurs at 650°C. The
Results and Discussion
We investigated the impurity contents in the Co films on Si͑100͒
substrates with and without in situ annealing. We also examined the
interface morphology and the phase formation of Co-silicide with
and without the Ti capping layer after in situ annealing. The varia-
tion of impurity contents in the Co films after in situ annealing was
observed using AES. Kim et al. showed the AES depth profile of the
Co film deposited by RPALD using CpCo͑CO͒2 precursor at 150°C
with H2 plasma. The deposition rate of the ALD process for the Co
film was ϳ1.1 Å per cycle with an ALD process window tempera-
ture between 125 and 175°C.18 We also previously reported the
process conditions of Co deposition in which the thickness of the Co
film increased linearly with the number of cycles used in the ALD
process window.18 In that process, the impurity contents in the Co
films were minimized using optimized conditions with H2 plasma at
a pressure range between 0.1 and 2 Torr and a plasma power of
300 W. The carbon and oxygen contents of the as-deposited Co
films were about 8 and 1 atom %, respectively, and the oxygen
content increased slightly at the interface of the Co film and Si
substrate. The Cp ligand in CpCo͑CO͒ reacts in a reducing ͑H2͒
2
atmosphere to form both stable and volatile hydrocarbons ͑CxHy͒
which are easily removed through the gas phase. The chemical re-
action leading to the deposition of the Co film is generally repre-
sented by the equation Co͑C5H5͒͑CO͒ ͑g͒ + 1/2nH2͑g͒ → Co͑s͒
2
+ 2CO͑g͒ + C5H5+n͑g͒.19 Although the Co film deposited by
RPALD using CpCo͑CO͒ precursor contained about 8 atom % car-
2
bon content, this is much lower than the approximately 50 atom %
carbon content of Co films deposited by the metallorganic CVD
͑MOCVD͒ method using the same precursor and a similar H2
pressure.20 These results indicated that the reactive hydrogen radical
in the H2 plasma is more effective at breaking Co–Cp bonds than
molecular hydrogen gas. Nevertheless, in order to lower the carbon
Figure 2. ͑Color online͒ AES depth profiles of in situ annealed Co films
deposited by RPALD using CpCo͑CO͒ with H2 plasma.
2
Downloaded on 2013-02-26 to IP 132.174.255.3 address. Redistribution subject to ECS license or copyright; see www.esltbd.org

Journal of The Electrochemical Society, 154 ͑10͒ H899-H903 ͑2007͒
H901
sectional TEM. The thickness of the in situ annealed Co film was
found to be about 110 Å, and this Co film showed a uniform surface
and a smooth interface as shown in Fig. 4a. Figure 4b shows that the
formation of the CoSi film appeared in the annealed Co film after ex
situ annealing at 500°C as shown in the XRD data ͑Fig. 3͒. Also, the
thickness of the CoSi film increased slightly. This indicated that Si
atoms diffused into the Co film and formed CoSi. In particular, in
the case of Co silicide formation, the volumetric change resulting
from a Co film changing to CoSi increased about 1.81 times during
silicide formation.21 After ex situ annealing at 600°C ͑Fig. 4c͒, dis-
continuous CoSi2 islands were formed, resulting in a rough surface
and interface of the CoSi2/Si substrate. In addition, cross-sectional
TEM of the Co film after ex situ annealing at 700°C exhibited CoSi2
recession and CoSi2 island growth corresponding to the XRD data,
as shown in Fig. 4d. The surface and interface morphologies of the
Co film without the Ti capping layer after ex situ annealing were
rough due to the agglomeration during the silicide formation
reaction.22 Later, we deposited a Ti capping layer to observe the
change of the surface and interface morphologies during CoSi2 for-
mation.
From the above result, we applied a Ti capping layer to prevent
the agglomeration of the Co film and to obtain a smooth surface and
interface characteristics. The Ti capping layer was deposited on the
in situ annealed Co film at 400°C by an e-beam evaporator and then
Ti capped Co films were treated by ex situ annealing.22 The inter-
face microstructure and the phase identification of Co films with the
Ti capping layer were observed by TEM and XRD analyses. Figure
5 shows the XRD patterns of Co films with the Ti capping layer
after ex situ annealing in the temperature range between 600 and
800°C. The diffraction peak of CoSi2͑200͒, which is coherent with
the Si substrate, began to appear after annealing at 700°C. However,
the diffraction peaks of CoSi such as ͑210͒ and ͑211͒ and the dif-
fraction peaks of CoSi2 such as ͑111͒, ͑220͒, and ͑311͒ were not
observed at all temperatures. Comparing the XRD data taken from
the Co films with and without the Ti capping layer, polycrystalline
CoSi2 characterized by the diffraction peaks of ͑111͒, ͑220͒, and
͑311͒ was observed in case of the Co film without a Ti capping layer
͑Fig. 3͒. However, the diffraction peak of CoSi2͑200͒ appeared in
Figure 3. ͑Color online͒ XRD patterns of in situ annealed Co films depos-
ited by RPALD using CpCo͑CO͒ with H2 plasma after ex situ annealing at
2
various temperatures for 1 min in N2 ambient.
CoSi phase was formed from the Co film by ex situ annealing at
500°C. After ex situ annealing above 600°C, the ͑111͒, ͑220͒, and
͑311͒ diffraction peaks of CoSi2 were observed without the ͑210͒
and ͑211͒ peaks of CoSi. Therefore, CoSi was completely trans-
formed to CoSi2 after ex situ annealing at 600°C. However, the
phase transition temperature of CoSi2 decreased from 650 to 600°C.
This indicates that the formation temperature of CoSi2 was slightly
decreased due to the decreased impurity contents ͑carbon 4 atom %
and oxygen 1 atom %͒ as a result of in situ annealing. We also
observed a sharp further increase of the intensities of the CoSi2
peaks up to an annealing temperature of 800°C.
To observe the interface and surface morphologies of Co silicides
at various temperatures, we examined Co silicides with cross-
Figure 4. XTEM micrographs of Co films
͑a͒ in situ annealed at 400°C and ex situ
annealed at ͑b͒ 500, ͑c͒ 600, and ͑d͒
700°C deposited by RPALD with H2
plasma.
Downloaded on 2013-02-26 to IP 132.174.255.3 address. Redistribution subject to ECS license or copyright; see www.esltbd.org

H902
Journal of The Electrochemical Society, 154 ͑10͒ H899-H903 ͑2007͒
Figure 5. ͑Color online͒ XRD patterns of Ti-capped Co films deposited by
RPALD using CpCo͑CO͒ with H2 plasma after ex situ annealing at various
2
temperatures for 1 min in N2 ambient.
the case of the Co film with a Ti capping layer, as shown in Fig. 5.
In particular, the formation temperature of the CoSi2 phase was
retarded about 100°C compared to the Co film without a Ti capping
layer. In the case of Co with a Ti capping layer, it was reported that
the phase transition temperature increased due to the penetration
effect of Ti into the Co film during silicide reaction.23 Furthermore,
the ͑111͒ and ͑200͒ diffraction peaks of TiN were observed above
700°C RTA in N2 ambient by XRD data, as shown in Fig. 5. The
reason why the Ti layer changed into a TiN layer after ex situ
annealing was that nitrogen atoms diffused into the Ti capping layer
during high-temperature annealing, and the Ti atoms reacted with
the nitrogen atoms.
As shown in Fig. 4, Co films without a Ti capping layer showed
rough surface and interface morphologies. To investigate the effect
of the Ti capping layer, a Ti layer was deposited on the Co film and
annealed at 700°C for 1 min RTA in N2 ambient. Figure 6 shows the
AES depth profile of the Co film with a Ti capping layer and the
cross-sectional TEM with two distinct layers on the Si substrate. The
top layer is considered to be a TiN layer where the nitrogen came
from the N2 annealing ambient and then reacted with the Ti capping
layer. The second layer is a Co-silicide layer where the atomic ratio
of Co to Si was approximately 1:2, as shown in Fig. 6a. Also, the
AES depth profile exhibits high oxygen impurity levels in the TiN
layer and carbon impurities between the TiN and Co-silicide layers.
In addition, the TEM of Co with a Ti capping layer film shows
highly uniform surface and interface morphologies as shown in Fig.
6b. The interfaces of both TiN/CoSi2 and CoSi2/Si are very planar
and smooth.
Figure 6. ͑Color online͒ ͑a͒ AES depth profile and ͑b͒ XTEM micrograph of
a Ti-capped Co film deposited by the RPALD method after ex situ annealing
at 700°C for 1 min in N2 ambient.
Acknowledgment
This work was supported by the National Program for Tera-level
Nanodevices of the Ministry of Science and Technology as one of
the 21st Century Frontier Programs.
Hanyang University assisted in meeting the publication costs of this
article.
References
1. H. Lee and H. Kim, Electrochem. Solid-State Lett., 9, G323 ͑2006͒.
2. S. Wolf, Silicon Processing for the VLSI Era, Lattice Press, New York ͑2002͒.
3. S. P. Murarka, Intermetallics, 3, 173 ͑1995͒.
Conclusion
In conclusion, Co films were deposited by RPALD using cyclo-
pentadienylcobalt dicarbonyl ͓CpCo͑CO͒ ͔ as the Co precursor with
4. K. Maex, Mater. Sci. Eng., R., 11, 53 ͑1993͒.
2
5. W. J. Freitas and J. W. Swart, J. Electrochem. Soc., 138, 3067 ͑1991͒.
6. C. Detavernier, R. L. V. Meirhaeghe, F. Cardon, R. A. Donaton, and K. Maex,
Microelectron. Eng., 50, 125 ͑2000͒.
7. R. T. Besser and F. Schrey, Appl. Phys. Lett., 67, 2164 ͑1995͒.
8. D. Sohn, J. Park, B. Lee, J. Bae, J. Byun, and J. Kim, Appl. Phys. Lett., 73, 2302
͑1998͒.
9. C. Fitz, M. Goldbach, and A. Dupont, Microelectron. Eng., 82, 460 ͑2005͒.
10. J. Lee, H. Yang, J. Lee, J. Kim, W. Nam, H. Shin, Y. Ko, J. Lee, E. Lee, and C.
Kim, J. Electrochem. Soc., 153, G539 ͑2006͒.
11. M. E. Gross, K. S. Kranz, D. Brasen, and H. Luftman, J. Vac. Sci. Technol. B, 6,
1548 ͑1988͒.
H2 plasma as a reactant at a process pressure range between 0.1 and
2 Torr and a plasma power of 300 W. The carbon and oxygen con-
tents of as-deposited Co films were about 8 and 1 atom %, respec-
tively. However, the carbon content in the Co film decreased from
about 8 to about 4 atom % after in situ annealing at 400°C. In the
case of the in situ annealed Co film without a Ti capping layer,
polycrystalline CoSi2 characterized by diffraction peaks of ͑111͒,
͑220͒, and ͑311͒ was observed after ex situ annealing at 600°C.
However, the surface and interface morphologies of the CoSi2/Si
were very rough. In the case of the Co film with a Ti capping layer,
the interfacial morphologies of both TiN/CoSi2 and CoSi2/Si were
very planar and smooth after ex situ annealing at 700°C.
12. B. Lim, A. Rahtu, and R. G. Gordon, Nature (London), 406, 1032 ͑2000͒.
13. M. Leskelä and M. Ritala, Angew. Chem., Int. Ed., 42, 5548 ͑2003͒.
14. T. Suntola, Handbook of Thin Film Process Technoloy, 1st ed., Institute of Physics,
London ͑1995͒.
15. M. Ritala and M. Leskelä, Nanotechnology, 10, 19 ͑1999͒.
Downloaded on 2013-02-26 to IP 132.174.255.3 address. Redistribution subject to ECS license or copyright; see www.esltbd.org

Journal of The Electrochemical Society, 154 ͑10͒ H899-H903 ͑2007͒
H903
16. J. Kim, D. Kim, H. Park, and H. Jeon, J. Electrochem. Soc., 152, G29 ͑2005͒.
17. J. Kim, S. Kim, H. Kang, J. Choi, M. Cho, K. Chung, S. Back, K. Yoo, C. Bae, and
H. Jeon, J. Appl. Phys., 98, 094504 ͑2005͒.
18. K. Kim, K. Lee, T. Park, Y. Lee, J. Kim, S. Yeom, and H. Jeon, Jpn. J. Appl. Phys.,
Part 2, 46, L173 ͑2007͒.
19. G. J. M. Dormans, J. Cryst. Growth, 108, 806 ͑1991͒.
20. H. Rhee and B. Ahn, J. Electrochem. Soc., 146, 2720 ͑1999͒.
21. K. Maex and M. V. Rossum, Properties of Metal Silicide, IMEC, Leuven, Belgium
͑1995͒.
22. G. Kim, J. Kwak, S. Lee, and H. Baik, J. Appl. Phys., 85, 1503 ͑1999͒.
23. D. Sohn, J. Park, B. Lee, J. Bae, K. Oh, S. Lee, J. Byun, and J. Kim, Tech. Dig. -
Int. Electron Devices Meet., 1998, 1864.
Downloaded on 2013-02-26 to IP 132.174.255.3 address. Redistribution subject to ECS license or copyright; see www.esltbd.org