Kinetic regularities of the chemical vapor deposition of nickel layers from bis-(ethylcyclopentadienyl)nickel

Article abstract of DOI:10.1134/S1070427212050102

Physicochemical regularities of the nickel layers chemical vapor deposition from bis-(ethylcyclopentadienyl)nickel were studied. Dependences of the growth rate of nickel layers on the deposition temperature, gas-fl ow linear rate, partial pressures of reagents, and substrate roughness, and also dependences of the thickness of a grown layer on time and on the position of a substrate on a susceptor were obtained.

Full text of DOI:10.1134/S1070427212050102

ISSN 1070-4272, Russian Journal of Applied Chemistry, 2012, Vol. 85, No. 5, pp. 742−746. © Pleiades Publishing, Ltd., 2012.
Original Russian Text © V.S. Protopopova, S.E. Alexandrov, 2012, published in Zhurnal Prikladnoi Khimii, 2012, Vol. 85, No. 5, pp. 741−745.
PHYSICOCHEMICAL STUDIES
OF SYSTEMS AND PROCESSES
Kinetic Regularities of the Chemical Vapor Deposition of Nickel
Layers from Bis-(Ethylcyclopentadienyl)nickel
V. S. Protopopova and S. E. Alexandrov
St. Petersburg State Polytechnical University, St. Petersburg, Russia
e-mail: vera_protopopova@mail.ru
Received April 6, 2012
Abstract—Physicochemical regularities of the nickel layers chemical vapor deposition from bis-
(ethylcyclopentadienyl)nickel were studied. Dependences of the growth rate of nickel layers on the deposition
temperature, gas-flow linear rate, partial pressures of reagents, and substrate roughness, and also dependences of
the thickness of a grown layer on time and on the position of a substrate on a susceptor were obtained.
DOI: 10.1134/S1070427212050102
One of promising directions of using nickel thin layers
is their application as catalytic layers for the synthesis of
carbon nanostructures, nanotubes [1] and graphenes [2],
and also as active magnetic layers [3] in various micro-
electronic devices.
lation equipped by a horizontal reactor with “hot walls”
intended for realization of CVD processes under reduced
pressure. The installation is described in detail in [10].
The experimental study of the nickel layers deposition
was carried out in the temperature range 665–925 K at the
following gas-phase parameters: partial pressure of bis-
(ethylcyclopentadienyl)nickel, from 8 up to 100 Pa, that
of hydrogen, from 0 up to 350 Pa, a flow of argon used as
a gas-carrier through the evaporator, 50 cm3 min–1, total
gas flow 100 cm3 min–1, and total pressure 840 Pa. Typical
deposition time was 60 min.As substrates we used n-type
silicon wafers (111) and (100) with a specific electrical
resistivity of 7.5 Oh cm, and polycrystalline Nb foil.
The method of chemical vapor deposition (CVD) al-
lows obtaining high-quality coats on large surface areas
and on products with complicated geometrical shapes,
including micro- and nanoporous surfaces. Organometal-
lic compounds are widely used as precursors in CVD pro-
cesses owing to their high volatility and low decomposi-
tion temperatures. In particular, carbonyls, β-diketonates,
acetylacetonates, and bis-π-cyclopentdienyl compounds
are used for obtaining nickel layers [4–8]. Traditionally
the deposition of metal layers is carried out in the pres-
ence of hydrogen [6, 7], as an addition of Н2 can reduce
the degree of uncontrollable alloying resulting layers by
carbon.
The results of preliminary experiments point to an es-
sential decrease in the thickness of layers in the direction
from the centre to the edges of a substrate (Fig. 1), which
probably is connected with a concentration gradient of
the precursor in a reactor. In this connection we have
chosen a field in a center of susceptor for the arrangement
of substrates, where the thickness of layers was constant
to within ±10%.
A principal possibility of using bis-(ethylcyclo-
pentadienyl)nickel [(EtCp)2Ni] as a promising precursor
for obtaining nickel layers was shown in [9]. The aim
of the present work was the study of physicochemical
regularities of the process including the determination
of the limiting stage type.
To determine the thickness of resulting Ni layers, we
have developed a technique based on measuring inten-
sity of a characteristic line of the element in the X-ray
fluorescence (XRF) analysis. The technique was realized
on a “Spectroscan MAX-GV” XFR spectrometer. The
spectrometer was graduated by standard samples. The
EXPERIMENTAL
Nickel layers were deposited in a technological instal-
742

KINETIC REGULARITIES OF THE CHEMICAL PRECIPITATION OF NICKEL LAYERS
743
both dependences is the presence of an incubation period
of 15 min at the deposition temperature 795 K and of 1
min at 845 K. It is necessary to note that its presence
cannot be connected with a time delay between opening
the evaporator and delivering the precursor to a reaction
zone, which is equal to 15 s for the used geometrical size
of the reactor and the chosen technological conditions.
Results of studying temperature dependences of the
growth rate of nickel layers on various substrates at vari-
ous partial pressures of the metalorganic precursor are
given in Figs. 3a and 3b.
Fig. 1. Distribution of thickness h (nm) of a grown nickel layer
across a substrate at Тdep = 800 K. (a) A substrate width (nm).
It is seen from Fig. 3a (curves 1 and 2) that at 75 Pa the
temperature dependence of the growth rate constructed in
thickness of nickel layers on these samples was measured
by the height of a scratch with the use of an atomic-power
microscope (Solver-PRO “NT-MDT”).
(a)
To determine type of the limiting stage of the nickel
layers deposition from (EtCp)2Ni, we have found experi-
mentally dependences of growth rate of coats on various
technology parameters.
The cases of nonstationary growth rates of metal lay-
ers are known from the literature [5, 11]. Special experi-
ments have shown that for the deposition temperatures
Тdep 795 and 845 K the thickness of deposited layers
linearly depends on the process duration (Fig. 2, curves 1
and 2) within the limits of experimental error, which did
not exceed 10%. Thus, the rate of layers deposition can
be determined by dividing the thickness of a deposited
layer by the experiment duration. The average growth
rate of nickel layers at 795 K was 152 nm h–1 and at
845 K, 190 nm h–1. Furthermore, a special feature of the
(b)
Fig. 3. Dependence of the growth rate V (nm h–1) of nickel
layers deposition on deposition temperature Т (K) inArrhenius
coordinates. (a) Р[(EtCp)2Ni] = 75 Pa, Р(Н2) = 210 Pa. Sub-
strate: (1) n-type Si(100), (2) p-type Si, (3) Nb. (b) Substructure
n-type Si(111), Р(Н2) = 210 Pa. Р[(EtCp)2Ni] (Pa): (1) 75, (2)
50, (3) 23.
Fig. 2. Dependence of thickness h (nm) of obtained nickel layers
on the deposition time τ (min). Р[(EtCp)2Ni] = 75 Pa, Р(Н2) =
210 Pa. Тdep (K): (1) 795, (2) 845.
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 85 No. 5 2012

744
PROTOPOPOVA, ALEXANDROV
likely with gas flow depletion along the reactor, which
is illustrated by the plot of the distribution of the
deposited layer thickness along a substrate (Fig. 4,
curve 2). This is also responsible for lower values of
the apparent activation energies for the substrates lay-
ing in the second raw along a gas flow. At deposition
temperatures below the temperatures corresponding
to the bend no gas flow depletion along substrates oc-
curs. It is proved by the distribution of the deposited
layer thickness along the reactor (Fig. 4, curve 1).
Arrhenius coordinates has a bend in the region of 810 K.
The qualitative character of the temperature dependences
does not change on the variation of (EtCp)2Ni partial pres-
sure (Fig. 3, curves 1–3). For a temperature range below
the bend the apparent activation energy of the process
was 115 ± 6 kJ mol–1 for the whole studied range of re-
agent partial pressures (23–75 Pa) and also for all silicon
substrates. It was found that for niobium foil substrates
the apparent activation energy was slightly less and was
equal to 99 ± 5 kJ mol–1 at the (EtCp)2Ni partial pressure
of 75 Pa (Fig. 3a, curve 3). Thus, the rate of nickel layers
deposition is independent of the crystallographic orienta-
tion of substrates for the whole range of technological
parameters under consideration.
The increase in the temperature corresponding
to the bend with decreasing metalorganic precur-
sor partial pressure is most likely connected with
the fact that in this case the gas flow depletion is
not so fast because of a lower conversion degree of
the precursor, as demonstrated by the exponential
character of the decrease in the growth rate of layers
with decreasing Р[(EtCp)2Ni] (Fig. 5). In spite of the
fact that the gas flow depletion begins later as the
(EtCp)2Ni concentration in a gas phase decreases,
the process of depletion along a susceptor increases
with temperature more intensively, as demonstrated
by a faster decrease in the growth rate and smaller
values of apparent activation energies.
It was found that the temperature corresponding to
the bend is shifted to the region of higher values as the
precursor partial pressure decreases (Fig. 3b). Thus, at
the (EtCp)2Ni partial pressure of 75 Pa the temperature
of 810 K corresponds to the bend, at 50 Pa, 850, and, at
23 Pa, 860. Furthermore, it was found that the slope of
a high-temperature section of the plot varies as a precursor
partial pressure changes.As the partial pressure increases
from 23 up to 75 Pa, the value of the apparent activation
energy increases from –42.3 ± 2 up to 37 ± 2 kJ mol–1 for
substrates arranged in the first raw along the gas flow; and
from –130.9 ± 6 up to –9.5 ± 0.5 kJ mol–1 for substrates
in the second raw along a gas flow. It was shown also that
in the high-temperature region the deposition rate is inde-
pendent of the crystallographic orientation of substrates.
The data of the mass-spectroscopic study of the
gas phase completely confirm the ssuggestion about
the gas flow depletion. It was experimentally shown
that the height of the peak corresponding to the mass
number 245 a.m.u. and characterizing the change of
The decrease in the apparent activation energy in
the temperature region over the bend is connected most
Fig. 5. Dependence of the growth rate V (nm h–1) of depo-
sition of nickel layers on the partial pressures (Pa) of bis-
(ethylcyclopentadienyl)nickel Р[(EtCp)2Ni] (1–3) at Р(Н2) =
210 Pa and of hydrogen P(H2) (4) at Р[(EtCp)2Ni] = 75 Pa. Тdep
(K): (1) 745; (2), (4) 795; (3) 845.
Fig. 4. Distribution of the deposited layer thickness h (nm)
along a substrate in the flow direction. Р[(EtCp)2Ni] = 75 Pa,
Р(Н2) = 210 Pa. (l) the substrate length (mm). (I) left-hand raw,
(II) right-hand raw of substrates. Тdep (K): (1) 800, (2) 855.
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 85 No. 5 2012

KINETIC REGULARITIES OF THE CHEMICAL VAPOR DEPOSITION
745
the precursor molecules (EtCp)2Ni concentration drops
up to zero at 740 K. Thus, at high temperatures the
delivery of reagents to the susceptor starts to play an
essential role, whereas at low temperatures this effect
was not observed.
Table 1. Growth rate of nickel layers at various gas-flow linear
rates in a reactor and constant molar flows of precursors and
gases. Experimental conditions: Тdep = 800 K, Р[(EtCp)2Ni] =
75 Pa, Р(Н2) = 210 Pa, p-tipe Si substrates
Reactor diameter,
mm
Growth rate,
nm h–1
Linear rate, m h–1
As it has been found, in the range of deposition tem-
peratures below the temperature of the bend the value
of the apparent activation energy was 115 ± 6 kJ mol–1
for the whole studied range of partial pressures and all
silicon substrates. This value is not characteristic for
diffusion processes and the stages limited by delivery
of reagents, but it is most typical of kinetic regimes of
passing heterogeneous processes. The kinetically limited
character of the process is proved also by the indepen-
dence of the growth rate of layers from the linear gas-flow
rate at 800 K (Table 1), which was varied by decreasing
the reactor diameter by application of quartz inserts. In
such experiments molar gas flows and partial pressures
of reagents were maintained at constant levels, and the
presence of the dependence would point to the diffusion
regime of passing the process.
47
41
31
22
3.46
4.55
110.1
107.3
104.4
107.9
7.96
15.79
Table 2. Growth rate of nickel layers at various substrate
roughnesses. Experimental conditions: Тdep = 775 K,
Р[(EtCp)2Ni] = 75 Pa, Р(Н2) = 210 Pa, n-type Si substrates
Average roughness
Growth rate, nm h–1
Ra, nm
0.3
57.8
3
56.8
205
275
80.1
Furthermore, it was found experimentally that at
temperature 775 K the rate of nickel layers deposition
increases as the roughness of substrates increases
(Table 2). It is known [12] that an increase in the roughness
of substrates results in a decrease in the growth rate of
layers deposition, if the process is limited by a diffusion
stage. When the process is limited by the delivery of
precursors in a deposition zone, the actual surface area
does not affect the growth rate of the deposited layers.
102.9
at Р[(EtCp)2Ni] = 75 Pa. As partial pressures of reagents
increase further, the growth rate of nickel layers slowly
decreases.
The exponential character of the growth rate depen-
dences on partial pressures of reagents in the case of
their low values is not characteristic for CVD processes
limited by chemisorption stages (including desorption) or
directly by a chemical reaction [12]. It is also proved by
the published values of apparent activation energies for
chemical vapor deposition of nickel layers from organo-
metallic compounds. Thus, for the processes limited by
the adsorption stages the characteristic values of the ap-
parent activation energy of 25 [7] and 50 kJ mol–1 [6] are
lower than the observed value of 115 ± 6 kJ mol–1. The
process limitation by the stage of the surface diffusion
of nickel atoms is also unlikely, as it is characterized by
the activation energy of 61 kJ mol–1 [13], which is also
less than the observed value.
To refine the limiting stage nature, we have studied the
dependence of the growth rate of nickel layers deposition
on partial pressures of the precursors (EtCp)2Ni and hy-
drogen. It was found experimentally that the dependences
of growth rate of nickel layers on the (EtCp)2Ni partial
pressure are the same for various ranges of deposition
temperatures (Fig. 5, curves 1–3) corresponding to the
temperatures before the bend (745 and 795 K) and after
it (845 K) for different types of substrates: silicon and
niobium. The dependence of the growth rate of layers
on the hydrogen partial pressure (Fig. 5, curve 4) has
a similar shape. A special feature of the found depen-
dences is the exponential character of the growth in the
Р[(EtCp)2Ni] range 8–70 Pa at Р(Н2) = 210 Pa and in
the Р(Н2) range 0–200 Pa at Р[(EtCp)2Ni] = 75 Pa. The
maximal growth rate of layers was observed in the range
of (EtCp)2Ni partial pressure 60–80 Pa at Р(Н2) = 210 Pa
and in the range of hydrogen partial pressures 200–220 Pa
In this connection it is most probable that the process
of heterogeneous nucleation is the limiting stage, as these
are such processes for which the exponential dependences
are characteristic for the rate of layers growth and the rate
of nucleation on the supersaturation value [14], which is
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 85 No. 5 2012

746
PROTOPOPOVA, ALEXANDROV
equal to a partial pressure of a reagent for irreversible
processes. The essential role of the nucleation is also
proved by the presence of an incubation period of 15 min
at 795 K and of 1 min at 845 K. It is possible that the
presence of the incubation period is connected with the
slowest process of the formation of “supercritical”-size
nuclei. It is proved by the results of studying morphology
of layers deposited at early stages, which have shown
the presence of nickel nanoparticles of equal size ar-
ranged separately. The suggestion about a homogeneous
nucleation is excluded, as a homogeneous formation of
nanopowders was not observed within all studied ranges
of technology parameters.
2000, vol. 88, no. 5, p. 4898.
2. Pollard,A.J., Nair, R.R., Sabki, S.N., et al., J. Phys. Chem.
C. Letters, 2009, vol. 113, no. 38, p. 16565.
3. Shalygina, E.E., Kozlovskii, L.V.,Abrosimova, N.M., and
Mukasheva, M.A., Fiz. Tverd. Tela, 2005, vol. 47, no. 4,
p. 660.
4. Andreev, O.N., Bespalenko, O.P., Marin, V.P., et al., Metal-
loorg. Khim., 1988, vol. 1, no. 5, p. 1184.
5. Fraser, B., Hampp,A., and Kaesz, H.D., Chem. Materials,
1996, vol. 8, p. 1858.
6. Brissonneau, L. and Vahlas, C., Chem. Vap. Deposition,
1999, vol. 5, no. 4. p. 135.
7. Kang, J.-K. and Rhee, S.-W., J. Mater. Res., 2000, vol. 15,
no. 8, p. 1828.
CONCLUSIONS
8. Ishikawa, M., Kada, T., Machida, H., et al., Jаpаn. J. Appl.
Phys., 2004, vol. 43, no. 4B, p. 1833.
(1) The process of nickel layers chemical vapor depo-
sition from bis-(ethylcyclopentadienyl)nickel under low
pressure proceeds in a kinetic regime in conditions under
study. The process is limited by the stage of heterogeneous
nucleation. The apparent activation energy of the process
is 115 ± 6 kJ mol–1.
9. Alexandrov, S.E. and Protopopova, V.S., J. Nanosci.
Nanotechnol., 2011, vol. 11, p. 8259.
10. Protopopova, V.S. and Alexandrov, S.E., Nauchno-Tekhn.
Vedomosti SPbGPU. Informatika. Telekommunikatsii.
Upravlenie, 2011, no. 3, pp. 145-150.
11. Premkumar, P.A., Bahlawane, N., and Kohse-Hoing-
(2) It has been shown that at high deposition tempera-
tures the gas phase is depleted along the reactor volume.
It results in a decrease in the growth rate.
haus, K., Chem. Vap. Deposition, 2007, vol. 13, p. 219.
12. Crystal Growth. Theory and Techniques, Goodman, C.H.L.,
Ed., London: Plenum Press, 1974, vol. 1.
13. Davydov, S.Yu., Fiz. Tverd. Tela, 1999, vol. 41, no. 1, p. 11.
REFERENCES
14. Chenthamarai, S., Jayaraman, D., and Subramanian, C.,
1. Choi, Y.C., Shin, Y.M., Lim, S.C., et al., J. Appl. Phys.,
Materials Chem. a. Phys., 2000, vol. 63, p. 163.
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 85 No. 5 2012