One-step electroless synthesis and properties of copper film deposited on silicon substrate

Article abstract of DOI:10.1002/pssa.200723305

A one-step electroless deposition method was developed to synthesize sponge-like copper film with favourable mechanical adhesion on silicon substrate. X-ray diffraction and X-ray photoelectron spectroscopy analysis and field emission scanning electron microscopy observations indicate that the asprepared copper film is polycrystalline and composed of copper nanoparticles with a diameter tanging from 45 nm to 60 nm, the thickness of the film being approximately 2 μm.Hydrogen obtained through the reaction of sodium hydroxide and silicon in ethylene glycol solution was employed to reduce in situ the copper ions to form metal copper films on the silicon substrate. The electrical and mechanical properties of the as-prepared copper film were measured. The results show that the as-prepared copper film possesses high resistivity and low Young's modulus due to its special structure.

Full text of DOI:10.1002/pssa.200723305

phys. stat. sol. (a) 205, No. 7, 1580–1584 (2008) / DOI 10.1002/pssa.200723305
a
p s s
www.pss-a.com
applications and materials science
One-step electroless synthesis
and properties of copper film
deposited on silicon substrate
Jing Ye, Qianwang Chen*, Ying Xiong, Yisheng Chai, and Peiran Chen
Hefei National Laboratory for Physical Science at Microscale and Department of Materials Science & Engineering,
University of Science & Technology of China, Hefei 230026, P.R. China
Received 3 July 2007, revised 24 March 2008, accepted 3 April 2008
Published online 4 June 2008
PACS 61.43.Gt, 62.20.de, 73.61.At, 81.15.Pq
*
Corresponding author: e-mail cqw@ustc.edu.cn, Phone: +86 551 3607292, Fax: +86 551 3607292
A one-step electroless deposition method was developed to Hydrogen obtained through the reaction of sodium hydroxide
synthesize sponge-like copper film with favourable mechani- and silicon in ethylene glycol solution was employed to re-
cal adhesion on silicon substrate. X-ray diffraction and X-ray duce in situ the copper ions to form metal copper films on the
photoelectron spectroscopy analysis and field emission scan- silicon substrate. The electrical and mechanical properties of
ning electron microscopy observations indicate that the as- the as-prepared copper film were measured. The results show
prepared copper film is polycrystalline and composed of cop- that the as-prepared copper film possesses high resistivity and
per nanoparticles with a diameter ranging from 45 nm to low Young’s modulus due to its special structure.
60 nm, the thickness of the film being approximately 2 µm.
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1 Introduction The deposition of metallic structures cause of the low processing temperature and the sub-
at a micro-/nanoscale on semiconductor substrates such as request for experimental equipment. Especially, electroless
silicon wafer has potential application in fabrication of deposition has emerged as the method of choice for metal-
electrical interconnections and microsensors in microelec- lization at present due to its fast deposition speed, good
tronics. For instance, copper film deposited on silicon sub- filling capability and good uniformity [16, 17]. However,
strate can be used as electrical interconnections ranging limitations still remain in electroless deposition; examples
from a few nanometres to several micrometres thick [1]. are the usage of formaldehyde, which acts as a reducing
Copper deposition has been developed rapidly recently [2– agent [18, 19], and HF solution, in which the copper can be
8] due to the potential application of copper with low elec- directly deposited through a redox reaction [20] with sili-
trical resistivity (~1.72 µΩ cm) in microelectronics and con without the need for reducing agents. Such methods
mircrosensors and high elecromigration resistance com- are environmentally hazardous due to the corrosiveness
pared with other metals [9]. Several methods for copper and high toxicity of both formaldehyde and HF. Recently,
metallization have been studied in previous reports, such as the deposition of a 50 nm thick layer of copper on silicon
the electroplating method, physical vapour deposition without the use of either formaldehyde or HF acid has been
(PVD) and chemical vapour deposition (CVD) [10–13]. reported. This method uses pre-sputtered deposited tita-
However, the costly experimental equipment and critical nium film, which is less toxic but more costly compared
experimental conditions of the PVD and CVD methods with formaldehyde and HF, as the reducing agent [21].
limit large-scale synthesis for industrial applications [14, Therefore, the deposition of copper films on specific sub-
15]. Therefore, it would be of interest if a preparation strate via a low-cost, environmentally friendly chemical
method controlling the crystal forms under mild conditions route is expected to be further developed.
were developed. Electroless deposition is considered to be
We report in this paper a one-step method for sponge-
a favourable method in the semiconductor industry, be- like copper film deposited on silicon substrate with favour-
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Original
Paper
phys. stat. sol. (a) 205, No. 7 (2008)
1581
able mechanical adhesion via a solvent thermal route, and deionized water to remove the impurities on the silicon
which is considered as a low-cost, non-toxic and mild wafer surface. In a typical procedure, 0.5 g sodium chlo-
chemical method for fabricating copper film. The primary ride (NaCl) and 0.4 g PVP (K30, M = 40000) were ultra-
w
principle used in this method is based on the reactivity of sonically dissolved in 10 ml EG, and stirred for 10 min to
silicon under basic conditions. In detail, silicon can react give a transparent viscous solution. After that, 0.05 mmol
with hydroxide ions to produce hydrogen, by which the cupric nitrate (Cu(NO ) ⋅3H O) was added into the solu-
3
2
2
cupric ion can be reduced:
tion with constant stirring followed by the addition of 1 ml
sodium hydroxide (0.5 mol/l) dropwise. The as-prepared
solution together with the pre-treated silicon wafer was
transferred into a Teflon-coated stainless steel autoclave
with a capacity of 15 ml. The autoclave was maintained at
160 °C for 48 h. After the autoclave cooled to room tem-
perature naturally, it was found that the silicon wafer was
covered with bright red surfaces and red-brown powder
was on the bottom of the autoclave. The silicon wafer was
then taken out of the solution and washed with deionized
water and alcohol several times and dried at room tempera-
ture.
-
Si + 2OH Æ SiO + H ≠ + 2e .
-
(1)
2
2
The standard electrode potential of copper is 0.3402 V:
2+ -
Cu + 2e Æ Cu ,
(2)
and thus the cupric ion can be deoxidized by hydrogen for
which the standard electrode potential is 0 V. So it is con-
ceivable to deposit copper on a silicon surface as long as
one controls the experimental parameters, such as the con-
centration of Cu2+, OH− and the dispersant, the reaction
temperature and so on.
The silicon wafer was characterized using X-ray dif-
fraction (XRD; Philips X-ray diffractometer with graphite
In our study, we successfully deposited sponge-like
copper film on silicon substrate using ethylene glycol (EG)
as the solvent, cupric nitrate as the Cu2+ source, sodium
hydroxide as the OH− source and polyvinylpyrrolidone
(PVP) as the dispersant. The as-prepared copper film is
polycrystalline and has uniformity in thickness. The elec-
tric and mechanical properties were investigated. The
sponge-like structure is responsible for high electric resis-
tivity and low Young’s modulus. To our knowledge, this is
the first time polycrystalline copper film has been depos-
ited on silicon substrate via a low-cost, non-toxic and mild
chemical route.
monochromatized Cu K radiation (λ = 1.54187 Å)),
α
X-ray photoelectron spectroscopy (XPS; VGESCALAB
MKΠ spectrometer with an excitation source of
Mg K = 1253.6 eV) and field emission scanning electron
α
microscopy (FE-SEM; JEOL JSM-6700 FSEM). The resis-
tance of the film deposited on silicon substrate was meas-
ured using the standard four-point probe method. The
Young’s modulus and hardness of the films were deter-
mined using a nanoindenter (Hysitron Com. Ms02100601).
3 Results and discussion The XRD pattern of the
sample is presented in Fig. 1a. Besides the (111), (222) dif-
fraction peaks of (111) orientation silicon wafer (JCPDS
89-5012), the three relatively weaker diffraction peaks can
be indexed as the (111), (200) and (220) diffraction peaks
of cubic copper with cell constant of 3.612 Å, which is
consistent with the values given in the standard card
(JCPD-04-0836). No other impurities, such as copper ox-
ide or copper silicate, were detected in the XRD analysis.
2 Experimental All the reagents were of analytical
purity, purchased from Shanghai Chemicals Company, and
used without further purification. The silicon wafer (heav-
ily boron-doped, 0.003–0.006 µΩ cm, p-type (111) orien-
tation) was pre-treated in sequence with isopropyl alcohol
(ultrasonic bath, 10–15 min), NH ⋅H O (25%)/H O
3
2
2
2
(30%)/H O (the ratio was 1:1:5 by volume, 60 °C, 30 min)
2
Figure 1 (online colour at: www.pss-a.com) a) XRD pattern of the as-prepared copper film prepared at 160 °C for 48 h. XPS spectra
of product: b) C 1s; c) N 1s.
www.pss-a.com
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

a
p s s
1582
Jing Ye et al.: One-step electroless synthesis and properties of copper film on silicon
It is concluded that the film deposited on the silicon wafer
The above results indicate that the sponge-like copper
surfaces is composed of phase-pure copper. In addition, film was formed at the copper–silicon interface area. On
according to the 1.11 intensity ratio of the first/second the basis of the data, a possible mechanism for the forma-
strongest diffraction peak (I /I ) for bulk Cu [22], the in- tion of the sponge-like copper film and the rough erode-
111 200
tensity ratio of I /I in the XRD pattern is about 2.3.
111 200
like surface of the silicon wafer can be tentatively pro-
The co-deposition phenomenon of other substances can posed. The formation process is illustrated in Scheme 1. In
be demonstrated by high-resolution XPS. The co-deposi- the first step, with hydroxide ions adsorbed on the surface
tion of PVP is confirmed by the XPS spectrum of C 1s of the wafer (Scheme 1, step a), the reaction between sili-
(Fig. 1b). The C 1s spectrum is composed of four peaks con and hydroxide ions occurs on the surface of the wafer
and the binding energies of these peaks are 284.6 eV (C–H (Eq. (1)). The dissolution of silicon dioxide in solution is
bonding), 286.0 eV (C–C bonding), 287.7 eV (C–N bond- responsible for the formation of the copper–silicon inter-
ing) and 289.0 eV (C=O bonding). These binding energies face:
can be attributed to four types of carbon atoms in PVP [23].
The co-deposition of NO− is confirmed by the XPS spec-
-
SiO + 2OH Æ SiO + H O .
2-
(3)
2
3
2
3
trum of N 1s (Fig. 1c). There are two nitrogen contribu-
tions in the N 1s core level spectrum: the binding energy
centred at 401.9 eV can be assigned to C–N bonding in
PVP while the binding energy centred at 411.1 eV can be
In the second step, the hydrogen atoms formed which have
strong reducing ability. It is quite easy to reduce the Cu(II)
ions. The continuous formation of copper deposited on the
substrate surface consists of three steps. First, the copper
ions interact with PVP. Second, copper ions exposed to the
hydrogen atmosphere are reduced to copper atoms on the
surface of the silicon substrate, and then nearby copper at-
oms aggregate at close range to form the primary nanopar-
ticles (Scheme 1, step b). Finally, the primary nanoparti-
cles interact with PVP deposited on the silicon substrate,
some of which coalesce with other nearby primary
nanoparticles to form large aggregates (Scheme 1, step c).
The powder on the bottom of the autoclave may be a result
of the reaction between copper ions and hydrogen atmos-
phere which diffuses to the solution. For the two reasons
that the copper nanoparticles interact with PVP and the
rough surface morphology of the substrate, the as-prepared
copper films formed by the deposition of copper nanopar-
ticles are not compact and possess special sponge-like
structures (Scheme 1, step d). And then the copper film
deposited on the silicon substrate is formed (Scheme 1,
step e). During the whole deposition process, PVP is ex-
pected to inhibit the further aggregation of copper nanopar-
ticles [24] and NaCl slows down the deposition speed of
copper [25].
assigned to NO− adsorption.
3
The morphology of the copper film is presented in
Fig. 2. Figure 2a shows the representative morphology of
the film, which appears smooth and homogeneous. Fig-
ure 2b shows a magnified FE-SEM image of the film. The
film is composed of loosely stacked copper nanoparticles
45–60 nm in size. With more careful observation, it can be
seen that some particles aggregate together to form larger
nanoblocks. Figure 2c shows a low-magnification FE-SEM
image of the cross-section of the film. The film has uni-
formity in thickness and attaches to the substrate closely.
Figure 2d shows a magnified cross-sectional image of the
film, from which the tri-layered structure can be observed
clearly. The three layers from the top down are sponge-like
copper film, copper–silicon interface area and silicon sub-
strate, respectively. The average thickness of the copper
film was calculated as approximately 2.2 µm.
To investigate the mechanical and electric properties of
the as-prepared copper film, the Young’s modulus and
electrical resistance of as-prepared sample was measured
using nanoindentation and four-probe measurements, re-
Figure 2 a) Representative FE-SEM image of the surface mor-
phology of copper film deposited on silicon substrate; b) magni-
fied image of the film; c) representative FE-SEM image of the
cross-section of the same copper film; d) magnified image of the
cross-section showing the tri-layered structure.
Scheme 1 Formation process of the copper film deposited on
silicon substrate.
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.pss-a.com

Original
Paper
phys. stat. sol. (a) 205, No. 7 (2008)
1583
Table 1 Plunge depth (h ), force (P ), Young’s modulus (E )
c
max
r
and hardness (H) values of the copper film (at different locations).
location
h (nm)
c
P
(µN)
E (GPa) H (GPa)
r
max
1
424
451
461
474
514
653
751
769
870
948
1013
1167
708
223
205
274
174
257
233
282
276
245
234
191
319
243
0.58
0.29
1.01
0.83
0.42
0.24
0.78
0.52
0.18
0.16
0.27
0.28
0.46
0.040
0.033
0.042
0.026
0.032
0.019
0.017
0.016
0.011
0.0093
0.0067
0.0059
0.021
2
3
4
5
6
7
8
9
10
11
12
Figure 3 Resistivity of copper film as a function of temperature
from 15 to 300 K.
average value
spectively. The submicrometre nanoindentation of thin
films on substrates is a common method to measure the
Young’s modulus and hardness [26, 27]. After the load–
deflection and unload–deflection curves were obtained by
the nanoindenter, the Young’s modulus and hardness val-
ues could be determined by fitting the experimental curves
through the nanoindenter software. As shown in Table 1,
the set of Young’s modulus values obtained for the copper
film (at different locations) exhibits a scatter showing that
the film is not compact and homogeneous. The average
value of Young’s modulus of the copper film is approxi-
mately 0.46 GPa, which is much lower than the value of
bulk polycrystalline copper (130 GPa), bulk copper film
reported by Jamting et al. [28] and copper films prepared
through various methods. For example, Read [29] meas-
ured electron-beam deposited and sputtered copper film us-
ing a nanoindenter, and the Young’s modulus was 99 GPa
and 109 GPa, respectively, and Farhat et al. [30] reported a
Young’s modulus of 102.46 2.5 GPa of as-sputtered
copper film by the nanoindentation method. The special
sponge-like structure of our prepared copper film plays an
important role in the decreased Young’s modulus and
hardness values and leads to little regularity between
plunge depth (h ) and plunge force (P ).
where ρ is the measured electrical resistivity, ρ is the re-
0
sidual resistivity and J and k are constants [31]. Usually, ρ
0
is dependent only on the amount of impurities and defect
concentration in pure metals. In our work, due to many
voids existing in the copper film, the value of ρ got to
0
650 µΩ cm. After subtracting ρ , the room temperature
0
electrical resistivity of the copper film is about 300 µΩ cm,
which is still significantly higher than that of bulk copper
(1.72 µΩ cm) and other films prepared by different meth-
ods, such as 32.34 0.50 µΩ cm of 11.5 nm thick copper
film thermally evaporated onto ~500 nm thick SiO on
2
Si(100) substrate [32] and 22 µΩ cm of 11 nm thick copper
film deposited by ion beam deposition [33]. The electrical
resistance in a metal thin film arises due to electron scatter-
ing by phonons, point defects, impurities, grain boundaries,
film surfaces and interfaces. First of all, the contact resis-
tance of nanoparticles maybe very high due to their partial
embedding in PVP impurities, which was demonstrated by
XPS analysis. Also, the thickness of the as-prepared sam-
ple (2.2 µm) is larger than the MFP of copper (39 nm).
And surface and interface scattering needs to be considered
in the increase of the resistivity of the copper film. Elec-
tron scattering originating from grain boundaries would re-
sult in the increase in resistivity of the copper film. The ex-
istence of a large amount of pores in the film leads to a de-
crease of the efficient cross-sectional area of copper which
notably increases the resistivity of the copper film. More-
over, the existence of many nanopores the width of which
is very small can also result in increased width of the con-
ducting paths, which is one of the reasons causing the high
resistivity. In general, the special sponge-like structure of
the copper film leads to the high electrical resistivity which
is different from samples prepared by other deposition
methods.
c
max
Due to the relatively large thickness of the as-prepared
film (2.2 µm) which is far larger than the mean free path
(MFP) of copper (39 nm), a body transport mechanism
should be applied, so the resistivity can be calculated using
the formula R = ρl/s (where R is resistance, l is the length
between the inner contacts and s is the cross-sectional area).
Figure 3 displays the temperature dependence of electrical
resistivity (ρ) of the copper film. It has a value of
950 µΩ cm at room temperature, decreasing linearly with
on cooling from 300 to 50 K. Below 50 K, it gradually
saturates to a value of about 625 µΩ cm at 0 K. Electrical
resistivity of a pure metal at low temperatures can be rep-
resented by the formula
4 Conclusions In this paper we have developed a
one-step deposition method to prepare copper films on sili-
con substrates. Our results indicate that the as-prepared
copper film has a sponge-like structure, about 2.2 µm thick
k
ρ = ρ + JT ,
0
(4)
www.pss-a.com
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

a
p s s
1584
Jing Ye et al.: One-step electroless synthesis and properties of copper film on silicon
[12] N. Awaya and Y. Arita, Jpn. J. Appl. Phys. Part 1 30, 1813
and composed of polycrystalline copper nanoparticles. Due
to its special sponge-like structure, the as-prepared copper
film possesses higher resistivity (300 µΩ cm) and lower
Young’s modulus (0.46 GPa) than bulk copper and the
films prepared by other methods. The preparation of more
compact and homogeneous films having lower electrical
resistivity will be the focus in our future work. It is also
expected that these copper films may have potential appli-
cations in microelectronics, not only for the requirement of
mild reaction conditions, but also for extension to other
metals such as silver, platinum and so on to deposit on the
silicon substrate.
(1991).
[13] J. J. Bladon, A. Lamola, F. W. Lytle, W. Sonnenberg, J. N.
Robinson, and G. Philipose, J. Electrochem. Soc. 143, 1206
(1996).
[14] T. Nguyen, L. J. Charneski, and D. R. Evans, J. Electro-
chem. Soc. 144, 3634 (1997).
[15] B. Johnson, R. Amster, and L. Vanasupa, J. Electron. Mater.
27, 923 (1998).
[16] E. K. Yung, T. L. Romankiw, and R. C. Alkire, J. Electro-
chem. Soc. 36, 206 (1989).
[17] J. T. Beechinor, M. O’Reilly, J. C. Patterson, S. Lynch,
E. Lafferty, P. V. Kelly, and G. M. Crean, Mater. Res. Soc.
Symp. Proc. 406, 353 (1996).
Acknowledgement This work was supported by the Na-
tional Natural Science Foundation of China.
[18] H.-H. Hsu, K.-H. Lin, S.-J. Lin, and J.-W. Yeh, J. Electro-
chem. Soc. 148, C47 (2001).
[19] C. J. Weber, H. W. Pickering, and K. G. Weil, J. Electro-
chem. Soc. 144, 2364 (1997).
References
[20] F. W. Kern Jr., M. Itano, I. Kawanabe, M. Miyashita, R. W.
Rosenberg, and T. Ohmi, in: Proceedings Advanced Wet
Chemical Processing II (Ultra Clean Society, Tokyo, 1990),
p. 113.
[1] X. Ye, M. D. Bonte, J. P. Celis, and J. R. Roos, J. Electro-
chem. Soc. 139, 1592 (1992).
[2] R. Padiyath, J. Seth, S. V. Babu, and L. J. Matienzo, J. Appl.
Phys. 73, 2326 (1993).
[21] R. K. Aithal, S. Yenamandra, R. A. Gunasekaran, P. Coane,
and K. Varahramyan, Mater. Chem. Phys. 98, 95 (2006).
[22] W. P. Davey, Phys. Rev. 25, 753 (1925).
[23] B. K. Park, S. Jeong, D. Kim, J. Moon, S. Lim, and J. S.
Kim, J. Colloid Interface Sci. 311, 417 (2007).
[24] I. Pastoriza-Santos and L. M. Liz-Marzan, Langmuir 18,
2888 (2002).
[3] C. S. Liu and L. J. Chen, J. Appl. Phys. 74, 5501 (1993).
[4] J. Choi, Z. Chen, and R. K. Singh, J. Electrochem. Soc. 150,
C563 (2003).
[5] E. T. Krastev, L. D. Voice, and R. G. Tobin, J. Appl. Phys.
79, 6865 (1996).
[6] J. T. Drotar, T. M. Lu, and G. C. Wang, J. Appl. Phys. 96,
7071 (2004).
[25] F. A. Harraz, T. Sakka, and Y. H. Ogata, Electrochim. Acta
47, 1249 (2002).
[7] V. M. Dubin, Y. Shacham-Diamand, B. Zhao, P. K.
Vasudev, and C. H. Ting, J. Electrochem. Soc. 144, 898
(1997).
[26] M. F. Doerner and W. D. Nix, J. Mater. Res. 1, 601 (1986).
[27] Y. Zhou, C. S. Yang, J. A. Chen, G. F. Ding, W. Ding,
L. Wang, M. J. Wang, Y. M. Zhang, and T. H. Zhang, Thin
Solid Films 460, 175 (2004).
[8] M. Pletea, W. Brückner, H. Wendrock, and R. Kaltofen,
J. Appl. Phys. 97, 054908 (2005).
[9] J. R. Shi, S. P. Lau, Z. Sun, X. Shi, B. K. Tay, and H. S. Tan,
Surf. Coat. Technol. 138, 250 (2001).
[28] A. K. Jamting, J. M. Bell, M. V. Swain, and N. Schwarzer,
Thin Solid Films 304, 308 (1997).
[10] Y. Zhang, E. Balaur, and P. Schmuki, J. Electroceram. 16,
65 (2006).
[29] D. T. Read, Meas. Sci. Technol. 676, 9 (1998).
[30] Z. N. Farhat, Y. Ding, and D. O. Northwood, Surf. Coat.
Technol. 24, 89 (1997).
C. T. Lin and K. L. Lin, J. Mater. Sci. Mater. Electron. 15,
757 (2004).
[31] F. B. Hildebrand, Introduction to Numerical Analysis
(McGraw-Hill, New York, 1956), Chap. 7.
J. Reid, S. Mayer, E. Broadbent, E. Klawuhn, and K. Ashti-
ani, Solid State Technol. 43, 86 (2000).
[32] H.-D. Liu, Y.-P. Zhao, G. Ramanath, S. P. Murarka, and
G.-C. Wang, Thin Solid Films 384, 151 (2001).
[33] J.-W. Lim and M. Isshiki, J. Appl. Phys. 99, 094909 (2006).
[11] B. Chin, B. Sun, T. Chiang, D. Angelo, I. Hashim, Z. Xu,
S. Edelstein, and F. Chen, Solid State Technol. 41, 41
(1998).
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.pss-a.com