Você está na página 1de 9

Mechanical properties of intermetallic compounds in

electrodeposited multilayered thin film at small scale by


nanoindentation
Abu Zayed Mohammad Saliqur Rahmana,, Pay Ying Chiaa , A. S. M. A.
Haseeba,
a

Department of Mechanical Engineering, University of Malaya


50603 Kuala Lumpur, Malaysia

Abstract
Mechanical properties of intermetallic compounds (IMCs) which were formed
in electrodeposited Cu/Sn and Cu/Ni/Sn multilayered thin film have been
investigated. The layers of Cu, Sn and Ni were formed by electrodeposition
technique using copper pyrophosphate, tin methanesulfonic and nickel Watts
baths, respectively. After synthesis,samples were subjected to high temperature aging at 150 C for 168 h. Two different types of intermetallics Cu3 Sn
and Cu6 Sn5 were formed in Cu/Sn. After adding ultra thin layer of Ni (70
nm) in between Cu and Sn layers, (Cu, Ni)6 Sn5 was formed after aging at
similar condition to that of Cu/Sn. Tin whisker growth was not observed in
both samples after preserving the samples in air for 365 days. Hardness and
elastic moduli of all three different types of IMCs were measured by using a
Hysitron Triboindenter 750 Ubi system. Hardness of the three IMCs Cu3 Sn,
Cu6 Sn5 , (Cu, Ni)6 Sn5 and Cu were found to be 5.99, 6.61, 7.43 and 1.55 GPa
respectively. The addition of Ni suppressed the growth of Cu3 Sn greatly.
This is expected to lead to better reliability of electronic interconnections as
Cu3 Sn is often associated with void formation.
Keywords: Multilayered structure, Intermetallic compound, Mechanical
Properties, Nanoindentation, Thin film, Interconnects

Corresponding author
Email addresses: zayed@um.edu.my (Abu Zayed Mohammad Saliqur Rahman),
chiapayying@siswa.um.edu.my (Pay Ying Chia), haseeb@um.edu.my (A. S. M. A.
Haseeb)

Preprint submitted to Materials Letters

February 4, 2015

1. Introduction
Growth of tin whisker has detrimental effect on electrical devices. Although effect of tin whisker in electronics is known since early of the 20th
century, growth mechanism is still not well understood. Rapid growth of tin
whisker on lead-free Sn solder alloy has been reported elsewhere [1, 2, 3, 4].
Recently, Young Sun et al. suggested the mechanism of tin whisker growth
[5]. Traditionally, lead was used to suppress the growth of tin whisker. However, lead was banned due to its detrimental effects on human health by European Union. Many electronic devices failed because of tin whisker growth
which causes short-circuit as it can conduct electricity to the neighboring
parts within the device. It may cause major havoc in large facility such
as nuclear reactor, satellite [6]. As future microelectronics inevitably moves
toward nano-scale devices, this could be one of the important issues to be
addressed to ensure reliability of the electronic devices. Multilayered thin
film interconnects could be one possible solution to this problem. This could
be applicable in miniature devices with a possibility to prevent tin whisker
growth.
We used electrodeposition technique to synthesize Cu/Sn and Cu/Ni/Sn
multilayered thin film[7]. Reliability of the interconnects greatly depends
on the mechanical properties of the IMCs. Therefore, it is very important
to understand the mechanical properties of these IMCs in thin film. Thus
we investigated mechanical properties of the IMCs by using nanoindentation
technique [8]. To the best of our knowledge, there is no study which reports
on the mechanical properties of IMCs formed on electrodeposited Cu/Sn and
Cu/Ni/Sn multilayered thin film at small scale by nanoindentation.
2. Experimental
The layers of Cu, Sn and Ni were formed by electrodeposition using copper
pyrophosphate, tin methanesulfonic and nickel Watts baths, respectively [9].
Deposition current density was set at 10 mA/cm2 for the copper bath and 20
mA/cm2 for tin and nickel baths. Cu/Sn and Cu/Ni/Sn thin films were prepared in the sequence of Cu/Sn/Cu/Sn/Cu and Cu/Ni/Sn/Ni/Cu/Ni/Sn/Ni
/Cu, respectively. After preparation, short reflow was done in a FT-02 convection reflow oven at 250 C for 60 s followed by high temperature aging at
2

150 C for 168 h in a Memmert oven. Samples were then mounted in epoxy
resin for microstructure investigation. Progress of solid state reaction in the
multilayered films was monitored by using a FEI Quanta 450 field emission
scanning electron microscope (FESEM) and energy dispersive x-ray spectroscopy (EDX). Mechanical properties of intermetallic compounds (IMCs)
were obtained by using a Hysitron Triboindenter 750 Ubi system at ambient
temperature. A Berkovich diamond indenter tip was used to measure hardness and elastic muduli of the IMCs. Indenter tip was held at the peak load,
4 mN for 20 s at each indent spot maintaining minimum drift rate at <0.05
nm/s. Hardness, H was defined by:
Pmax
A

H=

(1)

where Pmax is maximum indentation load and A is the resultant area of


contact corresponds to that load. The reduced elastic modulus Er , was
defined as:

s
Er =
(2)
2 A
where S is the stiffness of unloading curve and A is the projected area. Initial
unloading stiffness contact can be defined as:
S=

dP
dH

(3)

Elastic modulus can be obtained from the following equation if the sample
and diamond indenter tip are subjected to elastic deformations.
1
1 vs2 1 vi2
=
+
Er
Es
Ei

(4)

whereEs and vs are elastic modulus and Poissons ratio of the sample, respectively. Ei and vi correspond to elastic modulus and Poissons ratio of the
diamond indenter, respectively. For a diamond tip, Ei = 1140 GPa and vi =
0.07 [10]. On the other hand, vs was assumed to be 0.299 for Cu3 Sn, 0.309
for Cu6 Sn5 [11] , 0.3 for (Cu, Ni)6 Sn5 [12] and 0.33 for Cu [13].
3. Results and discussion
Figure 1a shows the cross-sectional image of electrodeposited Cu/Sn multilayered thin film after 168 h of aging at 150 C. Regions with the darkest
3

Figure 1: (Color online) FESEM cross sectional images of electrodeposited Cu/Sn (a) and
Cu/Ni/Sn (b) multilayer thin film after 168 h of solid state aging at 150 C. Elemental
analysis on a line (c) of IMC layer in Cu/Ni/Sn system by EDX (d). FESEM images of
Cu/Sn and Cu/Ni/Sn after 365 days of preservation in air.

contrast at the bottom shows Cu substrate, while regions with lighter contrast shows two different IMC phases. The IMC phases were identified based
on the ratio of concentration of constituent elements measured by EDX. The
IMC phase with darker gray represent Cu3 Sn phase while the lighter gray
represents Cu6 Sn5 [14, 15]. In Fig. 1b, regions with brighter contrast represents (Cu, Ni)6 Sn5 IMC phase. The thin gray layers in between (Cu, Ni)6 Sn5
phase and Cu substrate is shown to be Cu3 Sn phase. With the addition of
Ni, the thickness of Cu3 Sn decreases considerably as the growth of Cu3 Sn
phase seemed to be suppressed with the presence of Ni. Average concentration of Ni was found to be about 5% in the (Cu, Ni)6 Sn5 IMC from the
EDX analysis (Fig. 1d) which was measured on the line shown in Fig 1c.
Tin whisker growth was not observed in both samples. One of the reasons
for tin whisker growth is IMC formation that alter the lattice spacing in
the tin plating. This change in lattice spacing could cause stress to the tin
plating which form tin whisker as a process to release stress. However, in
this case, probably all amount of tin was converted into the IMCs because
there is enough supply of Cu for reaction from substrate and no amount of
tin was left to grow tin whisker. Fig. 1e and 1f shows the FESEM images
of the Cu/Sn and Cu/Ni/Sn, respectively after 365 days of preservation in
4

Table 1: Elastic Moduli and hardness of Cu3 Sn, Cu6 Sn5 (Cu, Ni)6 Sn5 and Cu obtained
by nanoindentation.

Material
Cu3 Sn

Cu6 Sn5

(Cu, Ni)6 Sn5

Cu

Elastic moduli Hardness


(GPa)
(GPa)
97.25 7.34
5.99 0.39
132.17 3.63
6.34 0.14
134.2 6.7
6.12 0.17
94.30 2.36
6.61 0.25
112.3 5.0
6.38 0.21
97.0 3.0
5.7 0.5
110.80 5.05
7.43 0.57
160.04.9
7.3 1.2
157.82 5.69
7.24 0.44
99.66 2.02
1.55 0.044
116 4.7
140 9
1.7 0.2

Ref.
This
[11]
[16]
This
[16]
[12]
This
[12]
[17]
This
[16]
[18]

work

work

work

work

air. Thickness of Cu3 Sn phase was increased (Fig 1e) while almost no change
in the thickness of (Cu, Ni)6 Sn5 is noticed. Tin whisker growth is still not
visible in both samples.
Figure 2 shows the representative load-displacement curves of Cu3 Sn,
Cu6 Sn5 and (Cu, Ni)6 Sn5 under the same conditions. Indents were made at
higher penetration depth (> 100 nm) to avoid the influence of surface free
energy. Multiple pop-in events were observed for Cu6 Sn5 and (Cu,Ni)6 Sn5
whereas no such events were seen for Cu3 Sn. P. -F Yang et. al also reported
similarly on observation of pop-in events in Cu6 Sn5 and Cu3 Sn[11]. Phase
change of the IMCs or crack might be possible reason.
Figure 3 shows the typical scanning probe microscopic (SPM) image of 3D
morphology after indent on Cu3 Sn, Cu6 Sn5 , (Cu, Ni)6 Sn5 and Cu. Difference
in indentation depth of the indents on IMCs and Cu are clearly visible.
Indentation depth depends on the hardness of the materials. Cu as the
softest material in this study, shows the largest indentation depth.
Figure 4 shows the hardness and elastic moduli of Cu3 Sn, Cu6 Sn5 , (Cu,
Ni)6 Sn5 and Cu, plotted for each indent. Hardness value for Cu keeps almost
constant for 4 indents with very low deviation from average. Average hardness value obtained for Cu is 1.55 0.044 and elastic modulus was found
to be 99.66 2.02 (GPa). By averaging measurement data of indentations
5

Figure 2: (Color online) Representative load-displacement curves from nanoindentation


on the cross-section of IMC layers.

Figure 3: (Color online) Typical 3D SPM images of (a) Cu3 Sn and Cu6 Sn5 (b) (Cu,
Ni)6 Sn5 and Cu after nanoindentation.

Figure 4: (Color online) Hardness and elastic moduli of three different types of IMCs
Cu3 Sn, Cu6 Sn5 , (Cu, Ni)6 Sn5 and Cu measured for each nanoindent.

from various spot on the IMCs, elastic moduli of the IMCs were found to be
97.25 7.34 for Cu3 Sn, 94.30 2.36 for Cu6 Sn5 and 110.80 5.05 for (Cu,
Ni)6 Sn5 . Corresponding hardness were 5.99 0.39 for Cu3 Sn, 6.61 0.25
for Cu6 Sn5 and 7.43 0.57 for (Cu, Ni)6 Sn5 . The results are furnished in
Table 1 with a comparison of other values from previous reports. Hardness
and elastic moduli were found to be in good agreement with other reports
shown in the table.
4. Conclusion
The thickness of the Cu3 Sn IMC decreased when Ni is introduced into the
Cu/Sn system. Mechanical properties namely elastic moduli and hardness
of these IMCs were measured by nanoindentation.The hardness values were
found to be 5.99, 6.61 and 7.43 GPa for Cu3 Sn, Cu6 Sn5 and (Cu, Ni)6 Sn5
respectively. Elastic moduli were measured to be 97.25, 94.30 and 110.80 for
Cu3 Sn, Cu6 Sn5 and (Cu, Ni)6 Sn5 respectively. Tin whisker growth was not
observed in both samples after preservation in air for 365 days.

Acknowledgment
The authors would like to acknowledge the financial support from HIR
Grant, University of Malaya (Project no. UM.C/625/1/HIR /MOHE/ENG/26).
References
[1] Chuang TH. Scripta Mater 2006; 55; 983-6.
[2] Chuang TH, Lin HJ, Chi CC. Scripta Mater 2007; 56; 45-8.
[3] Chen CM, Chen YJ. Mater Lett 2009; 63; 1517-20.
[4] Sarobol P, Koppes JP, Chen WH, Su P, Blendell JE, Handwerker CA.
Mater Lett 2013; 99;76-80.
[5] Sun Y, Hoffman NE, Lam PS, Li X. Scripta Mater 2011; 65; 388-91.
[6] Crandall ER. Factors Governing Tin Whisker Growth. Springer; 2013.
[7] Chia PY, Haseeb ASMA. J Mater Sci: Mater Electron 2014; DOI
10.1007/s10854-014-2398-9.
[8] Fisher-Cripps AC. Nanoindentation, New York: Springer; 2011, p. 2137.
[9] Goh Y, Haseeb ASMA, Sabri MFM. Electrochim Acta 2013; 90; 265-273.
[10] Oliver WC, Pharr GM. J Mater Res 1992; 7; 1564-83.
[11] Yang PF, Lai Ys, Jian SR, Chen J, Chen RS. Mater Sci Eng A 2008;
485; 305-10.
[12] Che FX, John HLP. J Alloys Comp 2012; 541; 6-13.
[13] Purswani JM. Electron Scattering at Copper Surfaces, New York: ProQuest; 2008, p. 43.
[14] Li JF, Agyakwa PA, Johnson CM. Acta Mater 2011; 59; 1198-1211.
[15] Li GY, Bi XD, Chen Q, Shi XQ. J Electron Mater 2011; 40; 165-175.
[16] Deng X, Koopman M, Chawla N, Chawla KK. Mater Sci Eng A 2004;
240-3.
8

[17] Song JM, Liu YR, Su CW, Lai YS, Chiu YT. J Mater Res 2008; 23;
2545-54.
[18] Chromik RR, Vinci RP, Allen SL, Notis MR. J Mater Res 2003; 18;
2251-61.

Você também pode gostar