Rapid growth of tin whiskers on the surface of Sn–6.6Lu alloy

(1)

Rapid growth of tin whiskers on the surface of Sn–6.6Lu alloy

T.H. Chuang,

*

_{H.J. Lin and C.C. Chi}

Institute of Materials Science and Engineering, National Taiwan University, Taipei 106, Taiwan

Received 6 June 2006; revised 18 August 2006; accepted 26 August 2006 Available online 5 October 2006

During the storage of a Sn–6.6Lu alloy in air for several days, large amounts of thread-like tin whiskers appear on the oxidized surface of Lu4Sn5precipitates in this alloy. Storage at 150C for 30 min causes hillock-type whiskers to form. The driving force for

whisker growth in this Sn–6.6Lu alloy is the compressive stress resulting from the diﬀusion of oxygen into the lattice of the Lu4Sn5

precipitates.

Keywords: Sn–6.6Lu; Rare-earth element; Tin whiskers; Hillocks

The formation of tin whiskers on electrodeposited Sn and Sn-alloys has been acknowledged for over 50 years[1]. Tin whiskers were found to be single crystals with a typical length of a few hundred micrometers which sprout from the Sn matrix and can get kinked during the growth process. Typical growth rates of tin whiskers in a bulk material are about 0.01–0.1 A˚ s1

[2]. However, some cases with more rapid growth rates have been reported. Tin whiskers were observed by Fur-uta and Hamaura to grow in a rapidly cooled Al–50%Sn ﬁlm at rates of about 0.5–5 A˚ s1 [3]. Higher whisker growth rates of about 2–4 A˚ s1 were found in electro-plated Sn–Mn ﬁlms[4]. Liu et al. reported that tin whis-ker growth can be accelerated to 3 A˚ s1 with an electrical current density of 1.5· 105A cm2 [5]. They also found that the whisker growth rates increase with rising temperature.

Several models for the growth mechanism of tin whis-kers, including dislocation theory [6] and recrystalliza-tion theory [3], have been proposed. However, other researchers have reported that tin whisker growth is re-lated to the outer oxide layer on the Sn surface and the internal compressive stress in the Sn matrix. Tu studied the tin whiskers on a bimetallic Cu–Sn thin-ﬁlm speci-men and proposed the existence of a biaxial compressive stress produced in Sn ﬁlm accompanying Cu6Sn5

forma-tion, which drives the extrusion of the whiskers from the outer oxide layer[7]. Sheng et al. further suggested that whiskers sprout from weaker spots of the oxide layer on

the Sn surface and that the roots of whiskers become localized stress relief centers[8].

It is known that rare-earth elements exhibit high chemical activity. Sn alloys containing rare-earth ele-ments should readily form an oxide layer that can give rise to the growth of tin whiskers. This study presents the eﬀect of rare-earth elements on tin whisker growth in a Sn–6.6Lu solder alloy.

For the preparation of the Sn–6.6Lu alloy, pure Sn (99.9%) and pure Lu (99.9%) were melted at 1000C under a 105Torr vacuum. The as-cast specimens were cut with a diamond saw, and their cross sections were ground with 2000 grit SiC paper and polished with 0.3 lm Al2O3 powder. Some specimens were stored at

room temperature in air, while others were aged at 150C in an air furnace. After various storage periods, the morphology of the tin whiskers that formed on the surface was observed by scanning electron microscopy (SEM). The chemical composition of the specimens was analyzed using an electron probe microanalyzer (EPMA).

The microstructure of the as-cast Sn–6.6Lu alloy contains many large precipitates, as shown inFigure 1(a). EPMA analysis indicates that the composition (at.%) of these precipitates is Lu:Sn = 44.4:55.6, which corre-sponds to the Lu4Sn5 phase. Figure 1(b) reveals that

after storage at room temperature in air for 4 h, the sur-face of the Lu4Sn5is covered with many bright particles,

and its chemical composition (at.%) changes to Lu: Sn:O = 37.2:29.3:33.5. This result indicates that the Lu4Sn5 precipitates have oxidized much more rapidly

than the matrix of the Sn–6.6Lu alloy. The bright

* Corresponding author. E-mail:[email protected]

Scripta Materialia 56 (2007) 45–48

(2)

particles are composed of nearly pure Sn, which implies that they are the sprouts of tin whiskers. With a further increase of the storage time to over 2 days, many thread-like whiskers appear in the oxidized Lu4Sn5

precipita-tion region, as shown in Figure 2(a). Some whiskers grow very quickly to a length of over 45 lm with an in-crease of storage time, as can be seen in Figure 2(b). The maximal growth rate of tin whiskers for the case of room-temperature storage in air is about 1 A˚ s1.

When the storage temperature is raised to 150C, tin whiskers form much earlier, as compared to the samples stored at room temperature. Figure 3(a) shows that some whiskers grew to a length of about 30 lm in the sample that was stored at 150C for only 10 min. The whisker growth rate in this case reaches about 500 A˚ s1. Such rapid and early growth of tin whiskers has never been reported previously in the literature. When the storage time is increased, a few hillock-type whiskers coexist with the thread-like whiskers in the oxidized Lu4Sn5region, as shown inFigure 3(b). A long

period of storage at 150C in air causes further growth of both the thread-like and hillock-type whiskers in the oxidized Lu4Sn5region (Fig. 4). It can be seen inFigure

4(a) that more than 500 lm3of tin has been extruded out of the Lu4Sn5 precipitate in the micrograph after

storage at 150C for 112 h through the growth process of hillock-type whiskers. In contrast, whisker growth was obviously reduced when the specimens were stored at room temperature for 3 days and at 150C for 2 h in a vacuum furnace of 103Torr as illustrated inFigure 5(a) and (b), respectively. The results imply that the for-mation of tin whiskers in this Sn–6.6Lu alloy is closely related to the oxidation of Lu4Sn5precipitates.

In order to clarify the mechanism for whisker growth in this rare-earth–Sn alloy, specimens were cut across the Lu4Sn5 precipitates. The three-dimensional cross

section in Figure 6(a) reveals a continuous oxide layer with a thickness of about 2 lm on the outer surface of a Lu4Sn5 precipitate in the Sn–6.6Lu stored at room

temperature for 4 h. The diﬀusion rate of oxygen, as estimated from the thickness of the oxide layer, is about 1.4 A˚ s1. The chemical composition (at.%) of the outer oxide layer is Lu:Sn:O = 30.9:35.1:34.0, which is similar to that analyzed for the surfaces of Lu4Sn5 phase in

Figure 1(b), However, Figure 6(b) shows that in the specimen stored at 150C for 90 min, the oxygen

Figure 1. Morphology of Lu4Sn5 precipitates with tin particles

distributed in Sn–6.6Lu alloy after air storage at room temperature for short periods: (a) 10 min, (b) 240 min.

Figure 2. Thread-like tin whiskers formed on the surface of Lu4Sn5

precipitates in Sn–6.6Lu alloy after air storage at room temperature for long periods: (a) 48 h, (b) 240 h.

Figure 3. Hillock-type tin whiskers formed on the surface of Lu4Sn5

precipitates in Sn–6.6Lu alloy after air storage at 150C for short periods: (a) 10 min, (b) 90 min.

(3)

penetrated through the Lu4Sn5/Sn–6.6Lu interface to

form a thick oxide layer of about 5 lm thickness on the outer surface and around the Lu4Sn5 phase in the

interior of the specimen, which corresponds to a diﬀu-sion rate of about 10 A˚ s1 for the oxygen into the Lu4Sn5phase. EPMA analysis indicates that the

envel-oped oxide layer possesses a composition (at.%) of Lu: Sn:O = 51.3:5.9:42.8. The results imply that the Lu4Sn5

precipitates react predominantly with oxygen to form an

LuO layer due to the high activity of the Lu element: 2Lu4Sn5+O2! 8LuO + 10Sn . The diﬀusion of oxygen

into the crystal lattice of the Lu4Sn5 phase leads to

lattice expansion, which is constrained by the surround-ing matrix. A compressive stress can be created, which extrudes the resulting Sn atoms after oxidation out of the surface of the Lu4Sn5phase. During storage at room

temperature, Sn atoms in smaller quantities are extruded from the weak spots of the LuO layer to form thread-like whiskers that are uniformly distributed over the whole surface of the Lu4Sn5phase. In this case, a higher

Sn concentration of above 30 at.% was detected in the outer oxide layer. In contrast, large amounts of Sn atoms in the LuO around the Lu4Sn5phase have been

extruded to form hillock-type whiskers, and the Sn con-centration in the enveloped LuO layer drops to a minor value of 5.9 at.%.

In conclusion, after storage at room temperature in air for several days, thread-like whiskers appear on the surface of Lu4Sn5precipitates in Sn–6.6Lu solder alloy.

The maximal growth rate of tin whiskers in this case is about 1 A˚ s1. In contrast, no whiskers can be found in the Sn–6.6Lu matrix. During air storage at 150C, the thread-like whiskers grow to a length of about 30 lm in 10 min, which corresponds to an amazingly high growth rate of 500 A˚ s1. After 30 min, hillock-type whiskers coexist with the thread-like whiskers in the Lu4Sn5region of Sn–6.6Lu alloy. The rapid growth

of tin whiskers in this rare-earth-element-containing alloy is attributed to the predominant oxidation of Lu atoms, which possess high chemical activity. The oxida-tion reacoxida-tion results in the release of Sn atoms, which are inserted in the LuO layer. The diﬀusion of oxygen into the Lu4Sn5 phase leads to a compressive stress,

which extrudes the resulting tin atoms out of the LuO layer.

Figure 4. Hillock-type tin whiskers formed on the surface of Lu4Sn5

precipitates in Sn–6.6Lu alloy after air storage at 150C for long periods: (a) 112 h, (b) 224 h.

Figure 5. Morphology of Lu4Sn5precipitates in Sn–6.6Lu alloy after

storage in a vacuum furnace of 103_{Torr: (a) at room temperature for}

3 days, (b) at 150C for 120 min.

Figure 6. Three-dimensional cross section of Lu4Sn5 precipitates in

Sn–6.6Lu alloy after storage in air: (a) at room temperature for 4 h, (b) at 150C for 90 min.

(4)

This work was sponsored by the National Science Council, Taiwan, under Grant No. NSC-94-2216-E002-015.

[1] C. Herring, J.K. Galt, Phys. Rev. 85 (6) (1952) 1060. [2] W.C. Ellis, D.F. Gibbons, R.C. Treuting, in: R.H.

Dore-mus, B.W. Roberts, D. Turnbull (Eds.), Growth and Perfection of Crystals, John Wiley, New York, NY, 1958.

[3] N. Furuta, K. Hamamura, Jpn. J. Appl. Phys. 8 (12) (1969) 1404.

[4] K. Chen, G.D. Wilcox, Phys. Rev. Lett. 94 (2005) 066104. [5] S.H. Liu, C. Chen, P.C. Liu, T. Chou, J. Appl. Phys. 95

(12) (2004) 7742.

[6] G.T. Galyon, IEEE Trans. Electron. Package. Manuf. 28 (1) (2005) 94.

[7] K.N. Tu, Acta Metall. 21 (4) (1973) 347.

[8] G.T.T. Sheng, C.F. Hu, W.J. Choi, K.N. Tu, Y.Y. Bong, L. Nguyen, J. Appl. Phys. 92 (2002) 64.