Effect of ENIG deposition on the failure mechanisms of thermomechanically loaded lead‐free 2nd level interconnections in LTCC/PWB assemblies

Abstract

Purpose -The purpose of this paper is to investigate the effect of electroless NiAu (ENIG) deposition on the failure mechanisms and characteristic lifetimes of three different non-collapsible lead-free 2nd level interconnections in low-temperature co-fired ceramic (LTCC)/printed wiring board (PWB) assemblies. Design/methodology/approach -Five LTCC module/PWB assemblies were fabricated and exposed to a temperature cycling test over a 2 40 to 1258C temperature range. The characteristic lifetimes of these assemblies were determined using direct current resistance measurements. The failure mechanisms of the test assemblies were verified using X-ray and scanning acoustic microscopy, optical microscopy with polarized light, scanning electron microscope (SEM)/energy dispersive spectroscopy and field emission-SEM investigation. Findings -A stable intermetallic compound (IMC) layer is formed between the Ni deposit and solder matrix during reflow soldering. The layer thickness does not grow excessively and the interface between the layer and solder is practically free from Kirkendall voids after the thermal cycling test (TCT) over a temperature range of 2 40 to 1258C. The adhesion between the IMC layer and solder matrix is sufficient to prevent separation of this interface, resulting in intergranular (creep) or mixed transgranular/intergranular (fatigue/creep) failure within the solder matrix. However, the thermal fatigue endurance of the lead-free solder has a major effect on the characteristic lifetime, not the deposit material of the solder land. Depending on the thickness of the LTCC substrate and the composition of the lead-free solder alloy, characteristic lifetimes of over 2,000 cycles are achieved in the TCT. Originality/value -The paper investigates in detail the advantages and disadvantages of ENIG deposition in LTCC/PWB assemblies with a large global thermal mismatch (DCTE $ 10 ppm/8C), considering the design and manufacturing stages of the solder joint configuration and its performance under harsh accelerated test conditions.

Enhanced thermal fatigue endurance and lifetime prediction of lead‐free LGA joints in LTCC modules

“…This was an expected result, since it is generally known that a (Cu,Ni) 6 Sn 5 layer forms typically between Sn‐based lead‐free solders and ENIG deposition after soldering at typical reflow peak temperatures, although a TEM investigation of the as‐soldered Sn1Ag0.5Cu and Sn4Ag0.5Cu joints proved that a very thin (≈ 20 nm) layer of (Ni,Cu) 3 Sn 4 existed between the Ni deposit and (Cu,Ni) 6 Sn 5 layer after reflow soldering (Suh et al , 2007). Moreover, some gold was occasionally detected within the (Cu,Ni) 6 (Sn,In) 5 layer after the reflow soldering and TCT in accordance with earlier studies (Islam et al , 2005; Sharif and Chan, 2006; Nousiainen et al , 2010b).…”

“…The typical crack path and the recrystallized microstructure of the test joints after the TCT over a temperature range of −40‐125°C are shown in Figure 2. The recrystallization behaviour and the intergranular (creep) cracking of the test joints were consistent with the thermomechanically loaded collapsible BGA joints with ternary SnAgCu alloys and the non‐collapsible solder joint configurations with the PCSB and SAC‐In solder (Hendersson et al , 2004; Mattila et al , 2004; Terashima et al , 2004a, b; Sundelin et al , 2008; Nousiainen et al , 2008a, 2009, 2010b). On the other hand, the recrystallized zone of the LGA joints with the Sn4Ag0.5Cu and Sn3Ag0.5Cu0.5In0.05Ni solders was localized next to the LTCC/solder interface in the more rigid assemblies with 1.2 mm thick LTCC modules and Arlon (Nousiainen et al , 2010a).…”

“…The as‐soldered microstructure of the Sn7In4.1Ag0.5Cu alloy in the pre‐tinned modules is shown in Figure 5(a). It consists of Cu 6 (Sn,In) 5 and Ag 3 (Sn,In) particles within the Sn matrix, in which a small amount (2‐4 at%) of indium exists as solute atoms, similarly to earlier studies with this alloy (Nousiainen et al , 2006, Nousiainen et al , 2009, 2010b). The only exception was the occurrence of small (<0.25 μ m) Au‐In‐Sn particles.…”

“…Considering that the joints spent more than 1,250 h at the high temperature extreme (125°C) of the TCT, a solid‐state phase transformation (Cu,Ni) 6 Sn 5 → (Ni,Cu) 3 Sn 4 seemed reasonable, since the occurrence of this phase transformation in ENIG‐plated lead‐free solder joints at elevated temperatures (≥ 150°C) was reported in the literature (Cheng et al , 2004; Kao and Duh, 2005; Sun et al , 2006). On the other hand, this phase transformation was not detected in the same ENIG‐plated LTCC modules with non‐collapsible BGA joints in the TCT over a temperature range of −40‐125°C (Nousiainen et al , 2010b). Since those joints were exposed to the high temperature for less than 650 h, it seemed that the growth kinetics of the IMC2 were relatively slow at a temperature of 125°C.…”

“…It reveals that the reason for this crack was a broken and missing piece of the glass solder mask on the Ag base metallization. In this area, solder was spread over the base metallization, resulting in ceramic cracking similar to primary ceramic cracking in non‐solder‐mask LTCC modules (Nousiainen et al , 2010b). This incident clearly demonstrated the importance of the solder mask in the solder joint configuration of LTCC modules.…”