A microindentation technique for determining strength of solder interface with silver metallization on co-fired multilayer ceramic substrate

Journal of Electronic Materials, Mar 2001 by Shang, Jian Ku, Huang, Rong-Fong, Wilcox, David L Sr

The use of microindentation to determine the strength of a solder interface was explored in low-temperature co-fired multilayer ceramic packages. Microindentations were made on cross-sections of multilayered structures at small stand-off distances from the interface under different indentation loads. Cracking behavior of the interface was observed following indentations. At low indentation loads and large stand-off distances, well-known indentation crack systems developed in the ceramic. Higher indentation loads and smaller standoff distances resulted in interfacial cracking. It is demonstrated that the indentation condition at the onset of the interfacial cracking may be used as a measure of the interfacial strength.

Key words: Microindentation, solder, interface, strength

INTRODUCTION

In the last few years, a series of new packaging technologies have emerged in the microelectronics industry by utilizing small volumes of solder alloys to mount individual devices or integrated circuits directly on the substrate or the printed circuit board of an electronic package. Notable examples include flipchip and ball-grid array technologies where the interconnection between the device and the board is made through a number of solder bumps or balls.1,2 Because of the low strength, low melting temperature and high fatigue susceptibility of the solder alloys, reliability of the solder interconnection remains a major concern. The problem is particularly serious in cases where solder alloys react aggressively with the device metallization to drastically weaken the interface between the solder alloy and the substrate, such as in the case of silver-metallized co-fired multilayer ceramic substrate.3

The reliability of the solder interface has been studied by a number of experimental techniques. Simple pull tests and lap shear tests are very popular but tend to be primarily qualitative because of the gross uncertainties associated with singular stresses or stress concentrations. Techniques based on fracture mechanics, such as the flexural peel and modified compact tension and double cantilever beam specimens, have been largely successful in providing quantitative measurements of the interfacial properties.4-7 However, these techniques often require large specimens in which a well-defined crack must be introduced. In newly developed packages based on solder bumps or balls of very small volumes, the solidification condition of the solder alloy can be quite different from that in large bulk specimens as required by the fracture mechanics techniques. As a result, it will be difficult to relate measurements made on large-size fracture mechanics specimens to the properties of the solder interface in real microsized devices. To probe the properties of solder interface on the microscale, a technique based on microsized specimens is needed.

In this paper, exploratory research investigating the use of microindentation as a tool to probe mechanical properties of solder interfaces on the microscale is described. Microindentations placed in the ceramic near the interface were found to induce interfacial cracking. The interfacial crack behavior is shown to correlate well with the indentation stress field. The onset of interfacial cracking is used to determine the strength of the solder interface.

BASIC CONCEPT

When a defect such as a crack is present on the surface, a ring-like crack pattern may develop if the magnitude of the radial stress exceeds the fracture stress associated with the flaw or the strength of the brittle solid.11-15

The basic hypothesis of this work is that if an interface is created in the vicinity of the concentrated load, as shown in Fig. 1, fracture of the interface may be observed when the magnitude ofthe stress reaches a critical value, i.e., the strength of the interface. For a given interface, such a postulation can be tested by systematically varying the external stress imposed on the interface. On a first-order approximation, this may be accomplished by varying the applied load, P, or the stand-off distance, S, between the interface and the loading point or both, according to Eq. 1. The strength of the interface, taken to be the external stress on the loading side of the interface, may be quantified by measuring the critical load at the onset of interfacial cracking, P^sub c^, for a fixed stand-off distance, or the critical stand-off distance, St, at a constant load.

EXPERIMENTAL PROCEDURE

Specimens used in this study were square arrays of solder connections formed between two metallized substrates, as shown in Fig. 2. The substrate was a glass ceramic, supplied as green sheets. The metallization was a thick-film silver paste. The solder was the tin-silver eutectic alloy containing 96.5 wt.%Sn and 3.5 wt.%Ag, obtained in the paste form.

Ceramic green sheets were metallized by screenprinting a layer of Ag paste selectively at the area where solder connections are to be formed. After drying to evaporate the solvent, the metallized tape was stacked together with a few layers of the green ceramic tapes and laminated under pressure. The laminate was then heated slowly to about 850 deg C to cofire the glass ceramic and the silver paste. The cofiring produced a dense ceramic plate covered with square arrays of silver pads. The thickness of the silver metallization after firing was about 10 Rm.