Characterization of the Growth of Intermetallic Phases by DSC - METTLER TOLEDO

Characterization of the Growth of Intermetallic Phases by DSC

The method presented in this article shows how DSC measurements of solders in copper crucibles can be used to analyze the formation of intermetallic layers between solder and copper in soldering and annealing processes. After calibration using polished microsections, the average thickness of layers can be determined from the DSC measurements.

 Polished microsection of an annealed solder join

 

Introduction

In microelectronics, soldering is the most frequently used bonding technology employed to connect the leads of electronic components to tracks on the printed circuit board. The soldering process itself has been known for more than 5000 years [1].

In this process, a low-melting alloy (the solder) connects one or more components together through melting and solidification processes. The formation of an intermetallic compound (a so-called IMC) between the solder and the component is the clearest sign that a connection has been made during the soldering process [2].

Essential for IMC formation are diffusion processes that take place between the reactants participating in the formation of the alloy. In this study, we investigated a solder alloy based on tin in contact with copper.

In the soldering process, an intermetallic phase is formed consisting of two zones. The zone next to the copper layer is the copper-rich Cu3Sn phase and that next to the solder side is the tin-rich Cu6Sn5 phase. The formation of these two zones can be seen in the polished metallographic microsection shown in Figure 1.

The diffusion processes necessary for the formation of the IMC are temperature dependent. They have a low activation energy, ΔE, and can therefore take place both in the melt and in the solid phase after solidification.

An increase in the thickness of the IMC is therefore observed on soldering.

Knowledge of the thickness of the IMC is necessary in order to make predictions about the reliability of solder connections. Since the intermetallic phase is usually more brittle than the alloys involved, the IMC must not become too thick because otherwise the attachment point becomes susceptible to mechanical stresses [3].

The temperature and time dependence of the layer thickness of intermetallic phases, y, is described by equation 1


where y0 is the initial thickness, A the growth constant, t the time, n the corresponding exponent, T the temperature (in Kelvin) and R the universal gas constant. The parameters for this equation can be determined by rather time-consuming experimental methods [4].

For example, the determination the activation energy and the growth constant A involves long periods of annealing in a thermostated oven at different temperatures. After this, a large number of metallographic polished microsections have to be prepared to measure the thickness.

The actual measurement using a microscope is also laborious because the microsections are often irregular. Several measurements have to be made to determine a mean thickness.

A fundamental problem associated with this method is that the information obtained is only from one cutting plane and not from the entire volume.

In this article, we describe a new method in which the thickness of the intermetallic phases can be indirectly measured by means of DSC without the time-consuming optical measurement of microsections.

The basic idea of the method is that the temperature of melting of the intermetallic phase containing copper is higher than that of the solder alloy (see Figure 2).

Due to this, the enthalpy of melting of the remaining solder decreases with the growth of the intermetallic phase. In the temperature range up to 270 °C, the DSC curve shows the melting of the solder but not the melting of the IMC. A comparison of the enthalpies of melting and solidification after a certain storage time in the melt provides information about the proportion of the IMC formed during the storage period.

The method allows the growth of the IMC to be measured both in the liquid and in the solid solder. In the second case, the sample is first melted and then cooled to an annealing temperature below the solidification temperature.

As shown in Figure 3, after annealing in the DSC, the sample is again melted and solidified. The growth of the IMC can be deduced from the enthalpy difference between the first melting peak and the second solidification peak.

 

Experimental Details

The material investigated was a commercially available lead-free solder paste (Heraeus Type F680 SA305C) with an alloy composition of 96.5% Sn, 3.0% Ag, and 0.5% Cu. The melting range of the solder was 217 to 219 °C.

The measurements were performed using a METTLER TOLEDO DSC 1. Samples of about 5 mg were measured in copper crucibles with a lid.

The formation of the IMC was investigated using the complex temperature program shown in Figure 3:

  •  The first five segments of the measurement program represent a soldering process. To simulate this process, the sample was first heated from room temperature to 270 °C at 20 K/min, held at this temperature for 2 minutes and then cooled at 20 K/min.
  • To investigate the annealing process, the sample was then annealed for different times (up to 100 hours) at 200 °C. The melting curve and following cooling curve were then measured. The enthalpies of the first melting process and the second solidification process were evaluated and their difference, ΔH, determined.


For comparison, after measuring the DSC samples, microsections were prepared to measure the thickness of the IMC under the microscope.

This was done not only by multiple determination of the phase thickness as usually described in the literature [6] but also by determining the area of the phase in the microsection using imaging software and dividing it by the width of the image. This allowed the IMC thickness to be determined with much better accuracy [7]. 

 

Results and Discussion

From equation 1, it follows that with isothermal annealing the thickness grows with the nth power of time. We showed that the thicknesses determined from the microsections grow at a rate proportional to the square root of the annealing time. The exponent in equation 1 is thus given by n = 0.5.


Conclusions


The growth processes of intermetallic Sn-Cu phases can be analyzed using DSC measurements. To do this, ΔH, the difference between the enthalpy of melting before annealing and enthalpy of solidification after annealing are determined.

This enthalpy difference is proportional to the thickness of the intermetallic layer. For the system investigated, the proportionality factor a is 19.48 mJ/µm. To determine the mean thickness of the IMC, ΔH need only be divided by a.

The preparation of polished metallographic microsections is no longer necessary.

 

Characterization of the Growth of Intermetallic Phases by DSC | Thermal Analysis Application No. UC 402 | Application published in METTLER TOLEDO Thermal Analysis UserCom 40