-
- News
- Books
Featured Books
- smt007 Magazine
Latest Issues
Current IssueComing to Terms With AI
In this issue, we examine the profound effect artificial intelligence and machine learning are having on manufacturing and business processes. We follow technology, innovation, and money as automation becomes the new key indicator of growth in our industry.
Box Build
One trend is to add box build and final assembly to your product offering. In this issue, we explore the opportunities and risks of adding system assembly to your service portfolio.
IPC APEX EXPO 2024 Pre-show
This month’s issue devotes its pages to a comprehensive preview of the IPC APEX EXPO 2024 event. Whether your role is technical or business, if you're new-to-the-industry or seasoned veteran, you'll find value throughout this program.
- Articles
- Columns
Search Console
- Links
- Events
||| MENU - smt007 Magazine
Effect of Cooling Rate on the Intermetallic Layer in Solder Joints
December 4, 2012 |Estimated reading time: 11 minutes
Editor's Note: This paper was originally published in the proceedingsof IPC APEX EXPO 2012 and the November 2012 issue of SMT Magazine.
Abstract While it has long been known that the Cu6Sn5 intermetallic that plays a critical role in the reliability of solder joints made with tin-containing alloys on copper substrates exists in two different crystal forms over the temperature range to which electronics circuitry is exposed during assembly and service, it has only recently been recognized that the change from one form to the other has implications for solder joint reliability. Under equilibrium conditions the change from the hexagonal to monoclinic form occurs in the cooling solder joint at 186°C. However, cooling rates after common commercial soldering processes are typically faster than the rate that would permit complete transformation under such equilibrium conditions. In this paper the authors report a study of the effect of cooling rates on Cu6Sn5 crystals. Cooling rates from 200°C ranged from 10°C/minute to 100°C/minute and the effect of isothermal ageing at intermediate temperatures was also studied. The extent of the phase transformation after each regime was determined using synchrotron X-ray diffraction. The findings have important implications for the manufacture of solder joints and their in-service performance.
Introduction
It is becoming increasingly apparent that the reliability of lead-free solder joints is not as consistent and predictable as that of the tin-lead solders they are replacing. That means that so far it has not been possible to establish for lead-free solders the acceleration factors in which designers can have as much confidence as they have in those developed for tin-lead solder. There are many reasons why the behaviour of lead-free solders is less predictable. Lead has a number of effects that served to moderate the behaviour of the tin. Tin is reactive towards most common substrates tending to form intermetallic compounds which have properties very different from the metals from which they are formed. Lead has no such tendency and appears to reduce the reactivity of tin by dilution. By dissolving in the tin to the extent of up to about 2.5% lead has the effect of increasing both strength and ductility and the physical structure of the tin-lead eutectic with layers of a lead-rich phase interleaved with layers of the tin-rich phase also seems to contribute to the general stability and consistency of its mechanical properties. By contrast neither of the additions to tin used in the most common lead-free solders, copper and silver, has any significant solubility in tin and instead both form intermetallic compounds with it. Intermetallic compounds tend to have low ductility and can so provide a low energy path for crack propagation. The role of intermetallics in lead-free solders has been further complicated by the realisation that one of the most common, Cu6Sn5, undergoes at least one polymorphic transformation over the temperature range to which it is exposed in normal assembly and service.
A phase change can have various consequences for the solder joint. The primary effect is the change in properties of the intermetallic as the crystal structure changes, but it has been proposed [1] that the stresses generated by the change in volume that accompanies the transformation could result in the in situ fracture of the intermetallic. Such fracture has been reported with the observation that it does not occur at all or to the same extent when Ni has substituted for some of the Cu in the Cu6Sn5 (Figure 1) [2]. It is expected that such cracking would have negative consequences for the reliability of the joint, particularly under conditions of impact loading and some evidence for that effect has been reported [3].
In this paper, ongoing studies of the stability and thermal expansion of the Cu6Sn5 intermetallic are reported, including the reproduction of some significant recently published results [4, 5]. Thermal expansion of Cu6Sn5, with and without partial substitution of the copper by nickel, was measured by dilatometry and X-ray diffraction. The conditions under which such volume changes might occur in Sn-Cu solder joints has been determined by monitoring by X-ray diffraction the crystal structure of Cu6Sn5 during a sequence of rapid cooling and isothermal ageing.
Figure 1: Cracking in the Cu6Sn5 at the interface between copper and an Sn-0.7Cu solder sphere. Cracking does not occur with an Sn-0.7Cu-0.05Ni solder sphere when Ni partially substitutes for Cu in the intermetallic [2].
Measuring Thermal Expansion
DilatometryCu6Sn5 and (Cu,Ni)6Sn5 powders were extracted from Sn-4Cu and Sn-4Cu-0.05Ni ingots by dissolution of the tin matrix in an ortho-nitrophenol/sodium hydroxide solution and crushed in agate mortar. The powder was pressed into a rod shape in a 10-mm diameter die and linear thermal expansion measured with a NETZSCH 402-C dilatometer. The sample was taken through three cycles from 25°C to 250°C at a rate of 1°C/minute with a 30-minute dwell at each end of each cycle. During the first cycle the powder is compacted so only measurements from the second and third cycle were analysed using the NETZSCH Proteus software. Hexagonal (101) pole figures of the Cu6Sn5 and (Cu, Ni)6Sn5 dilatometry sample rods indicated no particular orientation so that the data collected is taken as representing an average result for what are anisotropic crystals.
Results
Dilation percent and the rate of dilation during heating are plotted in Figure 2. The overlapping of the plots from the second and third cycles indicates a high level of reproducibility. The coincidence of the plots for the stabilized and unstabilized IMC at lower temperature indicates that the average coefficients of thermal expansion for both phases are similar. However, as the 186°C transformation temperature is approached the rate of dilation of the unstabilized IMC accelerates.
Figure 2: Normalized dilation and dilation rate for unstabilized and Ni-stabilized Cu6Sn5 during heating [4].
X-ray Diffraction
XRD measurements were made on the powder diffraction beam line of the Australian Synchrotron. Powder was loaded into a 0.3-mm diameter quartz capillary cell and on the stage heated with hot air at a rate of 6°C/minute to 30, 170, 200, 220, and 250°C. On reaching the desired temperature the sample was held for 1 minute for thermal stabilization and measurements taken for 10 minutes at 2θ scan angles of 10 to 80°C. A wave length of 0.0825 nm was used with Cu6Sn5 and 0.0773 nm with (Cu,Ni)6Sn5. Lattice parameters and unit cell volumes were calculated using REITAN-FP analysis software [6] with the refinement parameters optimized to minimize the residual Rwp, Rp and S factors. As a reference crystallography and atomic coordination International Centre for Diffraction Data 045-1488 was used for monoclinic and 047-1575 for hexagonal in association with the REITAN-FP.
Results
The disappearance of the minor peaks associated with the monoclinic phase in the unstabilized Cu6Sn5 between the X-ray diffraction peak profiles taken at 170 and 200°C (Figure 3a) can be taken as confirmation that somewhere between these two temperature the monoclinic η’ form undergoes polymorphic transformation to the hexagonal η form. The X-ray diffraction peak profile indicates that the (Cu,Ni)6Sn5 retains an hexagonal form throughout that range.
While the incorporation of Ni into the Cu6Sn5 structure appears to have no significant effect on the c axis lattice parameter it reduces the a axis parameter (Figure 4) and consequently the cell volume (Figure 5).
Transformation Kinetics Studies
Samples similar to those used in the thermal expansion study were subjected to two types of thermal treatment while on the powder beamline of the Australian Synchrotron.
Cooling
Measurements at 200°C for 5 minutes were followed by cooling to 50°C at rates of 100, 50, 30, 20, and 10°C/min-1 and then 5 minutes of measurements at 50°C (Figure 6).
Figure 3: X-ray diffraction peak profile for IMC samples heated from 30 to 250°C [4].
Figure 4: a and c axis lattice parameters in hexagonal Cu6Sn5 and (Cu,Ni)6Sn5 as a function of temperature [4].
Cooling Plus Isothermal Ageing
Measurements at 200°C for 5 minutes followed by cooling at a rate of 100°C/min-1 to 180, 160, 150, 140, and 100°C followed by 5 minutes of data collection at each of these temperatures for up to 12 measurements (Figure 7). Measurements of the unstabilized Cu6Sn5 were continued until complete conversion to the monoclinic form had occurred and measurements over a similar time frame were made for the stabilized (Cu,Ni)6Sn5.
Figure 5: Unit cell volumes as a function of temperature [4].
Figure 6: Cooling profiles to which samples of Cu6Sn5 and (Cu,Ni)6Sn5 were subjected [5].
Figure 7: Cooling and isothermal ageing profiles to which samples of Cu6Sn5 and (Cu,Ni)6Sn5 were subjected [5].
Figure 8. X-ray diffraction peak profiles for Cu6Sn5 samples from isothermal ageing experiment [5].
Results
From the peak profiles in Figure 8 it can be seen that for the samples of the unstabilized Cu6Sn5 cooled at 100°C/min-1 and aged at temperatures of 160, 150, and 140°C transformation to the monoclinic form (revealed by the appearance of minor peaks) begins almost immediately. At 180°C the hexagonal form is retained and in the samples cooled to 100°C a metastable hexagonal phase is retained. Figure 9 shows the summarised results of the XRD measurements for both the continuously cooled and isothermally held samples.
Discussion
Taking into account the results presented in Figures 8 and 9 a TTT diagram has been constructed as illustrated in Figure 10. This provides a means of predicting the phases likely to be present as a function of the thermal profile to which the solder is subjected. The times at which the transformation is completed are yet to be finally determined (dashed line).
Figure 9: Phase constitution of Cu6Sn5 after cooling rate and isothermal ageing [5].
Figure 10: Estimated boundaries of temperature-time-transformation (TTT) curves for Cu6Sn5[5].
Figure 11: Hypothetical plot illustrating a possible technique for minimizing the stress generated by the η→η’ transformation.
Based on the dilatometry measurements and the study of the polymorphic transformation under different thermal profiles a concept for risk minimization is being proposed that is illustrated schematically in Figure 11. On the basis that a volume change in either direction due to the polymorphic transformation could generate stress and trigger joint cracking, it is proposed that the cooling be managed in alloys which do not have the benefit of a Cu6Sn5 phase that has been stabilized by Ni to provide an opportunity for the transformation to occur at a temperature which the change in unit cell volume is at a minimum.
Conclusions
- In dilatometry studies the transformation from monoclinic η’ to the hexagonal η phase during heating can be observed as an expansion over a temperature range around the nominal 186°C transformation temperature.
- Cu6Sn5 with partial substitution of Cu by Ni retains its hexagonal crystal form over the range room temperature to 250°C.
- The partial substitution of Cu by Ni in the hexagonal Cu6Sn5 crystal reduces the lattice parameter in the a direction, but does not significantly affect the lattice parameter in the c direction.
- At cooling rates as slow at 50°C/min-1 metastable hexagonal η phase can be retained to 50°C.
- In isothermal interruptions to 100°C/min-1 cooling at 160, 150 and 140°C transformation from the metastable hexagonal phase to the monoclinic phase has begun by 400 seconds. At 180°C the transformation did not begin within an hour. At 100°C the metastable hexagonal η phase remained unchanged for up to 1,000 seconds.
- On the basis of the results of rapid cooling and isothermal ageing experiments it is possible to start to map a temperature-time-transformation (TTT) diagram for the η→η’ transformation in Cu6Sn5.
- On the basis of a presumption that the different coefficients of thermal expansion of the hexagonal η phase and the monoclinic η’ phase can be extrapolated into the temperature ranges where the phases are metastable it is hypothesized that whether the change from metastable η’ to η is associated with an expansion or contraction depends on the temperature at which it occurs. That creates an opportunity for minimizing potential damage to the joint integrity by allowing the transformation to occur at a temperature at which the difference in coefficient of thermal expansion is close to zero.
Acknowledgements
The work reported in this paper was conducted under an international cooperative research program between the University of Queensland, Australia, and Nihon Superior Company Limited, Japan.
References:1. T. Laurila, V. Vuorinen, and M. Paulasto-Kröckel, Materials Science and Engineering R 68, 1-2, 1-38, 2010.2. K. Nogita, Intermetallics 18, 145-149, 2010.3. H. Tsukamoto, T. Nishimura, S. Suenaga, S. D. McDonald, K. W. Sweatman, and K. Nogita, Microelectronics Reliability, 2011.4. D. Mu, J. Read, Y.-F. Yang, and K. Nogita, Journal of Materials Research 26, 20, 2660-2664, 2011.5. K. Nogita, C. M. Gourlay, S. D. McDonald, Y. Q. Wu, J. Read, and Q. F. Gu, Scripta Materialia 65, 922-925, 2011.6. F. Izumi, K. Momma, Solid State Phenomena 130, 15-20, 2007. Keith Sweatman is a graduate in metallurgical engineering with more than 30 years’ experience soldering technology, first with the International Tin Research Institute, then with Multicore Solders Limited and since 2001 with Nihon Superior Co. Ltd. as senior technical advisor. Over the past decade he has been particularly in involved in the implementation of lead-free technologies in electronics manufacturing
Tetsuro Nishimura is a graduate in metallurgy and is recognized as the pioneer in the development of micro-alloyed solders, holding patents on the use of trace additions of nickel that have made it possible for the tin-copper eutectic to be used for the production of reliable solder joints in mass production. This invention has earned recognition and respect around the world as the lead-free solder SN100C. Under his management Nihon Superior has expanded to its current position as a major supplier of lead-free solder to the global market.
Kazuhiro Nogita is a graduate in nuclear engineering and earned his Ph.D. for research on nuclear materials at Kyushu University. Since 1999 he has been engaged in research at University of Queensland, Australia, where he is currently associate professor and principal research fellow in materials engineering. He is also an associate professor at Osaka University where he is particularly involved in the application of the SPring8 synchrotron to materials research.
Stuart McDonald is a graduate in manufacturing and materials engineering and was awarded his Ph.D. for materials research at University of Queensland where he is currently a research fellow in materials engineering. In recent years his analysis of solidification microstructures has extended to lead-free solders, particularly those based on the tin-copper eutectic and the formation and growth of the intermetallic layer that develops at the solder/substrate interface.