Fig 1 - uploaded by Hongbo Qin
Content may be subject to copyright.
View of an as-reflowed BGA joint sample and loading mode of the joint under cyclic loading: (a) cross-sectional view of an as-reflowed BGA joint and its geometrical parameters (h = 300 lm and d = 480 lm); (b) schematic of an asreflowed single BGA joint under cyclic shear loading; and (c) the symmetrical trapezoidal wave of cyclic loading with displacement control.
Source publication
Low cycle fatigue performance of ball grid array (BGA) structure Cu/Sn-3.0Ag-0.5Cu/Cu joints with different standoff heights (h, varying from 100 to 500 μm) and two pad diameters (d, d = 320 and 480 μm) under displacement-controlled cyclic loading was studied by experimental method and finite element (FE) simulation. A prediction method based on th...
Contexts in source publication
Context 1
... (simplified as Cu/SAC305/Cu) joint samples by a reflow process with a typical ramp-soak-spike (RSS) curve offered by a BGA rework machine (RD-300). The peak temperature was 250 °C and the dwell time at 250 °C was 30 s. The measured heating rate from 120 to 250 °C was 1.5 °C/s and the cooling rate from 250 °C to 150 °C was about 2.3 °C/s [8]. Fig. 1 shows an as-reflowed sin- gle BGA structure solder joint and the loading mode of the joint under cyclic loading. Solder joint samples were prepared with three different standoff heights (h) of 100, 300 and 500 lm, and two pad diameters (d) of 320 and 480 lm, as shown in Fig. 1(a). After reflow soldering, the BGA structure solder joints ...
Context 2
... and the cooling rate from 250 °C to 150 °C was about 2.3 °C/s [8]. Fig. 1 shows an as-reflowed sin- gle BGA structure solder joint and the loading mode of the joint under cyclic loading. Solder joint samples were prepared with three different standoff heights (h) of 100, 300 and 500 lm, and two pad diameters (d) of 320 and 480 lm, as shown in Fig. 1(a). After reflow soldering, the BGA structure solder joints were cleaned in an ultrasonic cleaner with ethanol for 10 min to remove flux ...
Context 3
... cyclic shear load with displacement control mode was applied to the solder joints, as shown in Fig. 1(b), using a dynamic mechan- ical analyzer (DMA Q800, TA) at ambient temperature of 25 °C. The DMA Q800 system has a force resolution of 0.00001 N and a dis- placement resolution of 1 nm. A fixed gauge length of 14 mm was used when mounting the joint sample in the clamp of the DMA to minimize the effect of substrate length on the shear ...
Context 4
... to three standoff heights of 100, 300 and 500 lm were ±10-40, ±15-75 and ±20-100 lm respectively, which were equal to nominal shear strain amplitudes (c 0 , c 0 = w/ h) of 0.10-0.40, 0.05-0.25 and 0.04-0.20 respectively. The control waveform was a symmetrical trapezoidal wave with a dwell time of 3 s and a stress ratio of R = À1, see Fig. 1(c). To minimize the influence of loading rate on fatigue life of solder joints, the dis- placement loading rate (or ramp rate in the symmetrical trapezoi- dal wave) was set as 5 lm/s. Fractographies and microstructures of solder joints were examined by scanning electron microscopy (SEM, Quanta 200, FEI) equipped with an energy dispersive ...
Context 5
... solder balls are not easily supplied with arbitrary diameters. Only solder balls with limited diameter options, such as 400, 500, 600, 760 lm, are commercially available. To minimize the influence of contact angle (h), as shown in Fig. 1(a), on the anal- ysis, it is necessary to choose suitable diameters of solder balls for matching with the certain standoff heights to unify the contact angles. Therefore, the effect of joint geometry and volume can be reasonably converted to the influence of h (or the ratio of h/d) in this study. The geometry of solder joints can be ...
Context 6
... 7.04 g/cm 3 for molten SAC305; g is the gravitational acceleration constant (i.e., g = 9.80 m/s 2 ); z is the integral displacement of the potential energy in the z axis direction, i.e., the standoff height direction in this work; V is the volume of the material, which can be calculated by the geometry-based truncated sphere method, as marked in Fig. ...
Context 7
... data obtained from lower cycles do not appear to show much difference and may not be representative, while the data from high cycles require a longer testing time and more calcula- tions. Hence, 40 cycles were selected as the representative cycles in FE simulation here. Fig. 11(a)-(c) presents the distribution of the equivalent stress in a BGA joint after undergoing different cycle numbers, which shows almost no difference in the end of each cycle. However, the von Mises plastic strain in the solder near the SAC305/Cu interface rises continuously with increasing the cycle number, as shown in Fig. 11(d)-(f), ...
Context 8
... FE simulation here. Fig. 11(a)-(c) presents the distribution of the equivalent stress in a BGA joint after undergoing different cycle numbers, which shows almost no difference in the end of each cycle. However, the von Mises plastic strain in the solder near the SAC305/Cu interface rises continuously with increasing the cycle number, as shown in Fig. 11(d)-(f), where the maximum von Mises plastic strain increases from 0.1228 to 0.5365 after 40 cycles. In addition, simulation results have shown that the distri- bution of von Mises plastic strain is very similar with that of plastic strain energy density. Accordingly, the accumulated plastic strain in each cycle plays a critical role in the ...
Context 9
... of the plastic strain energy. As the plastic strain energy can be regarded as the energy stored in the solder, the increase of the plastic strain energy pro- vides the driving force for nucleation, growth and coalescence of dimples or microvoids, and further promotes the formation of crack, and finally leads to fatigue failure of the joints. Fig. 12(a)-(c) exhibits the distribution of the equivalent stress in the joints with different h after 40 cycles, where clearly the max- imum equivalent stress decreases with decreasing h under the same c 0 , from 59.7 MPa at h = 500 lm to 37.5 MPa at h = 100 lm (see the values corresponding to the red 1 color contour). However, the change of ...
Context 10
... h after 40 cycles, where clearly the max- imum equivalent stress decreases with decreasing h under the same c 0 , from 59.7 MPa at h = 500 lm to 37.5 MPa at h = 100 lm (see the values corresponding to the red 1 color contour). However, the change of the geometry of joints shows a very limited influence on the equivalent stress, as shown in Fig. 12(d)-(f). This reveals that the maximum von Mises plastic strain decreases significantly with decreasing h, from 3.552 at h = 500 lm to 0.048 at h = 100 lm after 40 cycles (i.e., the values corresponding to the red color contour). Moreover, the von Mises plastic strain exhibits relatively homoge- neous distribution in the solder joint with ...
Context 11
... von Mises plastic strain decreases significantly with decreasing h, from 3.552 at h = 500 lm to 0.048 at h = 100 lm after 40 cycles (i.e., the values corresponding to the red color contour). Moreover, the von Mises plastic strain exhibits relatively homoge- neous distribution in the solder joint with a small h (e.g., h = 100 lm), as shown in Fig. 12(f), while displaying a concentration along the SAC305/Cu interface in the joints with a large h (e.g., h = 500 lm), as shown in Fig. 12(d). Therefore, the magnitude of the plastic strain energy density accumulated in a solder joint with a smaller h is much lower; as a result, the solder joint with the smal- ler h has a longer fatigue life ...
Context 12
... (i.e., the values corresponding to the red color contour). Moreover, the von Mises plastic strain exhibits relatively homoge- neous distribution in the solder joint with a small h (e.g., h = 100 lm), as shown in Fig. 12(f), while displaying a concentration along the SAC305/Cu interface in the joints with a large h (e.g., h = 500 lm), as shown in Fig. 12(d). Therefore, the magnitude of the plastic strain energy density accumulated in a solder joint with a smaller h is much lower; as a result, the solder joint with the smal- ler h has a longer fatigue life compared with the joints with a rela- tively larger h value under the same nominal shear strain ...
Citations
... Qin et al. [17] used Surface Evolver to predict the geometry of solder joints after fixing CSH using spacers and varying the copper pad diameter. The fatigue life performance under cyclic shear load with displacement control was evaluated, and it was reported that the BGA solder joint fatigue life improves with a decrease in the geometric ratio of h/d under nominal shear strain amplitude; conversely, the fatigue life impaired with a decreasing h/d under the same shear displacement amplitude. ...
As electronic packages continue to get smaller, designing reliable solder joints is becoming an increasingly important part of the design process. The shape and height of the solder joint are major factors that can be optimized to improve the reliability of flip chip and ball grid array (BGA) packages. Optimizing the dimensions of the underlying copper pad, which influences solder height and shape, can lead to a longer time to failure for packages undergoing cyclic temperature loads. In this study, the influence of the copper pad design on the thermal fatigue life performance of an off-the-shelf BGA package is investigated. Printed circuit boards with different pad dimensions are employed to analyze the difference in the performance in accelerated thermal cycling tests (ATC) on a BGA package. In the second part of the paper, Surface Evolver is used to predict the shape of the solder joint for different combinations of copper pad diameters on the component and PCB side and its influence on fatigue life test performance is compared. The accumulated plastic work is used to predict the thermo-fatigue life of the different solder geometries.
... It is well known that solder joints are considered the most fragile part of packaging systems and electronic components [6][7][8][9][10]. Previous studies have found that the joint size has a significant impact on solder joint reliability, including the mechanical reliability, such as tensile [11], shear [12], creep [13], and fatigue [14] performances. While, both the solder volume and joint height change simultaneously in some studies [15], leading to the uncertainty of which dimension factor dominating the variation of mechanical performance. ...
... Thus, finite element (FE) simulation was carried out. For simplification, the geometry of Cu 6 Sn 5 layer was designed as a flat layer with a thickness of 5 lm [14]. The element type was SOLID186, and all elements in the FE models were 20-node hexahedron element. ...
In this study, the shear performance and fracture behavior of microscale ball grid array (BGA) structure Cu/Sn–3.0Ag–0.5Cu/Cu joints with same solder volume and different heights at increasing current density were systematically investigated by experimental characterization, theoretical analysis, and finite element simulation. The results showed that the shear strength of the solder joint decreased with increasing current density, while it increased with decreasing joint height. These changes were mainly due to Joule heating, the non-thermal effect of current stressing, and mechanical constraint in the solder joint. As current density increased, both Joule heating and the non-thermal effect of current stressing aggravated solder joint shear strength more and more severely. At the same current density, Joule heating’s deterioration on the shear strength was less serious in the smaller height joint, and the non-thermal effect was insensitive to the change in joint height. The higher shear strength of the smaller height joint was due to the larger mechanical constraint in the solder matrix induced by the substrate. Moreover, as current density increased, the fracture position changed from the solder matrix to the solder/IMC layer interface. The fracture mode shifted from ductile one to ductile–brittle-mixed one, and the critical current density corresponding to the fracture mode transition increased with decreasing joint height. The above findings indicate that the BGA structure solder joint with a smaller joint height is more reliable.
... Solder joints provide electrical, thermal, and mechanical connections among different components and various circuits in electronic products and devices, which are considered the weakest link in electronics [1]. The failure of one solder joint usually results in malfunction or total breakdown of electronics. ...
The shear performance and fracture behavior of microscale ball grid array structure Cu/Sn–3.0Ag–0.5Cu/Cu solder joints with increasing electric current density (from 1.0 × 103 to 6.0 × 103 A/cm2) at various test temperatures (25 °C, 55 °C, 85 °C, 115 °C, 145 °C, and 175 °C) were investigated systematically. Shear strength increases initially, then decreases with increasing current density at a test temperature of no more than 85 °C; the enhancement effect of current stressing on shear strength decreases and finally diminishes with increasing test temperatures. These changes are mainly due to the counteraction of the athermal effect of current stressing and Joule heating. After decoupling and quantifying the contribution of the athermal effect to the shear strength of solder joints, the results show that the influence of the athermal effect presents a transition from an enhancement state to a deterioration state with increasing current density, and the critical current density for the transition decreases with increasing test temperatures. Joule heating is always in a deterioration state on the shear strength of solder joints, which gradually becomes the dominant factor with increasing test temperatures and current density. In addition, the fracture location changes from the solder matrix to the interface between the solder matrix and the intermetallic compound (IMC) layer (the solder/IMC layer interface) with increasing current density, showing a ductile-to-brittle transition. The interfacial fracture is triggered by current crowding at the groove of the IMC layer and driven by mismatch strain at the solder/IMC layer interface, and the critical current density for the occurrence of interfacial fracture decreases with increasing test temperatures.
... In this way, the reliability of solder joints is highly dependent on the formation and growth of IMCs at solder/substrate interface [14]. However, during storage and exploitation of the electronic devices, thermal stress at solder joint interface causes IMCs to grow continuously, owing to the mismatches in physical properties such as thermal expansion coefficient and elastic modulus [15][16][17]. Excessive growth of IMCs could result in a deleterious effect. In other words, the thick IMCs will negatively influence interface integrity and result in mechanical failure of the joints [18,19]. ...
The growth of IMCs at solder/substrate interface becomes more important with the sustaining advancement of integrated circuit technology. The purpose of this study was to investigate the morphology and growth kinetics of the intermetallic compounds (IMCs) formed on the diverse copper lines (including electroless copper, electroplated copper, sputtered copper) of printed circuit board under different conditions (counting isothermal aging, thermal shock and multiple reflow). And the influence of IMCs on signal integrity after isothermal aging treatment was also discussed. The results indicate that IMCs emerged on copper lines with discrepant microstructure have evident differences in morphology and thickness while the microstructure, composition and growth thickness of the IMCs naturally get changed during variational conditions. And overgrown IMCs will further aggravate the signal transmission loss for the microstrip transmission line. These interesting evolutions in IMCs introduce advantageous assistance for further investigation of solder joint reliability in electronic packaging technology.
... 31,32 Finite element (FE) simulations have been applied to calculate local stresses and strains in solder joints to feed reliability models and perform lifetime predictions and design optimization. [33][34][35][36][37] The present report proposes using composite joints consisting of a solder matrix and thin metallic meshes to interconnect semiconductor chips to direct bonded copper (DBC) substrates in power modules. We have already proven the feasibility of this concept and investigated the microstructural features, quasistatic deformation and fracture behavior, and thermal conductivity of such soldermesh joints. ...
... The value of b for SAC305 solder can be narrowed down to a range between À1 and À2. 37 [33][34][35][36][37] The positive effect is caused by the thermal expansion coefficients of Cu and Ni being lower than that of SAC305 according to Table I. The mesh insert decreased the thermal expansion of the solder material in its vicinity. ...
... The reliability of soldered semiconductor chips typically improves with increasing solder joint thickness, which is in agreement with our experimental observations of 110-lm-and 180-lm-thick SAC305 solder joints and the results in Ref. 36. The solder-mesh composite joints had greater thicknesses of 320 lm and 360 lm. ...
Cu-to-Cu and Si-to-direct bonded copper (DBC) assemblies were bonded with SAC305 solder (reference) as well as with composite layers consisting of SAC305 solder matrices with integrated thin Cu and Ni meshes. The microstructural changes at the interfaces, such as the growth of intermetallic compound (IMC) layers and void formation, were monitored during annealing of the Cu-to-Cu samples at constant temperature of 150°C for 1000 h. All three types of sample exhibited sufficient shear strength above 40 MPa with preferable ductile fracture characteristics after annealing. The solder–mesh composite joints in the Si-to-DBC samples provided superior reliability against short-time temperature cycling between 80°C and 200°C. The cracking path was strongly distracted and branched through the mesh insert. The IMC growth rate was found to be reduced in the solder–mesh joints at the Si chip interface. Thermomechanical finite element simulations revealed that the mesh interacted with the solder matrix through plastic deformation and stretching, which could result in reduced thermal expansion of the composite overall.
... Qin et al. [69] utilized the following damage evolution to develop the fatigue life prediction model: ...
Due to the restriction of lead-rich solder and the miniaturization of electronic packaging devices, lead-free solders have replaced lead-rich solders in the past decades; however, it also brings new technical problems. Reliability, fatigue, and drop resistance are of concern in the electronic industry. The paper provides a comprehensive survey of recent research on the methodologies to describe the mechanical behavior of lead-free solders. In order to understand the fundamental mechanical behavior of lead-free solders, the visco-plastic characteristics should be considered in the constitutive modeling. Under mechanical and thermal cycling, fatigue is related to the time to failure and can be predicted based on the analysis to strain, hysteresis energy, and damage accumulation. For electronic devices with potential drop impacts, drop resistance plays an essential role to assess the mechanical reliability of solder joints through experimental studies, establishing the rate-dependent material properties and proposing advanced numerical techniques to model the interconnect failure. The failure mechanisms of solder joints are complicated under coupled electrical-thermal-mechanical loadings, the increased current density can lead to electromigration around the current crowding zone. The induced void initiation and propagation have been investigated based on theoretical approaches to reveal the effects on the mechanical properties of solder joints. To elucidate the dominant mechanisms, the effects of current stressing and elevated temperature on mechanical behavior of lead-free solder have been reviewed. Potential directions for future research have been discussed.
... Meanwhile, the continuous decrease in dimension of solder joints in electronic packages brings about various phenomena in solder joints, including the changes in microstructures and mechanical properties of joints, which is generally known as the size effect. Size effect in solder joints was widely observed under tensile [8], shear [9] and fatigue loads [10]. For solder joints under mechanical loading without electric current stressing, the creep behavior is also significantly influenced by the joint size [11][12][13][14]. ...
... With increasing trend of miniaturization of electronic devices and systems, the dimension of solder joints and pitches has been continuously scaling down. Thus far, many studies have fo und that there are obvious volume and size effects on the fracture and mechanical behavior of solder joints [1][2][3][4][5][6]. With the development of flip-chip (FC) technology, the size of solder joint has been scaled down significantly in recent years, while the amount of current carried by each joint has increased continuously [7] . ...
With increasing miniaturization of electronic devices and systems, the dimension of solder joints and pitches has been continuously scaling down, while the current density carried by solder joints increasing significantly, consequently a critical issue, electromigration (EM), has become a key reliability concern. The EM behavior in the joint is mainly dependent on the magnitude and distribution of the current density and thermal gradient. In this study, thermo-electrical finite element analysis was employed to characterize the influence of microstructure inhomogeneity on the current density and thermal gradient in micro-scale eutectic SnPb (63Sn37Pb) solder joints. Results show that, both geometry factor and microstructure inhomogeneity have obvious influence on the distribution of current density and thermal gradient in the flip chip solder joint. The current density, current crowding ratio and thermal gradient in the Sn-rich phase are much larger than that in the Pb-rich phase, thus the Pb atoms in Sn-rich phase are more prone to migrate under current stressing.
... On the other hands, the joint dimension is another important factor in the evaluation of the reliability of solder joints. Due to the shrinkage of solder joint size, the microstructure and mechanical behavior of solder joints are changed significantly [4][5][6][7][8][9], and there is a so-called "size effect". Some studies have shown the size effect on tensile [8], shear [4] and fatigue behavior of solder joints [9], but it is rare to see the work about the size effect on creep behavior [5,6], even the study of the size effect on creep of bulk solders is also quite limited [10][11][12][13]. ...
... Due to the shrinkage of solder joint size, the microstructure and mechanical behavior of solder joints are changed significantly [4][5][6][7][8][9], and there is a so-called "size effect". Some studies have shown the size effect on tensile [8], shear [4] and fatigue behavior of solder joints [9], but it is rare to see the work about the size effect on creep behavior [5,6], even the study of the size effect on creep of bulk solders is also quite limited [10][11][12][13]. ...
... However, in those previous studies, either the dimension of solder joints is not in micro-scale [15,16], or the current density in solder joints is lower than 5.5x10 3 A/cm 2 [13][14][15][16][17]. Meanwhile, due to the shrinkage of joint dimension, the mechanical behavior of solder joints changes significantly [11,12,[18][19][20][21][22][23][24], and the current density in solder joints increases dramatically [25], which can also influence the reliability of solder joints [25]. ...
The effect of electric current density on the creep deformation and fracture behavior of micro-scale Cu/Sn-3.0Ag-0.5Cu/Cu joints under electro-thermo-mechanical coupled loads was investigated in this study. Results show that all solder joints exhibit similar creep curves displaying typical three-stage creep characteristics regardless of the current density, which suggests that mechanical stress is the dominant factor controlling creep process of solder joints under electro-thermo-mechanical coupled loads. Moreover, the steady-state creep rate increases and the creep lifetime decreases with increasing current density, testing temperature and mechanical tensile stress. However, the creep stress exponent and activation energy do not show obvious change with the increase of current density, meaning that the creep mechanism of solder joints remains under electro-thermo-mechanical coupled loads with different current densities. The creep mechanism is determined to be dominated by lattice diffusion. Furthermore, the fracture position in solder joints changes from the solder matrix to the IMC/solder interface with the increase of current density, temperature and mechanical tensile stress.
