Warpage mechanism analyses of strip panel type PBGA chip packaging

https://doi.org/10.1016/j.microrel.2009.12.010Get rights and content

Abstract

The objective of this study is to analyze a warpage development mechanism by simulating a strip type packaging for plastic ball grid array. Molding compound and substrate materials were thermo-mechanically tested to obtain the mechanical properties by several test methods. Samples were fabricated using the same materials, and warpage developments were measured at room temperature after molding compound cure. Based on the tested materials property, the warpage developments were simulated by numerical calculations during cooldown process. The results were compared with the measurement data of the samples, and the warpage mechanism was investigated based on the elastic and viscoelastic simulation results. It was found that the relaxation behaviors of the molding compound and the substrate materials had significant effect on the warpage development. It was also found that the warpage development was dependent on the packaging geometry. The development mechanism was analyzed through the simulation calculations by combining different material properties modeling and geometries, and the results showed comprehensive consideration of the materials and the packaging design are essential to control the warpage.

Introduction

Typical process of plastic ball grid array packaging starts from silicon chip attachment on strip type substrate, followed by molding compound encapsulation through injection molding method. The fabricated strip undergoes hot temperature for molding compound cure, and cooldown to room temperature. Then the strip is cut into individual packaging units. In this process, the warpage development is frequently encountered as one of major technical problems in the packaging process. Obvious problem induced by the warpage appear when the strip proceeds into the singulation process of each unit. Even after the singulation, the warpage of individual unit poses various problems such as non-proper soldering on printed circuit board, leading to high stress generation in the joints. Even after the process, the environmental temperature changes to which the assembled unit is exposed create repeated warpage, which may fail the device function. Eventually, warpage is the main reason to threaten the device reliability. The trend toward high performance multi functional device in compact size demands very thin thicknesses of chip and board, and they deform easily even under small stresses raised during fabrication and assembly processes, and field service.

The warpage has been long time issue, and the stress and deformation control is still great challenge to packaging engineers. The analysis is not easy task due to the nature of complexity of packaging structure and multiple steps involved in the packaging process with different materials. Proper defining and understanding the warpage mechanism allow appropriate material selections and process design. For the mechanism study of the warpage development, numerical simulation has been popular choice over actually making and testing prototypes in terms of time and cost. However, the simulation results often contradict observed sample behavior, and the mechanism interpretation is not well defined yet. While the numerical scheme such as finite element method is competent, the problems in the analyses stems mainly from material property data, and the lack of thorough understanding on the engineering issues involved in the packaging process. It is widely accepted that the warpage develops due to the non-symmetry of the packaging structure and heterogeneity of the mechanical properties of component materials. Especially, unlike silicon, the molding compound and FR-4 substrate are polymer-based, and their properties are time- and temperature-dependent due to viscoelastic characteristics. Therefore, precise measurements and modelings of the material properties are major concerns for accurate simulation results. Packaging engineers often experience difficulties in acquiring reliable material property data, and mostly depend on manufacturers’ information. However, the data are occasionally found to be incorrect and insufficient for the simulation purpose. Even when the materials were directly tested for the calculation, the properties were modeled in various ways according to the researchers’ interests. Many literatures are available for the material characterization by different measurements and modelings for molding compounds to estimate the residual stresses in the microelectronics packaging. Teng and Hwang [1] tested the molding compound to model the material properties during cure by various cure condition parameters for the chemical shrinkage estimation. However, they used single Young’s modulus during cooldown to calculate warpage. Miyake et al. [2] also considered the cure process by addressing chemical shrinkage, and they used the dynamic tensile modulus for elastic shear modulus. Yang et al. [3] modeled the entire molding compound behavior by viscoelastic manner to include the cure effect as well as time and temperature effects by employing shear dynamic mechanical analyses. Tsai [4] modeled the molding compound behavior as temperature dependent elastic during cooldown based on the storage modulus obtained by the dynamic mechanical tests. Similar approach was often found in other literatures [5], [6], [7]. However, the storage modulus is another type of the modulus in frequency domain, and it cannot replace the Young’s modulus without proper conversion process. Interestingly, actual verification of the simulation results by comparing the sample measurements were rarely attempted in the previous literatures.

Unlike the molding compound, the FR-4 substrate had not been characterized much. Despite the fact that the substrate exhibits viscoelastic behavior [8], [9], [10], it has been modeled mostly as elastic [7], [11], [12], [13]. The first numerical study on the bare chip warpage development with viscoelastic FR-4 consideration was investigated by Kim [10]. Simultaneous consideration of the viscoelastic characteristics of the molding compound and the substrate in the warpage analyses was recently investigated by Kim and Lee using numerical approach for ball grid array chip [14]. To the author’s knowledge, there has not been any study on the warpage development by taking the viscoelastic characteristics of the molding compound and the substrate into account at the same time, and comparing the simulation results with experimental data for the mechanism analyses.

In this study, the warpage mechanism is investigated using strip type packaging for plastic molding ball grid array, a typical packaging structure for chip level. The structure is composed of the molding compound, FR-4 substrate, and silicon chip, especially prone to deformation creating various engineering problems. In this study, the fully cured molding compound and substrate material were carefully tested to obtain the elastic and viscoelastic mechanical material properties. The data was modeled and used for the simulations to calculate the warpage developments during cooldown in the fabrication process. The chemical shrinkage of the molding compound was also measured, and included in the calculations by adding extra temperature profile in the cooldown. Samples were made using the same materials, and the deformations induced during the cooldown were measured after the molding cure. The simulation results and the data were compared for the verification, and the warpage mechanism was analyzed by calculating various cases, especially based on the relaxation behaviors of the materials and the packaging structure geometry.

Section snippets

Material property measurements

The substrate material used in this study was 7409HGB (manufactured by Doosan Electro-materials, Korea). The single layer thickness of the substrate was 0.1 mm, and the material is basically a woven fabric reinforced epoxy resin composite. For the mold encapsulation, two different types of the epoxy-based molding compound (Type-A and Type-B manufactured by Jaeil Textile Industry, Korea) were provided, and used.

Sample manufacturing

Strip-type samples were fabricated using the substrate and the molding compounds Type-A and Type-B, which were named according to the molding compound type. The geometric dimension is shown in Fig. 10. The substrate thickness was 200 μ, the chip thickness 160 μ, and the molding compound thickness 910 μ, therefore the total thickness was 1110 μ. The width and length of the strip were 220 mm and 62 mm, respectively. The chip size was 4.8 × 3.4 mm, and total 42 chips were attached on the substrate. After

Comments on temperature dependent elastic modulus

The temperature dependent elastic modulus has been popularly employed in many literatures to analyze the temperature effects on the warpage development. We reconsidered the nature of the modulus, and found two major problems. First, the temperature dependent modulus has been commonly adopted from the storage modulus (E′) by DMA test under 1 Hz, and mistakenly used as Young’s modulus (E) without discrimination. As was indicated previously, the storage modulus is different from Young’s modulus,

Calculations and discussions

The warpage development was simulated based on the packaging geometry and the material property data obtained in the previous chapters. To take advantage of the symmetry, a quarter of the sample geometry (see Fig. 10) was modeled by ANSYS (V. 11.0) using the shell element 281, and the symmetry boundary conditions were applied along the center lines of the sample. The material properties obtained by the tests were implemented in the program. Poisson’s ratios of the molding compound and the

Conclusions

This study represented the mechanism of the warpage development by testing and simulating the strip type PBGA chip packaging structure. The molding compound and the substrate were tested and modeled as elastic and viscoelastic. The calculation results showed that the solutions based on the simultaneous modeling of the viscoelastic properties of the molding compounds and the substrate provided the best agreement with the data. The mechanism was studied by employing different types of the

References (16)

There are more references available in the full text version of this article.

Cited by (27)

  • A review of three-dimensional viscoelastic models with an application to viscoelasticity characterization using nanoindentation

    2012, Microelectronics Reliability
    Citation Excerpt :

    On the other hand, most mold compounds exhibit viscoelastic behavior even at room temperature, even though these polymers are filled as much as 90% by weight with silica filler particles. Kim et al. [5] studied the warpage mechanism by simulating a strip type packaging for plastic grid array, and compared the elastic and viscoelastic models for molding compound and substrate. They found that the relaxation behaviors of the molding compound and the substrate materials had significant effect on the warpage development.

  • Material property effects on solder failure analyses

    2011, Microelectronics Reliability
    Citation Excerpt :

    The objectives of this study are to examine the material modeling effects on the solder ball reliability analyses in the numerical simulation technology. The better accuracy of the viscoelastic calculations was already presented in the previous work [4]. Therefore, the analyses in this paper focus on the discrepancies of mechanism interpretations based on the calculation results when the materials are modeled in different ways.

  • Simulation and experimental analysis of large area substrate overmolding with epoxy molding compounds

    2011, Microelectronics Reliability
    Citation Excerpt :

    Consequently the simulation of thermo-mechanical stresses is restricted to temperature changes and subsequent failures like delaminations and popcorning [6,7]. Some research groups started to implement polymerization stresses into their models in order to investigate substrate warpage after molding [8–10]. Nevertheless there is the need to monitor also the pressure acting on the electronic parts during the polymer encapsulation process considering e.g. the spreading usage of MEMS sensors within automotive industry and their increased demand for signal stability, production tolerances and cost [11–13].

View all citing articles on Scopus
View full text