NUMERICAL EVALUATION OF THERMAL WARPAGE ON FLIP CHIP PACKAGE WITH - - PDF document

18TH INTERNATIONAL CONFERENCE ON COMPOSITE MATERIALS

NUMERICAL EVALUATION OF THERMAL WARPAGE ON FLIP CHIP PACKAGE WITH RESPECT TO LAYER RESIDUAL RATE

• W. Song1, Y. Byun2, T. Ku1, J. Kim2, M. Kim3, H. Kang3, B. Kang2*

1 Industrial Liaison Innovation Center, Pusan National University, Busan, S. Korea, 2 Dept. of

Aerospace Eng., Pusan National University, Busan, S. Korea, 3 BGA R&D Group, Samsung Electro-Mechanics Co., Ltd., Chungcheongnamdo, S. Korea

* Corresponding author (Hbskang@pusan.ac.krH)

Keywords: flip chip package, thermal warpage, finite element method, thermal expansion coefficient, glass fiber reinforced epoxy composite, layer residual rate

1 Introduction The reliability problems of flip chip (FC) packages subjected to temperature change during the packaging process mainly occur due to mismatches in the coefficients of thermal expansion [1]. FC package is generally consisted with main chip (Silicon), underfill and bare printed circuit board (PCB), as shown in Fig. 1. Resin molding compounds like underfill and glass fiber reinforced epoxy composites (GFRC) used in FC package strongly exhibit temperature-dependent material properties [2, 3]. In this study, the thermal warpage

• f FC package is evaluated using finite element

analysis (FEA). The thermal warpage of FC package was simulated and compared with respect to the variation of Cu and SR film residual rates in bare

Cu Land Via Solder SR Plugging Epoxy

Prepreg

ABF Die Bump

Numerical Evaluation Part

Fig.1. Schematic diagram of flip chip package. Table 1. The given condition on Cu & SR residual rate in bare printed circuit board.

Layer Residual Rate [%] Upper SR 89.2 Cu 56.6 Prepreg

• Cu

60.2 Prepreg

• Cu

47.0 Prepreg

• Cu

52.1 Lower SR 76.6

PCB, in case that the material selection and thickness of each layer are under the restriction without change. It is noted that the coplanarity of FC package can be enhanced with about 30% as only adjusting the residual rates of Cu and SR film in bare PCB. 2 Finite Element Analysis 2.1 Finite Element Model Numerical simulation is performed for the whole body of FC package with main chip, underfill and bare PCB. Finite element model of FC package is shown in Fig. 2. Bare PCB is constructed by SR film, GFRC prepreg and Cu layers. Main chip and underfill area are modeled by solid element and bare PCB is modeled by layered shell element. The nodes in the sharing region between underfill and bare PCB are merged in the finite element model. Solder bumps in the underfill region are ignored in this analysis due to the relatively weak mechanical characteristics of the bump. The residual rates on Cu

Fixed condition Chip (Silicon) Underfill Bare Printed Circuit Board

Fig.2. Finite element model and layer construction

• f bare printed circuit board.

80 120 160 200 240 0.0 5.0x10

3

1.0x10

4

1.5x10

4

2.0x10

4

2.5x10

4

Elastic Modulus [MPa] Temperature [

• C]

Prepreg (horizontal direction on the in-plane) Prepreg (vertical direction on the in-plane) Prepreg (thickness direction on the out-of-plane) Underfill

(a) Young’s moduli of GFRC and underfill

0.00 0.01 0.02 0.03 0.04 0.05 0.06 30 60 90 120 150

Effective Stress [MPa] Effective Strain

Cu

(b) Stress-strain relationship of Cu Fig.3. Mechanical properties of layer materials in flip chip package. and SR layer were given in Table. 1. In this study, the residual rates are main parameters to evaluate the thermal coplanarity of FC package. 2.2 Material Properties Mechanical properties of underfill and GFRC Prepreg are shown in Fig. 3, which are temperature dependent and measured by DMA Q800 of TA

• instruments. Thermal properties of underfill, GFRC

Prepreg, SR film and Cu are also shown in Fig. 4, which are also temperature dependent and measured by TMA 4000SA of Material Analysis &

• Characterization. The mechanical and thermal

properties of GFRC were tested in orthogonal directions due to the

• rthogonal

material characteristics of GFRC. Fig. 3 shows the stress- strain relation of Cu. Silicon is considered as an isotropic, perfectly-elastic and temperature- independent material with E: 169GPa, CTE: 3.0E-6. 2.3 Simulation Conditions

80 120 160 200 240 40 80 120 160 200

Coefficient of Thermal Expansion, α [ppm] Temperature [

• C]

Underfill SR film

(a) CTEs of underfill and SR film

80 120 160 200 240 10 20 30 40 50

Coefficient of Thermal Expansion, α [ppm] Temperature [

• C]

Prepreg (horizontal direction on the in-plane) Prepreg (vertical direction on the in-plane) Prepreg (thickness direction on the out-of-plane) Cu

(b)CTEs of GFRC and Cu Fig.4. Thermal properties of layer materials in flip chip package.

0.0 5.0k 10.0k 15.0k 20.0k 25.0k 30.0k 35.0k 40.0k

50 100 150 200 250 300

Temperature [

OC]

Time [sec]

125OC 1 hr 150OC 1 hr 165OC 1 hr 175OC 1 hr Peak 240OC Peak 220OC Heating & Cooling rate : 10OC/ min

25OC 25OC 25OC

Bare PCB Level Flip Chip Package Level

Fig.5. Thermal environment in the package process

• f FC package.

Thermal environment in the packaging process of FC package is shown in Fig. 5. In this simulation, the temperature change is assigned in the whole finite element models with room temperature (25oC) after peak value (240oC). To evaluate the thermal

Table 2. Parametric study conditions on Cu & SR residual rates.

Variation of Cu & SR Residual Rate [%] Layer No.1 No.2 No.3 No.4 No.5 No.6 No.7 No.8 No.9 U_SR 1 5 10

• 1

5 10 Cu 1 5 10

• 1

5 10 PPG

• Cu

1 5 10

• 1

5 10 PPG

• Cu
• 1

5 10 1 5 10 PPG

• Cu
• 1

5 10 1 5 10 L_SR

• 1

5 10 1 5 10

Thermal Warpage 60.8um 25OC B-B’ B-B’

Fig.6. Thermal warpage and deforming configuration of the given FC package model.

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09

Thermal Warpage(Coplanarity) [mm]

The given model No.1

• No. 2
• NO. 3
• No. 4
• No. 5

No.6

• No. 7
• No. 8
• No. 9

60.8um 43.2um 50.9um 39.8um 40um (30% ) (16% ) (34% ) 46.7um (23% )

Fig.7. Thermal warpage comparison warpage of FC package, the Cu and SR film residual rates in bare PCB were changed. The variations of Cu and SR film residual rates in bare PCB for each model (No.1 ~ 9) are shown in Table 2. Through the result of the FEA for the given residual rate model in

• Fig. 6., it is noted that bare PCB was deformed as

the convex configuration and main chip was deformed as the concave configuration due to the coefficient of thermal expansion mismatch. 3 Results and Discussions The thermal warpage of FC package was evaluated

58.2um No.1 50.2um 43.2um 59.1um No.2 No.3 No.4 54.5um 50.9um 56.8um 46.7um 39.8um No.5 No.6 No.7 No.8 No.9

Fig.8. Thermal warpage configurations with respect to variation of Cu & SR residual rate. with the variation conditions of the residual rates on Cu and SR film as shown in Table 2, which conditions were determined from the primitive FEA result for the given model. The results of the FEA of the model No.1 to 9 are shown in Fig. 7 and 8. As you can see Fig. 7, the adjustment of the residual rates of Cu and SR film in bare PCB can improve the coplanarity of FC package with about 30% for the temperature change. It is concluded that the proposed concept of the adjustment of residual rates on Cu and SR film can play a feasible role to reduce the thermal warpage in the design stage of FC package as the design guideline in the semiconductor industry.

3

Acknowledgements This work was supported by grants-in-aid for the National Core Research Center program through the National Research Foundation of Korea funded by the MEST (No. R15-2006-022-02002-0). Also the last author would like to thank the partial support by the Korea governments (MEST / KOTEF) through the Human Resource Training Project for Regional Innovation. References

[1] K. Oh and J. J “Submicro-displacement measuring system with Moire interferometer and application to the thermal deformation of PBGA package”. Trans.

• f the KSME(A), Vol. 28, No. 11, pp 1646-1655,

2004. [2] B. Oliveira and G. Creus “An analytical numerical framework for the study of ageing in fiber reinforced polymer composites”. Composite Structures, Vol. 65, pp 443-457, 2004. [3] Y. He “Thermomechanical and viscoelastic behavior

• f a no flow Underfill material for flip chip

applications”. Thermochimica Acta, Vol. 439, pp 127-134, 2005.