Characterization of plastic and creep behavior in thick aluminum wire for power modules

https://doi.org/10.1016/j.microrel.2021.114185Get rights and content

Highlights

  • Demonstration of tensile test for aluminum wires with different strain rates

  • Quantification of plastic and creep deformation behavior of aluminum wire

  • Validation of the deformation model covering with a wide temperature range

Abstract

One of the common reliability concern on most of power modules is a failure of its bonding. Stress, strain, and displacement fields around those bonding in power modules are required for highly-reliable design based on physically/mechanically based failure model. They can be accurately obtained from CAE tools such as finite element computer codes for stress analysis with mechanical properties of materials not only in the elastic region but also in the inelastic region. In this paper, for the reliability study of an aluminum bonding wire, an experimental study was performed to acquire its plastic and creep characteristics. The temperature-dependent constitutive equations for plastic and creep behavior of an aluminum wire are presented based on the isothermal tensile test data. These constitutive equations can be used at the temperatures ranging from 20 °C to 250°C, and are useful for estimating fatigue life predictions of bonding wires, when we utilize the failure models based on the physical quantities such as the inelastic strain range, the inelastic strain energy density range, the nonlinear fracture mechanics parameter T* range and so on.

Introduction

Power semiconductor modules are the key devices for power supply and have been in widespread use for many applications. They require high reliability because a failure of power modules leads to shutdown of the electronic system and even the overall system.

One of the common reliability concern on most of power modules is a wire bond interconnect. Metal wires are ultrasonically bonded to electrically connect the top pad of the semiconductor device to external electrodes. Temperature change causes thermal stress at the wire bond interconnects due to the mismatch of the coefficient of thermal expansion of material. The repeated thermal stress during operation can conclude thermal fatigue failure. The lifetime prediction methods of wire bond interconnects in power modules can be classified into two categories. One is based on empirical models developed from accelerated lifetime testing data. Another is based on physically/mechanically based failure models, but it is still limited due to the lack of detailed information of the materials. Apart from those failure-based approaches, the remaining useful life (RUL) approach can also estimate the remaining lifetime by monitoring failure precursor parameters, e.g. RON [1]. However, the RUL approach is limited for a specific module which includes the corresponding sensor functions, and requires both actual modules and a monitoring system to design the lifetime of the power module in advance.

In Chapter 2, we will review both the failure-based lifetime prediction methods of wire bond interconnect, and emphasize the importance of physical/mechanical engineering viewpoint, especially finite element stress analysis. Considering the temperature of power module operation, stress analysis requires not only the elastic material properties but also the inelastic material properties such as plastic and creep deformations of a bonding wire. Any corresponding information of materials is rarely available for the inelastic analysis of thick aluminum wires used in a power module. Thus, a newly-developed methodology for providing the material properties of Al wires required for physically/mechanically based failure model using finite element stress analysis was demonstrated in this paper. Tensile tests were performed at the different strain rates under the isothermal conditions. The experimental data of those tensile tests for aluminum wires are shown in Chapter 3. And, the methodology for obtaining both plastic and creep characteristics from the experimental data is explained in detail in Chapter 4. The obtained temperature-dependent constitutive equations for plastic and creep behavior of an aluminum wire are also presented, which can be used for finite element analysis at the temperatures ranging from 20 °C to 250°C. The concluding remarks are given in Chapter 5, which emphasize a paradigm shift from estimating residual lifetimes of in-service power modules using the purely empirical failure models to estimating the lifetimes in the design phase of power modules using the physically/mechanically based failure models.

Section snippets

A brief review of the lifetime prediction methods of wire bond interconnects

The main failure mode of wire bond interconnect is either wire-liftoff or heel crack. A lot of papers have been published for the failure mode of wire-liftoff, in which a bonding wire is delaminated from a top pad of semiconductor chip. Most of the lifetime models are proposed as a function of the temperature range ∆T for estimating the wire-liftoff lifetime Nf for power modules [2]. Among them, the typical lifetime model is as follows:

Bayerer's model [3]:Nf=A1Tn1exp.β2/Tmintonβ3Iβ4Vβ5Dβ6

This

Tensile tests for a bonding wire

The tensile tests were performed to acquire the inelastic behavior of a wire for power modules. The tested wire is 400μm in diameter aluminum wire TANW type made by TANAKA Denshi Kogyo K.K. Fig. 1(a) shows the testing apparatus for tensile tests under isothermal conditions. An aluminum wire is subjected to tensile load in a heat chamber. Fig. 1(b) shows a wire specimen fixed with jigs inside the heat chamber. The stiffness of the testing machine is 6kN/mm, which is under the lowest condition

Methodology for obtaining both plastic and creep characteristics from experimental data

Here we will show how to obtain both plastic and creep characteristics from the experimental data shown in Chapter 3.

The total strain εtotal obtained in an isothermal tensile test is decomposed asεtotal=εe+εp+εcwhere εe, εp and εc represent elastic, plastic and creep strains, respectively. We employ the Ludwik type constitutive equation [18] for plasticity and the Norton-Bailey type constitutive equation for transient creep [19,20]. They are given as follows:σ=σ0+αεpβεc=Aσntm

In Eq. (10), σ0, α

Concluding remarks

In this paper, the temperature-dependent constitutive equations for plastic and creep behavior of an aluminum wire are presented based on the isothermal tensile test data obtained from the present study. These constitutive equations can be used at the temperatures ranging from 20°C to 250°C. The constitutive equations proposed in this study are useful for estimating fatigue life predictions of bonding wires, when we utilize the failure models based on the physical quantities such as the

CRediT authorship contribution statement

Nobuyuki Shishido: Methodology, Data curation, Writing – Original Draft, Writing – Reviewing and Editing.

Yoshiki Setoguchi: Investigation.

Yuto Kumagai: Formal analysis.

Masaaki Koganemaru: Conceptualization, Software.

Toru Ikeda: Project administration, Resources.

Yutaka Hayama: Validation.

Noriyuki Miyazaki: Supervision.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

This work was supported by JSPS KAKENHI Grant Number JP18K03863.

References (27)

  • A. Hanif et al.

    A comprehensive review toward the state-of-art in failure and lifetime predictions of power electronic devices

    IEEE Trans. Power Electron.

    (2019)
  • M. Held et al.

    Fast power cycling test for insulated gate bipolar transistor modules in traction application

    Int. J. Electron.

    (1999)
  • Q. Yu et al.

    Analytical and experimental hybrid study on thermal fatigue strength of electronic solder joints (2nd report, evaluation by isothermal mechanical fatigue test)

    Transact. Jpn Soc. Mech. Eng. (Part A)

    (1998)
  • Cited by (0)

    View full text