Stress analysis of Cu/low-k interconnect structure during whole Cu-CMP process using finite element method

https://doi.org/10.1016/j.microrel.2013.01.011Get rights and content

Abstract

The stresses of Cu/low-k interconnect structure during the whole Cu-CMP process were studied based on finite element method to analyze the interfacial delamination and fracture of the low-k layer. Effects of the polishing down pressure, low-k modulus, barrier modulus, Cu film thickness, and the coefficient of friction (COF) on the stress distribution were investigated. Simulation results revealed that the probabilities of both interfacial delamination and fracture of low-k layer during all three polishing steps were raised as increasing the polishing down pressure, barrier modulus, and the COF; while increase of low-k modulus made the probabilities decreased. The COF mainly affected interfacial delamination. During bulk Cu removal step, it can be seen that the decrease of Cu thickness made the probability of interfacial delamination increased, but had little effect on the fracture of low-k layer. Among three polishing steps, it was during the barrier polishing step that the risk of interfacial delamination between barrier layer and low-k layer was the largest, at the corner of interconnect structure interface; while the probability of fracture of low-k layer was the highest during overpolishing step, at the corner of low-k layer surface. Moreover, during the same polishing step, effects of the same parameter on interfacial delamination and on fracture of low-k layer were compared.

Introduction

As the feature size of integrated circuits (ICs) continues to decrease, the resistance–capacitance (RC) time delay and the crosstalk become to be the major problem for the fabrication of new generation ICs. Therefore, low-k dielectrics and copper are introduced in order to reduce the parasitic capacitance and the resistance [1], [2]. Different from the conventional metallization technique, copper is integrated into an IC manufacturing process by using the technique of dual Damascene [3], [4], [5]. Then chemical mechanical planarization (CMP) is applied to remove the extra metal material and make the wafer surface globally planarized [6]. However, the application of low-k dielectrics has brought some problems, such as delamination and fracture of low-k material during CMP because of poor adhesion and weak mechanical strength [2], [7], [8]. The CMP-induced delamination was extensively studied by many researchers [9], [10], [11], [12]. Balakumar et al. [9] investigated the mechanism of peeling and delamination during Cu-CMP process using Cu/SiLK single and dual stack structure. Leduc et al. [12] demonstrated that CMP-induced delamination was driven by the work done against the friction force. Accordingly, some methods had been reported to reduce the CMP-induced delamination, including decrease of the CMP polishing down pressure, improvement of the mechanical strength of low-k materials, and enhancement of the adhesion strength between low-k layer and other layers. In addition to experimental researches, simulation investigations had also been performed using finite element (FE) analysis [2], [13], [14], [15]. Compared with experimental approaches, the FE method is convenient for analyzing the effects of CMP process parameters on the stress distribution of Cu/low-k interconnect structure, which is the key factor to cause the delamination and fracture of low-k layer during CMP. It is well known that there are three polishing steps for the whole Cu-CMP process, they are bulk Cu removal step, barrier polishing step, and overpolishing step [9]. The delamination and fracture of low-k layer may happen at any one step. However, most of the existing researches are focused on the bulk Cu removal step.

Consequently, in this paper, stress analysis of Cu/low-k interconnect structure during all the three polishing steps were implemented with FE method in order to thoroughly understand the effects of CMP process parameters on the delamination and fracture. The studied process parameters involved the Cu film thickness, the polishing down pressure, the low-k dielectric modulus, the barrier layer modulus, and the coefficient of friction (COF). Furthermore, the probabilities of delamination and fracture during different polishing step were compared.

Section snippets

Method

Most of low-k dielectric materials used in IC devices are brittle materials, such as porous silica (p-SiO2), fluorine-doped silica (F-SiO2), carbon-doped silica (C-SiO2), and methylsilsesquioxane (MSQ). The fracture of brittle materials always generates at a tensile stress condition, and the maximum principal stress (Rankine) criterion predicts the fracture of brittle materials well [2]. So Rankine criterion is used in our work to analyze the fracture of low-k layer during CMP process, which is

Simulation results and discussion

To study the effects of low-k modulus on interfacial delamination and fracture of low-k layer during CMP process, the low-k modulus changed from 1 GPa to 10 GPa; the polishing down pressure was 6.89 kPa (1 psi), and TaN was used as barrier layer. Fig. 3 shows the maximum ISS between low-k layer and barrier layer, and the maximum MPS within low-k layer as a function of low-k modulus. It can be seen that both maximum ISS and MPS were increased as the decrease of low-k modulus during all three

Conclusions

Simulation models were established to study the failures of Cu/low-k interconnect structure during the whole Cu-CMP process based on FE method. Changes of maximum ISS and maximum MPS were analyzed to investigate the interfacial delamination between barrier layer and low-k layer and the fracture of low-k layer, respectively. The simulation results revealed that during all three polishing steps the increases of polishing down pressure, barrier modulus, and the COF made the probabilities of both

Acknowledgements

The authors appreciate the financial support from 973 Project (Grant No. 2009CB724201), Innovative Research Groups of the National Natural Science Foundation of China (Grant No. 51021064), and National Natural Science Foundation of China (Grant No. 91023016).

References (18)

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

Cited by (7)

View all citing articles on Scopus
View full text