The Behaviors of Alkaline Slurry during the CMP of TSV Backside Heterogeneous Microstructure

Abstract

In the integrated circuits industry, 3D integration technology has been extensively studied. Through silicon via (TSV) technology is the core technology of the 3D integration. Chemical mechanical planarization (CMP) process, due to its high ability to achieve local and global planarization, is introduced into the traditional 1 st back side via reveal (BVR) process during the back side thinning and flattening of TSV wafers. In this paper, CMP experiments were performed to study the selective removal of the back side TSV heterogeneous microstructure using alkaline slurry with hydrogen peroxide as oxidant. Atomic force microscope (AFM) measurements were used to evaluate the surface topography, and energy dispersive spectrometer (EDS) was used to characterize the surface chemistry. Scratch experiments were also conducted by AFM method to characterize the removal mechanism on the Cu surface during CMP. Based on the experimental results, the behaviors and mechanism of H 2 O 2 , BTA and pH value in different CMP slurries were discussed. The results show that a slurry contains 0.3 wt%-0.6 wt% H 2 O 2 , 15 mM of BTA and pH value of 10.5-11 is suitable to achieve a smooth polished surface. Findings in the research will have a guiding significance to the research of TSV back side CMP and BVR process.

A Research of the Micro-Damage of the Backside TSV Heterogeneous Microstructure after CMP Process with Alkaline Slurry

“…The results show that BTA can help to obviously reduce the corrosion effect from slurry to Cu surface (static corrosion efficiency can reach to 95% in H 2 O 2 based slurry) and the adsorption of BTA involves both physisorption and chemisorption, which produces (BTAH) ads and Cu-BTA on Cu surface, respectively. And in our previous work, 5,21 it has been put forward that the addition of BTA can form a "soft" adsorption film on Cu surface, which is easier to be removed by mechanical effect due to the AFM scratch experiments. Since the surface materials are removed "gently", the surface quality is improved, which is side proved by the reduction of Cu surface roughness.…”

“…The thickness values of Amorphous A and B of different CMP treated samples were measured and are shown in Table I. There are follow discussions: BTA content barely doesn't change the thickness since it has barely no effect to the Si surface; large H 2 O 2 content leads to large thickness of B since H 2 O 2 promotes the passivation of Si surface, thickness of A doesn't change much when H 2 O 2 content is 0.3 wt% or 0.9 wt%, but decreases obviously when there's no H 2 O 2 since the absence of H 2 O 2 passivation effect leads to a high removal rate of Si 5,21 surface and a small thickness of Amorphous A; when pH value varies from 9-12, larger pH value leads to a thicker B layer since large hydroxide ion concentration promotes the dissolution of the surface passivation layer, thickness of A layer doesn't change much when pH = 9 and pH = 10.5, but decreases when pH value increases to 12, we believe that the large hydroxide ion concentration when pH = 12 can also apply dissolution effect to the monocrystalline Si despite of passivation materials and leads to a thin Amorphous A layer, which needs to be further investigated.…”

“…3 It is supposed to expose the Cu pillars to a certain height and achieve a smooth and flat surface, 4 and 1 st back side via reveal (BVR) process is right after the removal of substrate Si by precision grinding. As discussed in the previous work, 5 there are some shortcomings existing in the traditional wet etching process, and CMP is expected to be more promising in the 1 st BVR process. As a precise global planarization technology, CMP can eliminate the height error and the overall height difference of Cu pillars 4 caused by deep hole etching and achieve a sub-nanometer roughness flat surface.…”

“…During the CMP process used in 1 st BVR, Cu and Si are simultaneously polished, and CMP slurry is a key element during this process of heterogeneous material removal. As conducted in our previous work, 5,21 alkaline CMP slurries with or without hydrogen peroxide (H 2 O 2 ) were studied, the polishing effects of the slurries with different pH values, H 2 O 2 content and benzotriazole (BTA) content were characterized and the mechanisms were analyzed. But in our previous work, there isn't enough work about mechanism investigation, especially about the back side surface and subsurface micro-damage.…”

“…But in our previous work, there isn't enough work about mechanism investigation, especially about the back side surface and subsurface micro-damage. Since as mentioned in our previous work, 5 slurry plays an important role during CMP selective removal, the regular patterns of the damage with the changing of different slurries' parameters are important to be investigated.…”