Positron Annihilation Lifetime Spectroscopy (Pals) Application in Metal Barrier Layer Integrity for Porous Low-K Materia
- PDF / 4,560,578 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 33 Downloads / 229 Views
POSITRON ANNIHILATION LIFETIME SPECTROSCOPY (PALS) APPLICATION IN METAL BARRIER LAYER INTEGRITY FOR POROUS LOW-K MATERIALS Simon Lin a, d), Jia-Ning Sun b), David W. Gidley b), Jeffrey T. Wetzel a), K.A. Monnig a), E. Todd Ryan a, c), Simon Jang d), Douglas Yu d) and M.S. Liang d) a) International SEMATECH, Austin, TX b) Department of Physics, University of Michigan, Ann Arbor, MI c) Advanced Micro Devices, Austin, TX d) Taiwan Semiconductor Manufacturing Co, HsinChu, TAIWAN, R.O.C.
ABSTRACT Positron Annihilation Lifetime Spectroscopy (PALS) (1, 2) is a useful tool to pre-screen metal barrier integrity for Si-based porous low-k dielectrics. Pore size of low-k, thickness of metal barrier Ta, positronium (Ps) leakage from PALS, trench sidewall morphology, electrical test from one level metal (1LM) pattern wafer and Cu diffusion analysis were all correlated. Macroporous low-k (pore size >= 200A) and large scale meso-porous low-k (>50~200A) encounter both Ps leakage and Cu diffusion into low-k dielectric in the 0.25µmL/0.3µmS structures when using SEMATECH in-house PVD Ta 250A as barrier layer. For small scale meso-porous (>20~50A) and micro-porous (10%) of mesoporous low-k (> 20A), both PALS and SANS can detect the same order of pore sizes. For micro-porous (=200A) were evaluated. They are JSR LKD-5109 (~20A), Dow Corning XLK 2.2 (~50A), Honeywell Nanoglass-C (~70A) and UCL-C (~200A), an early version of UCL film. The morphologies of the pore structures of these four porous low-k materials (with or without cap) after cleavage are shown in Figure 1. In order to have clearer pore structure comparison among them all, images after FIB for all four materials were also prepared as shown in Figure 2. Although FIB might be able to smooth the surface, the exposed pores after FIB indicate their relative pore sizes. Trench sidewall morphologies were also examined after etch and ash processes as shown in Figure 3. Columnar structures on the sidewall are transferred pattern after etch and ash from developed photo resist. Pores are observed on the trench sidewall from the large scale mesoporous low-k and larger. However, for small scale meso-porous and smaller low-k (pore size 1E-11Amp). Comparable capacitance distribution for LKD-5109, XLK 2.2, NG-C and UCL-C probably suggest the same failure root causes. Nanoglass-C was evaluated by 0.35µm L/ 0.4µm S structure using a different test vehicle. Its capacitance comparison with others is after structure correction. Cu diffusion through barrier layer into low-k is verified by SEM, TEM and EDS results. Figure 8 shows Cu diffusion in UCL-C and NG-C but not in LKD-5109 and XLK 2.2. The barrier integrity problems are therefore concluded as pore size differences based on the same metallization porcesses. The tested die of UCL-C shows bridge current of 1E-6.8 Amps and capacitance of 7.4 pF. The results also indicate that entrapped Cu atom is not so mobile to cause complete comb shortage under the test conditions, 1V at100MHz and 400µm finger length. A9.7.4
The comparison metrics among diff
Data Loading...
