Determination of the Planarization Distance for Copper CMP Process
- PDF / 1,013,521 Bytes
- 6 Pages / 417.6 x 639 pts Page_size
- 44 Downloads / 246 Views
Mat. Res. Soc. Symp. Proc. Vol. 566 © 2000 Materials Research Society
Previous work has examined pattern-dependent planarization issues in oxide CMP processes. The polish rate at a specific site on a patterned wafer has been modeled as an appropriately weighted function of the pattern density over a characteristic region defined by the 'planarization length or distance' [2,31. Such models assumed perfect planarization efficiency and led to linear step height reduction with time relationships. Models which incorporate compressible pads and partial loading of the down regions led to exponential step height reduction with time relationships after the initial linear component [4]. These models, however, neglect changes in effective pressure due to global non-planarity created by topography. An example of this effect occurs during the polishing of next-level metal which incurs preexisting topography as a result of array erosion at the previous level as shown in Figure 2. The resulting array serpentine lines after polish are thicker as a result of initial recess and shielding by the neighboring field region which lowers the effective pressure on the array. Half Overlap: M2=4.510.5 over M1:1/1
11 0.8 -0.6-
.iS:IPAY
R:4 25(0
1F0N4
AUPLCNM
%--,=--
0.2-
U/)
+=Over NoM Topography
%ian 0 50
ILl
>.
,
io.31
1000
1500
20,00
2500
1500 2000 1000 X: Spatial Distance (urn)
2500
10M_1
Data Loading...
