### Table 5: Raphael 3-D #0Celd solver results for total capacitance extraction of a single

1999

"... In PAGE 29: ...Fill Impact on Extraction Table5 shows capacitance extraction results obtained with the Raphael 3-D #0Celd solver from TMA #2F Avant!, for an isolated conductor #28i#29 with or without #0Cll insertion in empty regions of adjacentlayers, and #28ii#29 with or without same-layer neighbor conductors. 17 The simulation shows that ignoring the possibility of metal #0Cll can result in underestimation of total line capacitance by more than 50#25.... ..."

Cited by 19

### Table 5: Raphael 3-D #0Celd solver results for total capacitance extraction of a single

1999

"... In PAGE 29: ...Fill Impact on Extraction Table5 shows capacitance extraction results obtained with the Raphael 3-D #0Celd solver from TMA #2F Avant!, for an isolated conductor #28i#29 with or without #0Cll insertion in empty regions of adjacentlayers, and #28ii#29 with or without same-layer neighbor conductors. 17 The simulation shows that ignoring the possibility of metal #0Cll can result in underestimation of total line capacitance by more than 50#25.... ..."

Cited by 19

### Table 6: TMA#2FAvant! Raphael capacitance extraction results: total capacitance

1999

"... In PAGE 30: ... When there is no o#0Bset, the #0Cll geometries lie directly under the victim conductors. Table6 shows that the total capacitance values for the middle conductor #28B#29 #0Ductuate by less than 1 percent over all four combinations of #0Cll pattern and o#0Bset. The critical factor is that the #0Cll is presentinthe #0Crst place.... ..."

Cited by 19

### Table 3: TMA/Avant! Raphael capacitance extraction results: total capacitance for the middle victim conductor B.

1998

"... In PAGE 8: ... When there is no offset, the fill geometries lie directly under the victim conductors. Table3 shows that the total capacitance values for the middle con- ductor (B) fluctuate by less than 1 percent over all four combinations of fill pattern and offset. The critical factor is that the fill is present in the Victim A, C Total Capacitance (10,15F) Fill layer offset Fill geometry e = 3:9 e = 2:7 N 10#021 3.... ..."

Cited by 19

### Table 6: TMA#2FAvant! Raphael capacitance extraction results: total capacitance

1999

"... In PAGE 30: ... When there is no o#0Bset, the #0Cll geometries lie directly under the victim conductors. Table6 shows that the total capacitance values for the middle conductor #28B#29 #0Ductuate by less than 1 percent over all four combinations of #0Cll pattern and o#0Bset. The critical factor is that the #0Cll is presentinthe #0Crst place.... ..."

Cited by 19

### Table 2: Model parameters are explicitly represented so that engineers may change these parameters and experiment with different behaviours (after Raphael, 2000)

2000

"... In PAGE 7: ... This is because the behaviour of a structure, which is represented by a model, may change significantly when model parameters are changed. Table2 contains model parameters that can be changed. The same assumption classification is used as before.... ..."

### TABLE VII TMA/AvANT! RAPHAEL CAPACITANCE EXTRACTION RESULTS: TOTAL CAPACITANCE FOR THE OUTSIDE VICTIM CONDUCTOR A OR C Victim A, C Total Capacitance (10 -F) Fill layer offset g eometry c . C = 2. 7

1999

Cited by 19

### Table 5 shows capacitance extraction results obtained with the Raphael 3-D eld solver from TMA / Avant!, for an isolated conductor (i) with or without ll insertion in empty regions of adjacent layers, and (ii) with or without same-layer neighbor conductors.17 The simulation shows that ignoring the possibility of metal ll can result in underestimation of total line capacitance by more than 50%. This can in turn lead to inaccurate RCX, delay calculation, and timing analysis results. We conclude that the presence or absence of ll geometries must be modeled during performance-driven layout optimization. Such modeling must be e cient and \transparent quot;; since there are many iterations through the layout optimization loop, we must be careful with the time complexity of ll insertion and the increases in data volume.

"... In PAGE 29: ... Table5 : Raphael 3-D eld solver results for total capacitance extraction of a single victim conductor. The conductor on layer i is 20 1.... ..."

### Table 5 shows capacitance extraction results obtained with the Raphael 3-D eld solver from TMA / Avant!, for an isolated conductor (i) with or without ll insertion in empty regions of adjacent layers, and (ii) with or without same-layer neighbor conductors.17 The simulation shows that ignoring the possibility of metal ll can result in underestimation of total line capacitance by more than 50%. This can in turn lead to inaccurate RCX, delay calculation, and timing analysis results. We conclude that the presence or absence of ll geometries must be modeled during performance-driven layout optimization. Such modeling must be e cient and \transparent quot;; since there are many iterations through the layout optimization loop, we must be careful with the time complexity of ll insertion and the increases in data volume.

"... In PAGE 29: ... Table5 : Raphael 3-D eld solver results for total capacitance extraction of a single victim conductor. The conductor on layer i is 20 1.... ..."

### Table 2: Raphael 3-D field solver results for total capacitance extrac- tion of a single victim conductor. The conductor on layer i is 20 1. Line-to-line spacing is 1, line width is 1, line thickness is 1.5, and dielectric height is 1.5. Metal fill features on layers i?1 and i+1 are 10 1 with side-to-side spacing of 1 and end-to-end spacing of 4. The dielectric permittivity was set to both 3.9 (for SiO2) and 2.7 (cf. recent announcements by Sematech [12] of new low-permittivity dielectric technologies). Layers i?2 and i+2 are set to be 40 40 ground planes.

1998

"... In PAGE 7: ...14] K. Wampler and T. Laidig, personal communication, Sept. 1997. Appendix: Fill Impact on Extraction Table2 shows capacitance extraction results obtained with the Raphael 3-D field solver from TMA/Avant!, for an isolated conductor (i) with or without fill insertion in empty regions of adjacent layers, and (ii) with or without same-layer neighbor conductors. The simulation shows that ignoring the possibility of metal fill can result in underestimation of total line capacitance by more than 50%.... ..."

Cited by 19