Dust Particle Diagnostics in Rf Plasma Deposition of Silicon and Silicon Oxide Films (Invited)
- PDF / 2,471,092 Bytes
- 11 Pages / 414.72 x 648 pts Page_size
- 35 Downloads / 218 Views
plasma reao•chamber
(a)
(b)
vacuum
RF electrode with gas showerhead CCD camera
vacuum /chamber
matching network
S/
5.7 cI 47 c i
7
interference filter
d=2.4cm plasma
Ji
j
generator glass substrate grounded 45cm x 35cm x 1mm reactor walls
windows argon laser beam at 488 nm
bea am 'ander
Figure 1. a) Top and b) front view of the laboratory plasma reactor Film a) in Fig. 2 deposited at low input power shows a very good uniformity but at a low deposition rate of 1.5 A/s. In contrast, film b) deposited at higher input power with powder formation has a deposition rate of 3.4 A/s, with a strongly degraded uniformity. The inhomogeneities are concentrated at the edges of the substrate where the thickness increases rapidly over the last 3 cm. The scattered intensity profiles show clear differences with and without substrate: it would appear that the substrate greatly influences the powder traps. The regions above the substrate where the scattered intensity is higher correspond to regions where the film thickness is also thicker. This shows that inhomogeneous powder distribution gives a non-uniform film thickness. a) power = 50 W
b) power = 100 W
11
9 7 5 3 Position from reactor edge (cm)
1
.
E 1.9.. CdU
S
0
position (cm)
35
Figure 2. Global interferograms for films deposited a) at 50 W (no powder observed); and b) 100 W input power (with powder). The thickness profile corresponds to film b) along the line shown. Curve A represents the vertically-integrated intensity scattered from the powder measured from the CCD image shown; curve B is for the case without substrate. 548
11
9 7 5 3 1 position from reactor edge (cm)
Fig. 3 Interferograms of amorphous silicon films deposited at 13.56 M}Lz; the deposition conditions are identical to Fig. 2 b) except for a glass frame placed around the substrate. Curve A corresponds to the scattered intensity profile with the additional substrate strips and curve B to the case with the substrate alone. In a second experiment, the substrate was surrounded by strips of substrate in order to cover the whole of the bottom area of the ground electrode. The film obtained in this configuration had good homogeneity even at elevated deposition rates (3.6 A/s) since the inhomogeneities were completely confined to the substrate strips.The additional pieces of substrate displace the powder traps that appear at the transition between the substrate and the electrode. This is not a trivial result, because if the powder location were determined purely by fringing electric fields at the reactor edge, then the dielectric inserts would have no effect. Moreover, a 3 mm gap between a glass plate and the substrate reintroduced thickness gradients on the substrate. Geha et al [6] have also demonstrated the use of dielectric inserts to 'tune' the plasma potential in order to reduce particle trapping. The displacement of the powder clouds may be due to a shift in plasma potential wells, or to thermophoretic forces [7] which would arise if the dielectric surfaces were heated by the plasma
Data Loading...
