Deposition, Defect and Weak Bond Formation Processes in a-Si:H

  • PDF / 425,127 Bytes
  • 6 Pages / 414.72 x 648 pts Page_size
  • 33 Downloads / 244 Views

DOWNLOAD

REPORT


There is a need to understand the

25

deposition of hydrogenated amorphous silicon (a-Si:H) in order to minimise its defect density and instabilities for applications such as solar cells, and thin film transistors. It is

-' 20 0

widely observed that the defect density and valence band tail width (due to weak Si-Si bonds) pass through a minimum at deposition temperatures of T, ; 250 0 C (Fig. 1)[1-4]. This minimum depends somewhat on the

50 9 E 9growth > 80 LM *

plasma conditions and the rise at lower

0

Ro

* x 'r=

* ,.

70

".

0.1W 5W

growth rate 15-15W __

.

10ow Beyer* W

Roca

0.1W 5W

rate

low 15W/

[]

"amasakiQ

Yamasa"i temperatures may move to higher T, for , 60 6 higher growth rates. The cause of the rise at 50-2 low temperatures is of particular interest. 5 There are various suggestions for this 1018 minimum. Ganguly and Matsuda [5] suggest a *,\\ "A growth that it arises because 250'C is a maximum in 101711.-, , rate the surface diffusion length of the growth species Sill3. Street [6] and Winer [7] '. suggest that defects are in a hydrogen- . 1016 mediated equilibrium with the Si network, . and this equilibrium fails a lower T, because 1015 100 200 300 400 500 60 0 hydrogen diffusion is too slow to diffuse even Deposition temperature (C) one atomic spacing. After growth, there is general agreement that the defect density Fig.1 Variation ofhydrogencontent, Urhachslope depends via denityof the defect pool model on a fixed aid Variation hydoge co with t hdeposition slop eakbonsqueche-indurngand defect densityof of a-Si:H density of weak bonds, quenched-in during temperature [4-7]. growth [8]. Thus in many ways, weak bonds are a more fundamental property of the network, and a growth model should describe not just the origin of defects (dangling bonds) but also of weak bonds. Finally, we note that the growth surface is largely saturated with hydrogen, but

897 Mat. Res. Soc. Symp. Proc. Vol. 507 01998 Materials Research Society

the bulk has a much lower H content. The H evolution process which converts the surface layer to bulk a-Si:H in a subsurface growth zone is not described properly in any existing model. SURFACE GROWTH MODEL

SiH3

desorption

reflection,

recombination as Si2H6

Plasma enhanced chemical vapour deposition (PECVD) H abstraction involves the creation of a growth giving SiH4 physisorption ° o/ species in the plasma, the reaction 7'= P3-s PQQIQOQOs, of the growth species at the a-Si:H surface, and the change of the icking (growth) surface Si-H layer into the bulk aSi:H structure in the subsurface 0 I I I I I I I I I I I I I I I I I I I I I zone. There is general agreement a-Si:H surface with Si-H bonds from plasma diagnostics that the growth species for high quality a- Fig. 2. Surface processes of a Sill radical. 3 Si:H is the monoradical Sill3 [911]. The a-Si:H surface is essentially fully hydrogen covered by either -Si-H, =Sill2 or -Sill3 groups. Sill3 radicals become physisorbed on these groups by forming a 3-centre Si-H-Si bond. This model has been studied in detail by Mat