Environmental Engineering Reference
In-Depth Information
Doping Effects.
The defect structure of Cr
2
O
3
is not fully ascertained but the
fact that Cr
2
O
3
is a p-type semiconductor suggests that the oxygen defects are
oxygen interstitials. Cr
2
O
3
scale growth kinetics further suggests that countercur-
rent diffusion of Cr ions and oxygen along grain boundaries results in oxide
formation within the scale. It is also reported that Cr
2
O
3
is an intrinsic semicon-
ductor above 1200
C. So, Y being trivalent, no doping effect is expected if this
element is dissolved in Cr
2
O
3
. Substitution of Ce or Th in the Cr
2
O
3
lattice, which
can have valencies of 4
°
, would lead to a reduction in the concentration of
interstitial Cr
3
ions and so to a reduction in the outward transport of chromium,
leading to n-type behavior. Thus, dispersions of the reactive elements, where the
metal can assume trivalent state and occupy cation vacancies, can improve oxida-
tion resistance. The main weakness of this model lies in the apparently equal
effects of a wide range of both metallic and oxide dispersion additions. Moreover,
addition of Y or Y
2
O
3
has often been found to affect the rate, which is not ex-
pected from this model.
Formation of a Partial or Complete Blocking Layer.
In this model, it is postu-
lated that a reduction in chromium flux toward the outer interface takes place
owing to the precipitation of rare earth oxides in the alloy matrix acting as a ''set
of sieves in series,'' thus blocking the supply of Cr. Further, in the oxidation
study [40] of TDNiCr (Ni-20%, Cr-2vol%, ThO
2
) it is proposed that the incorpo-
ration of the dispersoid particles in the scale reduces its growth rate by decreasing
the cross-sectional area of Cr
2
O
3
available for transport of chromium. However,
the elegance of this model falls down when additions of even a small quantity
of reactive oxide produce significant reduction in the oxidation rates, whereas a
substantial amount of reactive element oxide is required to bring about a reduction
in the cross-sectional area of the scale.
Short-Circuit Diffusion Model.
This model suggests that the dispersoid particles
on the alloy act as nucleation sites for the first formed oxides, thus decreasing
the internuclei spacing. So the time required for the lateral growth process to
form a complete prohibitive layer of Cr
2
O
3
and terminate the formation of base
metal oxides is reduced. The dispersoid particles may only block the short-circuit
diffusion paths (dislocations) for Cr. The presence of dispersoid in the alloy re-
duces the subsequent oxide grain size, and if it is reduced below a certain level
corresponding to the inter-dislocation distance in a pile-up, a single dislocation,
being unstable in the grain, will move to the grain boundary where it ceases to
be a cation diffusion path. Thus, if dislocations are the short-circuit diffusion
paths for transport of Cr ions in Cr
2
O
3
, reduction in the size of the oxide grain
below a certain limit may reduce the density of such paths and hence the diffusion
rate. On the other hand, the oxygen diffusion rate via grain boundaries will be
increased because larger grain boundary area causes a reversal of the oxide
growth process. Accordingly, the oxide-forming reaction will shift from the
