Prakt. Met. Sonderband 38 (2006) 147 
1GPa for Element content [At%] 
100 
80 Cr 20 — 01s Y 3d -—— Fe 2p —— Al2s 
60 
40 
20 
Depth 
0 0 50 100 150 200 250450 650 850 1050 1250 1450 1650 2050 [nm] 
Figure 5 Shallow and deep SDPs of PM2000 obtained from XPS 
Discussion 
The analyses of T91 and HT-9 after 200, 400 and 600h test in LBE show the typical multiple 
layer structure [4, 5]. But the layer compositions cannot be related to a clear oxide 
composition as observed for gas oxidation [6] where the outer layer is know as Fe;O, and the 
inner layer is know as (Fe,Cr),03 spinel. Therefore more measurements are needed in order 
to clearly characterize the layers and their microstructures. Slight silicon enrichment between 
the layers is found in both materials. HT-9 shows lower silicon enrichment than T91. Since 
HT-9 has a higher carbon content than T91, it may be assumed that the silicon in HT-9 is 
used up by silicon-carbides and cannot diffuse into the oxide layers. The oxide layers on HT-9 
Therefore are slightly thicker than those on T91. Therefore it seems that more silicon in the layer 
e material impedes the diffusion which is already described in the literature [7]. EP823 has much higher 
inum and silicon content than T91 or HT-9. Therefore it follows that the silicon enrichment in the layers 
smuth, no is much higher, and the total thickness of the oxide is lower. So far it is not clear why the inner 
N layer of EP823 is not as homogeneous as in T91 and HT-9. It is assumed that the silicon 
2d on the diffusion and enrichment is microstructure (grain orientation and structure) related, leading to 
de clearly this inhomogeneous behavior. To support this hypothesis further work is needed. The nano 
However, indentation measurement shows hardness and E-modulus results for the steels as expected 
on all of them (~4GPa hardness, ~200GPa E-modulus). Throughout the inner oxide layer the 
hardness does not increase as much as expected for a Fe-Cr-oxide. For example, dense 
Cr,0; has a hardness of 21GPa [3]. Also the E-modulus is lower than that in the steel. The 
plausible explanation is that the oxides contain a large amount of pores which decreases the 
hardness and E-modulus. In ref. [8] it is found that the hardness and E-modulus decrease 
with growing oxide, with the assumption that the increasing layer thickness leads to increased 
porosity. The outer oxide layer shows slightly increased hardness and further decreasing 
E-modulus. This effect might be based on different porosity and/or structure in the outer layer 
than in the inner layer. The SEM images clearly show different structures in the cross section. 
The inner layer looks “rougher” than the outer layer which could relate to a higher porosity.