16  Prakt. Met. Sonderband 50 (2016) 
2 Corrosive Failures — Intergranular Attack in Retaining Rings cannot be co 
always be sor 
A number of retaining rings used in fuel gas piping assemblies of large stationary heavy-duty ih: 
industrial gas turbines were found broken upon reception at the gas turbine manufacturer's plant. From this po] 
While the pipes were made of AISI 321 austenitic stainless steel, 1.4541, X6CrNiTil8-10, the out of specifi 
specified material of the subject locking rings is the martensitic stainless chromium steel Springs Or Sır 
X39CrMo17-1, 1.4122. In fact, however, the failed snap rings consisted of the lower-chromium resistance ar 
X39Cr13, 1.4031. The fractured retaining rings all originated from a particular North American polishability 
supply chain. In contrast, the same pipe assemblies delivered by an alternative second-source particularly h 
European supplier did not show these remarkable fractures in their retaining rings. Also, they were environments 
indeed made of the specified higher-chromium material. It was later discovered that the supplier any case, 1.4 
that shipped the faulty product employed a peculiar wet cleaning process without proper post- Inconosions 
cleaning drying. The conclusion of the metallurgical failure analysis was that the subject retaining | 
rings failed by intergranular corrosion due to sensitisation from heat treatment and wet cleaning What 18 MO 
residues that remained on the product and caused corrosion during shipment. The erroneous protection In 
material selection, violating the specification, contributed to the failure [6]. or wet clean 
material mix: 
The subject locking rings exhibit clear evidence of intergranular attack. A chlorine peak was found from wet cle 
in the energy dispersive X-ray spectroscopy (EDX) spectrum of a spot analysis within the corrosion to Intergr ath 
product on the fracture surface. It is remarkable that there is no chromium peak in this spectrum. It austenitic Sta 
appears the corrosion product consists mainly of iron oxide, i.e. there must have been a rather 
severe corrosive environment experienced by the subject retaining rings that prevented the usual The metallur 
formation of a protective chromia (Cr,0;) layer. Evidence of intergranular attack was verified with am 
scanning electron micrographs. could have bı 
The sensitisation by chromium depletion of the immediate vicinity of grain boundaries, resulting Also a contri 
from secondary chromium carbide precipitation at grain boundaries, is so severe and excessive in . 
the case at hand, that the relatively poor lateral resolution of an energy dispersive X-ray Shy 
spectroscopy (EDX) map was not insufficient to show this precipitation of comparatively large A 
chromium carbides, probably Cr,;Cs. The exact nature of these carbides was not verified by STEM- 
EDX (scanning transmission electron microscopy) or diffraction-based phase identification somewhat ag 
techniques, because it was unnecessary for the determination of the root cause of failure. Sven 14122 
in any case i 
Sensitisation results from the migration at elevated temperatures of carbon and chromium to Arguably, th 
precipitation sites at grain boundaries, were there is more inter-lattice space for these rather large pan f 
carbide phases to precipitate. There, carbon and chromium combine to form said carbides. This they desi 
readily do due to their high affinity to each other. Selgn one, 
particular ap 
In many cases of industrial heat treating practise, the elevated temperatures required for carbide Se 
precipitation, i.e. the secondary chromium carbide precipitation range between ca. 450°C and heat treatme 
850°C, is provided either by the tempering heat treatment after quenching, when the final required chromium € 
strength of components is adjusted, or even by slow cooling from hardening temperature, if the boundaries 
cooling rate is not high enough and the precipitation range is passed too slowly, or, finally, by imme diately 
insufficiently fast heating to hardening temperature. This means that even if the tempering corrosion re: 
temperature of alloy 1.4122 is well below 600°C in order to retain a high enough strength level, the alloy and is 
carbide precipitation range might still be reached. In other words, in both the specified alloy 1.4122, might fail br 
and the actually selected 1.4031. secondary chromium carbide precipitation at grain boundaries ;