Microstructures
by George Langford, Sc.D., Massachusetts Institute of Technology, Cambridge, MA, 1966
Copyright©
2005 by George Langford
Low Alloy Steels - Lesson 1 - Fifth specimen
AISI 1045 steel at 500X etched
The arrangement of the soft, ductile ferrite in this 1045 steel led to a fatigue failure.  The specimen was used in the as-hot-rolled condition.  It is shown here at 500X; Nital etch.

Even though the ferrite is the minority microconstituent, it forms a continuous envelope around the prior austenite grains.  The faitigue crack (not shown here) followed this ferrite network because of the lower shear strength compared to the pearlite.  This piece wasn't any better than a completely ferritic steel.

What would you do to make things right ?

























Remedy: rearrange the ferrite and cementite by quenching and tempering the steel to obtain a uniform distribution of spheroidal cementite particles in a fine grained ferrite.  This is also called tempered martensite. 

However, many makers of fatigue sensitive parts follow a different tack.  They use annealed alloy steel.  The reason is that most steels have a fatigue limit, the stress below which no cracks ever develop. In annealed alloy steel, the fatigue limit is higher than the yield strength.  The part will therefore show signs of yielding when it has been exposed to stresses that could lead to fatigue failure.  A case in point is a lifting swivel, a ball bearing device used with a crane that permits the load to be rotated without twisting the cable.  The one that I examined was originally made with very little clearance between the upper and lower portions.  An inspector could verify that the swivel was safe to use by trying to insert his business card between the two halves.  If it would not go in, the swivel had never been overloaded and was therefore free of fatigue.  This avoided the expensive task of taking the swivel all apart in order to check its central shaft for cracks.  The lack of any quenching and tempering treatment on this sensitive part meant that it would also never be likely to contain any cracks resulting from the transformation stresses involved in forming martensite.  This approach isn't universal because the part has to be proportionately larger in its load-bearing cross sections to compensate for the low strength of the annealed material.