Contrary to many railroaders’ and rail mills’ opinions, there are no innocuous inclusions in rail. Any simple oxide particle, any complex oxide particle, any sulfide particle, or any complex oxide-sulfide particle in a rail will initiate a crack in the rail if it happens to be located in a critically stressed section of rail.

Dr. M. Nabil Bassim at the University of Manitoba demonstrated this fact in several research studies (1, 2, 3), indicating that the mechanism for crack initiation for each type of inclusion in rail is different.

• Aluminum oxide particles generate a tensile residual stress field around each alumina particle. It takes less applied stress on the rail to initiate a crack in the pearlite surrounding the alumina particle.

• Conventional oxides and complex oxides are inherently brittle. They crack easily under minimal loading. The resultant crack will extend into the surrounding pearlite matrix.

• Sulfide particles have absolutely no bond with their surrounding pearlite matrix material. They contribute no "strength" to the pearlite when it is stressed. The action is analogous to cracking open a pea pod — the peas do not hold anything together. The shear crack resistance of any sulfide particle is equivalent to a void of the same size and shape.

Controlled laboratory test work on specially prepared rail steels by Drs. J. Kalousek, D.M. Fegredo, E.E. Laufer, and M.T. Shehata generated the conclusion that the total volume fraction of sulfide particulate in the steel was a detrimental factor for avoiding or minimizing rail head wear (4, 5, 6).

Small inclusions in a railhead’s near-surface layer will definitely be overload-compression shear-stressed. At least one major heavy-haul railroad considers the extent of surface corrugation damage (or spalling) in premium rails to be one of the most serious causes for rail damage and replacement. Since near-surface railhead material is continuously wheel-loaded, it is subject to unavoidable near-infinity shear strain. It is not hard to extrapolate laboratory test results to recognize that any non-metallic particle in that high-strain field will severely limit the pearlitic structure’s ability to continue plastic deformation without initiating a surface spall crack.

Current inclusion-related quality specifications for rail intended for heavy-axle-load service merely prohibit the presence of single, large ultrasonically detectable inclusions in the rail, and only partially limit the size and aggregation of inclusions that can be seen in macro-etched transverse section test surfaces.

Rail breaking dynamically under stress reacts consistently with predictions derived from modern fracture mechanics, so long as the perceived flaw is a single, isolated flaw. The analysis becomes murky whenever two or more inclusion flaws are in close proximity. There is no effective procedure for assessing the real enhanced flaw potential of a cluster or an aligned string of non-metallic inclusions in a rail.

All so-called undersize particles, regardless of their clustering or aggregation, are commonly considered to be benign at the time of rail acceptance. The fact that initiation of rail failure may be traced to a clustered inclusion site is of little help to a railroad arguing with a supplier mill that the inclusion cluster in the supplied rail initiated failure.

Four conclusions can be made about non-metallic inclusions in rail:

• There is no such thing as a truly benign inclusion in rail.

• The only reason any non-metallic inclusion will not initiate rail failure during its track service life is because it didn’t experience sufficient service stress pulsing to initiate a crack.

• Rail mills may have little effective control over the size and distribution of inclusions in their finished rail, but they can control and minimize the total volume fraction of inclusions in their rail.

• Complete elimination of non-metallic inclusions in rail will never completely prevent rail failure or surface corrugation damage in track service. However, further minimization of inclusions will slow the incidence of sudden fracture and corrugation damage in current production rail.

Sulfur and Hydrogen Content in Rail
Traditionally, excess hydrogen in finished rail generated a high probability that the rail would spontaneously form serious hydrogen "fish-eye" flaws in the rails’ head sections, causing sudden failure in track. Over the years, the problem of excess hydrogen was greatly reduced by slow cooling of finished rail in insulated cooling boxes. The hydrogen problem has since been virtually eliminated by vacuum degassing of every molten heat of rail steel.

Throughout this period, there has been some recognition of the presumed fact that sulfide particles in a finished rail can act as hydrogen sink traps, and sulfide particulate would tend to prevent the formation of fish-eye flaws. Many rail mills continue to use this rationale for arguing that a rail with less than 0.025% or possibly 0.035% sulfur is a rail with some potential for forming fish-eye flaws. This argument is not universally true. Rail mills with consistent vacuum treatment facilities can consistently produce rail with less than 1 p.p.m. hydrogen. The mills that produce low hydrogen content rails can safely limit their sulfur contents to the 0.005% - 0.015% range. Heavy-haul railroads using these ultra-low sulfur and consistently properly degassed rail steels have not experienced any hydrogen fish-eye-type failures.