In a sentence, thermal cycling, or normalizing, is the metallurgical technique of reducing visible grain size by repeated cycling of steel from near its critical temperature to ambient temperature. Several years ago I wondered how traditional Japanese smiths were able to produce such fine grain size even in cases when normalizing was not officially part of their workflow and process.
As I watched traditional Japanese swordsmiths at work and as my own forging techniques adapted to a better “feel” for the steel I began to realize that normalizing could be built right into the forging process.
Research indicates that extremely coarse grain size is found in steel that has been heated too hot or too long, and generally any time a modern steel, for example a file, is heated above its critical temperature it will have a coarser structure than it did out of the factory rolling mill. Experiments by other bladesmiths also indicate it is possible to recover a reasonably fine grain structure by careful and accurate normalizing prior to hardening a blade. Three times at slightly decreasing heats seems to be the optimal range for significant levels of improvement in grain structure.
This will not be a treatise on normalizing or grain structure, but simply an interesting observation in light of my current knowledge and techniques. Nor will it even suggest a replacement for the standard-advised and well-recommended triple normalization before quenching and tempering a blade.
Recently I have had the opportunity to work with some century-old shear steel which has shed some new light on my understanding of both what was historically practiced in the west before the widespread availability of cast steel and what might have been the case centuries ago when koto swords were being forged in Japan.
I conducted some initial cold break tests on some horse-drawn carriage leaf springs made from possibly pre-industrial steel in preparation for forging and was surprised to find the grain was very coarse, only slightly below the point that would be considered a completely ruined blade. All of the leaf springs I tested yielded similar results, indicating to me that the nuances of heat treating were not known or, at least, not applied in the production of spring steel for this type of end use a century ago. I will continue to test and observe as I come across more samples to see if the results are coherent with this theory.
The second interesting discovery occurred as I began to work with these same samples of steel, using my normal workflow and approach. I found that by the time I reached the end of the sunobe and hizukuri stages a break test revealed grain sizes that were well below the original test results, often into the size range of hardened factory files, an excellent grain for blades.
I realized that the traditional practices of forging at lower and lower temperatures towards the end of the process, and forging out into a black heat were serving as normalizing without requiring it as a discrete step. To further clarify, I will outline my current workflow and its built in thermal cycling methods.
All throughout the process, whether forging or forge welding, be very aware of the maximum temperature of each heat, working at the minimum temperature possible for each operation. Welding is highest, wakashinobe a bit lower, sunobe a bit lower, hizukuri a bit lower, finishing and planishing still lower, and final straightening in the black (but not brittle blue) or even cold. Within each operation the heat range can be slowly dropped as well, for example starting sunobe at bright orange and ending at orange, starting hizukuri at orange and working through red orange down to red.
Also known as normalizing when undertaken as a specific operation, in this case it is done as part of the forging process. Within each heat during hizukuri, and more so as the end approaches, the piece can be worked to a lower heat as shape gets closer to finished. Each heat is allowed to drop right into the black range as planishing, straightening, adjusting, and checking is done. At least three to four intentional heats like this is enough but making it a natural habit eliminates the need to count.
Annealing vs Normalizing
I don’t anneal simple carbon steel as it becomes too gummy and clogs hand files, but a good practice is to finish hizukuri with one very low red normalizing heat of the entire blade at once to catch any localized hardening from the water on the anvil. After air cooling ara-shiage can be started right away. For extra care on sensitive steels, normalizing can be carried out after ara-shiage in preparation for yaki-ire.
**The heating time has been edited out and some of the tang work is missing due to battery issues.
The blade shape is based on the Aizu Shintogo kata: islandblacksmith.ca/2014/04/aizu-shintogo-kunimitsu-tanto-kata/