In order to suppress heat treatment distortion of quenched parts of industrial machinery, oil quenching is often applied instead of water quenching. If heat treatment distortion can be suppressed even by water quenching and applied to industrial machinery parts, further high strength and a carbon-neutral manufacturing environment can be realized by rapid cooling. Therefore, quenching experiments were performed on carburizing members by water cooling and oil cooling, respectively, and the state of martensite formation and heat treatment distortion was investigated. In oil cooling, martensite formation proceeds over several minutes, and during that time, the temperature distribution of the processed member is also equalized, forming a relatively uniform martensite shell, and it was confirmed that the heat treatment deformation was reduced. In water cooling, martensitic formation is reached in tens of seconds, and there is no time to equalize the heat of the processed member, and uneven physical parameters such as relative speed cause large heat treatment deformation. On the other hand, it was clarified that the temperature distribution was actively controlled using physical parameters such as relative velocity for more than 10 seconds as a control factor, and that there was a possibility of controlling heat treatment deformation at the same time as martensite formation.

For the purpose of grain size refinement, the Gange’s method executing repeated transformation was applied to the commercial SCM440 of a diameter of 19mm with a length of 4000mm workpiece using the continuous heating cooling system of the actual manufacturing equipment. As a result, the prior austenite grain size was refined from grain size number 7.5 as received workpiece to grain size number 11.5. Therefore,　the mechanical properties of thus prior austenite grain refined samples and the conventional heat treated product were investigated.　The balance of the strength as 0.2％ proof stress and the toughness as room temperature impact absorbed energy of the prior austenite grain refined samples was improved compared with those of the conventional heat treatment product and the ductile-brittle transition temperature was also lowered from form －75℃ to －115℃.

In steels with low Ms temperature, austenite cannot transform to martensite completely and untransformed austenite retains within martensite matrix. Since the value of saturation magnetization（SM）is very sensitive against the existence of non-magnetic phases, the measurement of SM was applied on the quantitative evaluation of small amount of austenite that has retained within martensite matrix. The SM-method was also applied to 90％ cold rolled pure iron and it was confirmed that dislocation does not give any influence on the SM values. During tempering of martensite containing carbon, complicated carbide reactions occur but no change was found on the SM values of tempered martensite. As a result, in steels with bcc structure, the SM value J^* is not dependent of microstructure but depends on only the amount of alloying element X（mass％）in steels, as expressed by the following equation:

J^* [T] ＝2.153－0.118C－0.064Si－0.030 (Mn＋Cr) －0.010Ni－0.020Mo－0.024Cu－0.072Al

The volume fraction of retained austenite V（％）is obtained by the following equation as a function of the measured SM value J.

V（％）＝100×（1－J/J^*）

Abnormal grain growth during carburizing of steel can cause problems such as increased heat treatment distortion and reduced fatigue strength. Since grain pinning using precipitates is effective for suppressing abnormal grain growth, determination of the particle size distribution of the precipitates is vital. Indirect methods such as the small-angle scattering method and direct methods such as transmission electron microscope（TEM）observation are used for analysis. However, the former has problems with accuracy in size determination, and the latter with ensuring representativeness. In this study, we investigated an evaluation method using a scanning electron microscope（SEM）, which can observe a wider area than a TEM. By using an ultra-low accelerating voltage SEM at a low accelerating voltage and selectively collecting low-energy secondary electrons, AlN can be distinguished from the matrix by its bright contrast. We applied this method to gears of two types of gear steel with different former γ grain sizes and determined the histogram of the AlN size in each gear.