Dear editor and reviewers:

Here is our supplementalfile that mainly consists of 4 parts, and wishes to help you of better understanding our research.

1. Surface hardness measurements

Methodology: each sample prepared for hardness measurements are cut into a plate with 40 mm long, 15 mm wide and thickness of 8 mm. Rockwell hardness tests were employed on the surface of each specimen after grinding through abrasive paper in turns of 400 and 800-grit. 16 points were uniformly selected for the hardness tests.

Results: Fig. S1 shows the hardness distribution of the experimental steels, from which the hardness of annealed sample is around 22 HRC and is much lower than that of quenched samples, this indicates that the ferrite is the main phase in MSSs of annealing state. The harnesses of quenched samples are all higher than 56 HRC, which implies the martensite is acquired by quenching. In addition, with the increase in austenitizing temperature, the hardness of MSSs firstly increased, and then declined.

Fig. S1 Surface hardness distribution of experimental steels (in HRC).

2. EDS-line scanning analysis of undissolved M23C6

Methodology:Fieldemission scanning electron microscopy (FE-SEM) of type LEO-1530 equipped with an Oxford Instruments energy dispersive spectrometer was applied to perform a line scan of one typical undissolved M23C6 carbide in 1030 sample.

Results: the image of undissolved carbide with line scanning results is shown in Fig. S2. It reveals a gradually increase ofCr content and decrease of Fe content from the matrix to the carbides, indicating the existence of Cr enrichment in the carbide area, and is matched with the mapping test in Fig. 3( see manuscript).

Fig. S2 SEM image with line scanning results of undissolved M23C6 in 1030 sample.

3. TEM information of undissolved M23C6

Methodology: Specimens for TEM examination were firstly grinded to a thickness of about 0.05mm with a diameter of 3mm, and then were electro-polished in a twin-jet machine in a solution of 4%perchloric acid and 95%alcohol at about -30℃.

Results: the TEM image of the undissolved M23C6 carbide in 1030 sample is shown in Fig. S2. The large round particle with 1μm in diameter (Fig. S2 (a)) is accordance to the M23C6 in Fig. 2 (c), the SADP of the corresponding phase also prove that the carbide is FCC structure and with a lattice parameter of 1.064nm, this is well matched with the results of XRD analysis, and indicates that the undissolved carbides in quenched sample are M23C6 carbides.

Fig. S3 TEM image of undissolved M23C6 carbides: (a) bright field of M23C6, and (b) SAED pattern of corresponding carbide.

4. The potentiodynamic polarization curves of each specimen

Methodology: the potentiodymic polarization measurements were performed as we described in our article.

Results: As we have noted in article, the potentiodynic polarization curves of each heat treated sample were based on at 3 parallel specimens. The results showed that for the samples that experienced same heat treatment methods, the pitting potential were slightly various and distributed among the scale of ±6~±40mV, and the scattering of corrosion potential (Ecorr) are all no more than 80mV, this indicates that the polarization test is normal and trustful.

Fig. S4 Potentiodynamic polarization curves of each parallel specimen experienced different heat treatment: (a) annealed sample, (b) 980 sample, (c) 1030 sample, (d) 1080 sample and (e) 1130 sample

Captions:

Fig.S1 Surface hardness distribution of experimental steels (in HRC).

Fig. S2 SEM image with line scanning results of undissolved M23C6 in 1030 sample.

Fig. S3 TEM image of undissolved M23C6 carbides: (a) bright field of M23C6, and (b) SAED pattern of corresponding carbide.

Fig. S4 Potentiodynamic polarization curves of each parallel specimen experienced different heat treatment: (a) annealed sample, (b) 980 sample, (c) 1030 sample, (d) 1080 sample and (e) 1130 sample.