X. OUYANG et al.: COMPOSITION DESIGN OF 316LN AUSTENITIC STAINLESS STEEL FOR LIQUID-HYDROGEN STORAGE ... 729–735 COMPOSITION DESIGN OF 316LN AUSTENITIC STAINLESS STEEL FOR LIQUID-HYDROGEN STORAGE BASED ON HIGH-FLUX PREPARATION OBLIKOV ANJE KEMIJSKE SESTA VE AUSTENITNEGA JEKLA VRSTE 316LN ZA SHRANJEV ANJE TEKO^EGA DU[IKA NA OSNOVI PRIPRA VE Z VISOKO U^INKOVITIM TALILOM Xin Ouyang 1,2 , Xuemin Wang 1* , Xinming Hu 2,3 , Mengnan Xing 2,3 , Chenxi Liu 2,3 , Zongxu Pang 2,3 1 Collaborative Innovation Center of Steel Technology, University of Science and Technology Beijing, Beijing 100083, PR China 2 State Key Laboratory of Metal Material for Marine Equipment and Application, Anshan, Liaoning 114009, PR China 3 Iron and Steel Research Institute of Ansteel Group, Anshan, Liaoning 114009, PR China Prejem rokopisa – received: 2024-04-24; sprejem za objavo – accepted for publication: 2024-09-06 doi:10.17222/mit.2024.1167 Designing multi-element alloy compositions to achieve target performance was the first step in the development of modern ma- terials, but the traditional trial-and-error experiment seriously restricted the development of new materials due to its low effi- ciency. In this investigation, the composition design of 316LN austenitic stainless steel for enhanced liquid-hydrogen storage us- ing multi-crucible synchronous metallurgy in a high-flux experiment was proposed. Sixteen groups of as cast austenitic stainless steel samples with different compositions were smelted using high-flux material preparation. Then, through the observation of their microstructures, the chemical composition that best matched the target performance was finally selected. The results show that high-throughput experiments can greatly improve the efficiency of composition design optimization of new stainless steel products. At the same time, this investigation also analyzed the elemental composition of -ferrite and the method for effec- tively controlling the -ferrite content. In 316LN stainless steel, Cr and Mo were easily enriched in ferrite grains or grain bound- aries, forming Cr and Mo enriched regions. This resulted in a gradient transition of Cr and Mo from the ferrite region to the aus- tenite region, forming galvanic corrosion. In this study, the distribution of Cr, Mo and other elements in 316LN stainless steel was studied by means of a metallographic microscope, electron probe microanalysis, transmission electron microscope and scanning electron microscope. In addition, the relationship between the ferrite content and chemical composition was explored. Finally, it was determined that high-temperature, long-term sensitization treatment is an effective method for controlling the -ferrite content. Keywords: austenitic stainless steel, multi-crucible synchronous metallurgy, -ferrite, sensitization heat treatment Oblikovanje kemijske sestave kovinske zlitine z ve~ legirnimi elelementi in ciljanimi lastnostmi je prva faza razvoja novih modernih materialov. Tradicionalni eksperimentalni pristop preizkus-napaka-nov preizkus zelo omejuje napreden razvoj novih materialov zaradi njegove majhne u~inkovitosti. V tem ~lanku avtorji opisujejo dizajn kemijske sestave nerjavnega austenitnega jekla vrste 316LN za hitro shranjevanje teko~ega du{ika. Pri tem so predlagali za izdelavo te vrste jekla uporabo sinhrone metalurgije ve~ talilnih loncev in uporabo visoko u~inkovitega talila. Vzorce {estnajstih skupin litega nerjavnega jekla z razli~no kemijsko sestavo so izdelali v prisotnosti zelo u~inkovitega talila. Nato so opazovali mikrostrukturo nastalih litin in na koncu izbrali kemijsko sestavo tiste, ki se je ujemala z najbolj `eleno mikrostrukturo in ciljanimi lastnostmi. Rezultati so pokazali, da izvedeni visoko produktivni preizkusi lahko mo~no izbolj{ajo u~inkovitost optimizacije dizajna novih izdelkov iz nerjavnega jekla. Isto~asno so avtorji raziskovali in analizirali povezavo med kemijsko sestavo in vsebnostjo -ferita. V nerjavnem jeklu 316LN se feritna kristalna zrna in kristalne meje enostavno (hitro) obogatijo s Cr in Mo. Posledica je nastanek galvanske korozije zaradi postopnega prehoda vsebnosti Cr in Mo s feritnih na austenitna podro~ja. V tem ~lanku so avtorji porazdelitev Cr, Mo in drugih elementov v izdelanih nerjavnih jeklih tipa 316LN {tudirali s pomo~jo svetlobnega mikroskopa, vrsti~nega in presevnega elektronskega mikroskopa, elektronskega mikroanalizatorja, in eksperimentalno ena~bo za odvisnost med kemijsko sestavo in vsebnostjo delta ferita. Ugotovili so {e, da je visoko temperaturna in dolgotrajna senzibilizacijska toplotna obdelava u~inkovita metoda za kontrolo vsebnosti -ferita. Klju~ne besede: austenitno nerjavno jeklo; sinhrona metalurgija v ve~ talilnih loncih; -ferit; senzibilizacija toplotne obdelave 1 INTRODUCTION When developing new materials with specific target properties, the traditional material research used to adopt the empirical trial-and-error method, and explore the best performance of materials by adjusting the sample com- position and production process discretely through a small number of experiments. However, this research and development method, which relied heavily on a large number of pilot experiments, was time-consuming and expensive. At the same time, due to a small amount of experimental data, the investigation on the relationship between material composition and performance was not comprehensive and thorough. With the rise of material genome engineering, 1–3 the use of high-throughput ex- periments, computer performance prediction and other methods to accelerate the development of new materials attracted more and more attention. The high-throughput Materiali in tehnologije / Materials and technology 58 (2024) 6, 729–735 729 UDK 669.14.018.8:57.012.3 ISSN 1580-2949 Original scientific article/Izvirni znanstveni ~lanek Mater. Tehnol. *Corresponding author's e-mail: wxm@mater.ustb.edu.cn (Xuemin Wang) material preparation method significantly improves the efficiency of material preparation due to its multi-station, automated and parallel characteristics. It can allow a rapid and large-scale synthesis of materials. Compared with the traditional single-sample production mode, its preparation efficiency is significantly improved. 4–5 With the accumulation of experimental data, material database plays an increasingly important role in the process of new material research and design. By screening the data- base and finding candidate materials closest to the target performance requirements, high-throughput experiments have gradually become a powerful method for obtaining large amounts of high-quality and self-consistent experi- mental data. 2 EXPERIMENTAL PART High-flux smelting method for controlling chemical composition is mainly used for the preparation of homo- geneous as cast samples with complex alloy composi- tions, including additive manufacturing, multi-mode casting, multi-crucible synchronous metallurgy, etc. In this study, multi-crucible synchronous metallurgy was used to prepare as cast samples. 6,7 The test process in- cluded alloy sample batching, arc melting, grinding and polishing, microstructure observation, rolling, heat treat- ment and mechanical property inspection. The high-flux alloy preparation system adopted had a 16-station auto- matic arc melting furnace, as shown in Figure 1. For the alloy melting process of this experiment, we adopted a X. OUYANG et al.: COMPOSITION DESIGN OF 316LN AUSTENITIC STAINLESS STEEL FOR LIQUID-HYDROGEN STORAGE ... 730 Materiali in tehnologije / Materials and technology 58 (2024) 6, 729–735 Table 1: Chemical composition of a steel ingot prepared in the second round of high-flux preparation (in mass fractions (w/%)) Com- posi- tion range C 0.04 - 0.08 Si = 0.75 Mn 1.00 - 2.00 P = 0.03 S = 0.015 Cr 16.00 - 18.00 Ni 11.00 - 13.00 Mo 2.00 - 3.00 Test plan B1 0.04 0.4 1.6 0.015 0.005 18 11 2 Scheme 1: investigate the ef- fect of Cr on ferrite content when the content of austen- ite-forming elements is the lowest B2 0.04 0.4 1.6 0.015 0.005 17 11 2 Scheme 2: investigate the ef- fect of Mo on ferrite content when the content of austen- ite-forming elements is the lowest B3 0.04 0.4 1.6 0.015 0.005 16 11 2 B4 0.04 0.4 1.6 0.015 0.005 16 11 3 B5 0.06 0.4 1.6 0.015 0.005 18 11 2 Scheme 3: investigate the ef- fect of Cr on ferrite content at high C B6 0.06 0.4 1.6 0.015 0.005 17 11 2 B7 0.06 0.4 1.6 0.015 0.005 16 11 2 Scheme 4: investigate the ef- fect of Mo on ferrite content at high C B8 0.06 0.4 1.6 0.015 0.005 16 11 3 B9 0.04 0.4 1.6 0.015 0.005 18 13 2 Scheme 5: in the case of high Ni, the effect of Cr on ferrite content is investigated B10 0.04 0.4 1.6 0.015 0.005 17 13 2 B11 0.04 0.4 1.6 0.015 0.005 16 13 2 Scheme 6: investigate the ef- fect of Mo on ferrite content in the case of high Ni B12 0.04 0.4 1.6 0.015 0.005 16 13 3 B13 0.06 0.4 1.6 0.015 0.005 18 13 2 Scheme 7: investigate the ef- fect of Cr on ferrite content when the content of austen- ite-forming elements is the highest B14 0.06 0.4 1.6 0.015 0.005 17 13 2 B15 0.06 0.4 1.6 0.015 0.005 16 13 2 Scheme 8: investigate the ef- fect of Mo on ferrite content when the content of austen- ite-forming elements is the highest B16 0.06 0.4 1.6 0.015 0.005 16 13 3 Figure 1: High-flux material multi-crucible synchronous smelting system: a) high-purity metal block raw materials; b) high-throughput preparation system; c) control interface; d) multi-crucible synchronous metallurgical platform; e) as cast metal ingots with different composi- tions high-throughput alloy preparation system, which realized a multi-station, automated and efficient operation in all stages of material preparation, thus significantly reduc- ing the average experimental preparation and operation time needed for a sample. For example, the system avoided the cumbersome steps of multiple vacuum pumping of the traditional arc melting method, and suc- cessfully overcame the bottleneck of time-consuming material preparation. Compared with the traditional sin- gle- or multi-sample preparation methods, the overall ef- ficiency of the high-throughput alloy preparation system is increased by at least 10 times, showing significant ad- vantages in the field of material preparation. 3 RESULTS According to the literature, 8–9 the ratio of -austenite to -ferrite was comprehensively affected by -austenite forming elements (such as Ni, Mn, N, etc.) and -ferrite forming elements (such as Cr, Mo, Si, etc.). The forma- tion of -ferrite was mainly sensitive to the segregation of C, Ni, Cr and Mo. Carbon easily forms carbides with chromium, and the element segregation caused by car- bides promotes the formation of -ferrite. Nickel inhibits the precipitation of -ferrite by improving the stability of austenite and reducing the martensitic transformation temperature. When the molybdenum content is high, it increases the tendency of brittle phases such as -ferrite to precipitate. In this study, the contents of sensitive ele- ments were adjusted through 16 groups of high-through- put tests to find the target chemical composition with the lowest ferrite content. The test scheme is shown in Ta- ble 1, and the ferrite measurement results are shown in Figure 2. Results showed that the ferrite content in the metallographic structures of B2, B3 and B7 test ingots was low. The three groups of ingots were heated, rolled and solution heat treated. The ferrite content of the hot-rolled test plates was measured. The observation re- sults for the ferrite content are shown in Figure 3. The common feature of the three components was the fact that the contents of Ni, Cr and Mo were controlled ac- cording to the middle and lower limits. From the per- spective of meeting the mechanical properties, it was X. OUYANG et al.: COMPOSITION DESIGN OF 316LN AUSTENITIC STAINLESS STEEL FOR LIQUID-HYDROGEN STORAGE ... Materiali in tehnologije / Materials and technology 58 (2024) 6, 729–735 731 Figure 2: Metallographic photograph of ferrite content of the steel ingot prepared during the second round of high-flux processing recommended to select the chemical composition of the B7 test plate with a high carbon content as the target component. However, with the increase in the carbon content, the ferrite content of the test plate also in- creased. Other means must be used to optimize the alloy elements so that they achieve a good match between the mechanical properties and microstructure. 4 DISCUSSION In this study, the stability of ferrite was destroyed with chemical methods to reduce the content of ferrite in austenitic stainless steel. 10 Large ferrite was divided into the granular and unstable state. Due to the particularity of 316LN stainless steel, it can be sensitized at a higher temperature for a long time (about 850 °C). According to the composition characteristics of carbides and ferrites, long-term heat preservation in a high-temperature envi- ronment accelerated the precipitation of chromium-en- riched ferrites, resulting in M 23 C 6 distributed at the aus- tenite grain boundaries as discontinuous small particles. In addition, carbon atoms quickly dissolve back into aus- tenite at high temperature, so it was easy to destroy the stable structure of original ferrite during subsequent high-temperature annealing or solid solution treatment. 11 In this process, the large ferrite was cut into a lot of gran- ular small pieces that become very unstable due to the re-dissolution of carbon atoms at subsequent high tem- peratures, which was conducive to the rapid dissolution of ferrite. Figure 4 shows the microstructure of sample B7 after the sensitization treatment at 850 °C for 4 h. It can be clearly seen from Figures 4a) and 4b that the large strip ferrite in the center after sensitization was divided into granular carbides. M 23 C 6 carbide belongs to the face-cen- tered-cubic (FCC) structure, so its diffraction pattern shows the characteristics of this structure. According to the observation results for the transmission structure and diffraction spots, the diffraction ring or spot group of carbides precipitated at the grain boundaries was cen- tered around the optical axis of the transmission electron microscope, showing a circular or nearly circular distri- bution; diffraction spots showed a central symmetrical pattern, proving that the carbide was mainly composed of M 23 C 6 . 12 Figure 5 shows the ferrite content and morphology of sample B7 before and after sensitization and solid so- lution. It shows that the microstructure etched with po- tassium hydroxide electrolysis does not only clearly show the outline of grain boundary, but also the residual -ferrite. On Figure 5a, we can clearly see that the color of the -ferrite region is black, which is due to the signif- icant damage of the passive film in this region. There were two main reasons for this failure: first, because of its low Perner force and stacking fault energy, austenite experienced slip of edge dislocation relatively easily, X. OUYANG et al.: COMPOSITION DESIGN OF 316LN AUSTENITIC STAINLESS STEEL FOR LIQUID-HYDROGEN STORAGE ... 732 Materiali in tehnologije / Materials and technology 58 (2024) 6, 729–735 Figure 3: Measurement results for the ferrite content of hot-rolled test plates while -ferrite was more prone to cross-slip of screw dis- locations due to its higher stacking fault energy. This led to the inhomogeneity of the passive film on the surface of the material, and then affected the stability of the pas- sive film so that the dissolution rate of the passive film in the -ferrite region was much higher than that of austen- ite. 13 Secondly, during the hot rolling process, the flat- tened -ferrite concentrated a large number of high-den- sity dislocations, making its energy higher than that of the austenite matrix. This led to a difference in the elec- trochemical properties between the austenite matrix and the residual -ferrite, forming a corrosion micro cell, in which -ferrite was used as the anode and the austenite matrix was used as the cathode; as a result, -ferrite was more vulnerable to corrosion, resulting in a black ap- pearance. 14 It can be seen from Figure 5 that the ferrite content decreases by about 50 % after the sensitization and solid solution processes, meeting the requirements of 3% , but there is still a gap to achieve a reduction to 1% . The main reason for this is the fact that the sensitization treatment does not only destroy ferrite, but also promotes the formation of carbides. During the sensitization pro- cess, the formation of carbides was accompanied by a lo- cal high enrichment of Cr and Mo, resulting in the segre- gation of alloy elements in the matrix. During the subsequent solution heat treatment, these previously formed carbides disintegrated, and the elements in them were re-dissolved into the matrix. However, due to the local high concentration of Cr and Mo in the sensitiza- tion stage, their diffusion dissolution during the solid so- lution heat treatment was limited. Figure 6 shows the microstructure of sample B7 after sensitization treatment and solid solution treatment at 1060 °C. It can be seen from Figure 6a that carbides were eliminated after the solution treatment, but there was still a certain amount of carbides in the local -fer- rite areas. As shown in Figure 6b, after corrosion, it was obvious that a / grain boundary with a width of 1.5–2 μm formed a corrosion-dissolved structure. The corro- sion resistance of the -ferrite, austenite and two-phase interface was different, leading to the formation of gal- X. OUYANG et al.: COMPOSITION DESIGN OF 316LN AUSTENITIC STAINLESS STEEL FOR LIQUID-HYDROGEN STORAGE ... Materiali in tehnologije / Materials and technology 58 (2024) 6, 729–735 733 Figure 5: Ferrite morphology and content of 316LN stainless steel be- fore and after sensitization: a) rolled state; b) 850 °C×4hsensitized state; c) 1060 °C×1hsolid solution Figure 4: Sensitized microstructure of 316LN stainless steel: a) metallographic photos of macromorphology; b) partially enlarged scanning elec- tron microscope photograph; c) TEM morphology and diffraction pattern vanic corrosion. The content of alloying elements in the -ferrite structure was high, and the electrode potential was high. The content of alloying elements in the austen- ite structure was low, and the electrode potential was high. Due to the unequal electrode potentials of metals, a corrosion battery was formed, causing the galvanic cur- rent to flow from the high potential area to the low po- tential area, thus increasing the metal dissolution rate at the two-phase interface, and resulting in local corro - sion. 15 In order to explain the cause of galvanic corrosion, the sample was electropolished, as shown in Figure 6c, to retain ferrite and all the surrounding structures. Then the compositions of the polished original structure and the corroded structure were analyzed, as shown in Ta- ble 2. During electrolytic corrosion, the electrode metal rod with a high potential was used as the anode, and its dissolution rate was faster. Stainless steel was used as the cathode, and its dissolution rate was slower, while the surface was protected by the anode. 16,17 In addition, the degree of galvanic corrosion was reduced by adjusting the area ratio of anode and cathode and reducing the di- electric conductivity with a 10-% KOH aqueous solution so that the structure of the transition region of the two-phase interface could be retained. It can be seen from the table that in the microstructure of metallo- graphic corrosion, the -ferrite region contained high amounts of Cr and Mo, with maximums of 20.59 w/% and 5.18 w/%, respectively. This was due to the ferrite with high Cr and Mo amounts. These amounts of Cr and Mo gradually decreased from the center of ferrite to the austenite zone, and the amounts of Cr and Mo in the aus- tenite around ferrite were about 18 % and 2.9 %, respec- tively. The amount of Mo was also higher than that in typical 316LN, indicating that the composition around -ferrite had reached an equilibrium state. Table 2: Analysis of energy spectra from Figure 6 (in mass fractions (w/%)) No Cr Mo Ni Si Mn 1 20.59 5.18 9.35 0.56 1.77 2 17.81 2.90 12.23 0.80 1.73 3 20.25 4.27 9.42 0.70 2.17 4 16.38 2.87 12.47 0.62 2.43 5 20.36 4.78 9.48 0.57 1.89 6 17.64 2.93 12.12 0.57 2.56 7 17.75 3.84 10.13 0.47 1.59 8 20.67 4.58 9.55 0.51 2.41 9 17.10 2.24 12.92 0.53 1.82 5 CONCLUSIONS 1. In this study, the composition optimization design of 316LN austenitic stainless steel for liquid-hydrogen storage and transportation equipment was advanced with high-throughput experiments, and comprehensive design X. OUYANG et al.: COMPOSITION DESIGN OF 316LN AUSTENITIC STAINLESS STEEL FOR LIQUID-HYDROGEN STORAGE ... 734 Materiali in tehnologije / Materials and technology 58 (2024) 6, 729–735 Figure 6: Microstructure of 316LN stainless steel after sensitization and solid solution: a) metallographic structure; b) metallographic structure (EPMA); c) electropolished state (EPMA) efficiency was increased by more than 20 times. Stain- less steel ingots differing by a single component were prepared in large quantities using the multi-crucible syn- chronous metallurgy method, which was at least 10 times faster than the traditional single/few sample preparation. This work proved the feasibility of doubling the speed of material research and development with low cost pro- posed by the Materials Genome Initiative, where the high-throughput experimental method can become an ef- fective general strategy to accelerate the optimization de- sign of material composition. 2. Reducing the Creq/Nieq ratio within the composi- tion range, especially controlling the amounts of Cr and Mo, allowed us to fundamentally reduce the formation of ferrite and control its content. When the composition was fixed, the ferrite content had to be controlled from two aspects: sensitization treatment and solution treat- ment. 3. Under existing process conditions, the method of high-temperature sensitization and solid solution treat- ment was more effective in reducing ferrite, meeting the requirement of -ferrite=3% .H o w e v e r ,i tcould not achieve -ferrite=1%o rcompletely eliminate it. A large number of carbides were generated during the sen- sitization heat treatment; however, although the carbides could disintegrate ferrite at the same time, the enriched Cr and Mo were difficult to diffuse during the subse- quent solid solution heat treatment. As a result, a smooth transition of the concentration gradients of Cr and Mo was easily formed, thus causing galvanic corrosion. Acknowledgment The authors acknowledge the support of the Funda- mental Research Business Funds for Central Univer- sities (FRF-BD-22-02). 6 REFERENCES 1 Z. Wang, M. Wang, Q. L. Yong, et al., Materials Informatics and its Application in Materials Research, Materials China, 36 (2017)2 , 132–140, doi:10.7502/j.issn.1674-3962.2017.02.08 2 Y. Z. Lv, Z. X. Qin, X. Lu, Current Research Status of High- Throughput Development of Amorphous Alloys, Materials Reports A: Review Papers, 31 (2017) 9, 112, doi:10.11896/j.issn.1005- 023X.2017.017.016 3 Y . H. Liu, Z. H. Hu, Z. G. Su, et al., Current Status of High Through- put Experimental Technology and its Application in Research of Pre- cious Metals, Science China Technological Sciences, 62 (2019)4 , 521–545, doi:10.3969/j.issn.1004-0676.2018.z1.014 4 W. Jiong, et al., Machine learning assisted high-throughput experi- ments to accelerate hard high entropy alloys, Progress in Materials in China, 39 (2020) 4, 269–277, doi:CNKI:SUN:XJKB.0.2020-04-004 5 X. Wen, F. Zhang, et al., Second Phase Strengthening in High-En- tropy Alloys, Materials China, 38 (2019) 3, 53, doi:10.7502/ j.issn.1674-3962.2019.03.06 6 M. Bao, J. Qiao, Research Progress of Hexagonal Close-Packed High-Entropy Alloys, Materials China, 397 (2018) 4, 314–319, doi:10.7502/j.issn.1674-3962.2018.04.03 7 C. Li, Y . Hou, et al., Diffusion Dynamic Analysis on Selective Laser Melting Process of Fe/Ni Powder, Materials Review, 34 (2020) Z1, 117–130 8 J. Li, Y . Zhuang, W. Yin, Formation and control of ferrite in 316H stainless steel, Steel, 57 (2022) 11, 123–130, doi:10.13228/j.boyuan. issn0449-749x.20220293 9 S. Aras, R. Ertan, Studying the strength of dissimilar joints of AISI 430 and 301 stainless steel welded at different welding parameters, Mater. Tehnol., 58 (2024) 2, 123–130, doi:10.17222/ mit.2023.1054 10 Z. Pang, et al., Residual in S31609 austenitic stainless steel corro- sion behavior of ferrite, China Metallurgical, 31 (2021) 3, 59–65, doi:10.13228/j.boyuan.issn1006-9356.20200451 11 Y . S. Sato, H. Kokawa, Preferential precipitation site of sigma phase in duplex stainless steel weld metal, Scripta Materialia, 40 (1999)6 , 659, doi:10.1016/S1359-6462(98)00483-7 12 X. Huangfu, X. Wang, W. Wang, Study on characteristics of residual ferrite in solidification structure of austenitic stainless steel, Iron and Steel, 47 (2012) 4, 74, doi:10.13228/j.boyuan.issn0449-749x.2012. 04.006 13 H. Dong, Comparative study on microstructure evolution of austenite and ferrite during cyclic deformation, Dalian: Dalian University of Technology, 2020, doi:10.26991/d.cnki.gdllu.2019.001550 14 J. Wang, Calculation of electronic work function and its application in the study of electrochemical problems on the surface of materials, Hefei: University of Science and Technology of China, 2016 15 Z. Peng, L. Cai, F. Peng, et al., P92 steel effect of ferrite on the composition and volume fraction of precipitates aged at 700 and 750 °C, Acta Metalica Sinica, 48 (2012) 11, 37, doi:CNKI:SUN:JSXB.0. 2012-11-007 16 B. Schlüter, G. Barkleit, F. Schneider, et al., Analysis of intergranular corrosion attack of stainless steels by means of atomic force micros- copy and optical microscopy. Part2: Influence of impurity content, Mater. Corros., 51 (2015) 2, 115, doi:10.1002/(SICI)1521- 4176(200002)51:2<115::AID-MACO115>3.0.CO;2-N 17 G. Barkleit, F. Schneider, et al., Analysis of intergranular corrosion attack of stainless steels by means of atomic force microscopy and optical microscopy. Part 1: Influence of treatment, Mater. Corros., 51 (2000) 2, 109, doi:10.1002/(SICI)1521-4176(200002)51:2<109: :AID-MACO109>3.0.CO;2-B X. OUYANG et al.: COMPOSITION DESIGN OF 316LN AUSTENITIC STAINLESS STEEL FOR LIQUID-HYDROGEN STORAGE ... Materiali in tehnologije / Materials and technology 58 (2024) 6, 729–735 735