+Advanced Search
网刊加载中。。。

使用Chrome浏览器效果最佳,继续浏览,你可能不会看到最佳的展示效果,

确定继续浏览么?

复制成功,请在其他浏览器进行阅读

Effect of Recrystallization of Steel on Stability of Interme-tallic Phases in Molten Al  PDF

  • Zhang Jiarui 1
  • Yu Zhongdi 2
  • Wu Jinping 2
1. School of Materials Science and Engineering, Northeastern University, Shenyang 110819, China; 2. Xi'an Rare Metal Materials Institute Co., Ltd, Xi 'an 710016, China

Updated:2023-12-25

  • Full Text
  • Figs & Tabs
  • References
  • Authors
  • About
CN CITE
OUTLINE

Abstract

The influence of steel's recrystallization on the stability of intermetallic compounds (IMCS) layer was investigated in molten Al at 700 °C. Results show that high degree of structural rearrangement leads to the formation of non-protective IMCS layer on the 304SS surface and its corrosion kinetics coincides with linear law. Compact and stable Fe2Al5 layer delays the failure of the 410SS in liquid Al. The stability of IMCS is affected by steel's recrystallization in three aspects: (1) energy for IMCS nucleation and growth are reduced, especially for Fe4Al13; (2) grain orientation transforms to close-packed direction which slows down the diffusion rate of Al ion; (3) stress mismatch is increased at the IMCS/steel interface, especially for the austenitic 304SS.

Hot-dipping aluminum is an economic method for boilers and exhaust steel pipes protection in high-temperature environment[

1–3]. By forming continuous alumina layer, oxidation of the matrix is prevented. However, thick intermetallic compounds (IMCS) formed at Al/steel interface threaten the stability of the coating, especially under thermal shock case. Quite different thermal expansion of the IMCS and steel increases interfacial stress concentration, promoting microcracks initiation within this layer[4–6]. As the reaction of steel with liquid-Al is thermodynamically spontaneous, the research on aluminized steel degradation by IMCS formation is meaningful to guide its production [7].

Over the last decades, the evolution of IMCS on aluminized mild steel has been researched widely. Double-layer structure of the inner tongue-like Fe2Al5 (η phase) and outer flocculent Fe4Al13 (θ phase) are observed in ferrite steel after dipping in molten Al[

8–11]. The growth of this layer is affected by outer θ phase dissolution and element interdiffusion in the inner η phase. This model is suitable for most low-alloying steels with bcc structure, while some abnormal results are reported for the austenitic or martensitic steel in recent years[12–15]. Chen et al[16] studied the effect of austenitic process on the ST12 steel in liquid aluminum. After austenitizing treatment, tongue-like Fe2Al5/α-Fe transforms to flat Fe2Al5/γ-Fe interface. Similarly, Li et al[17] observed a flat IMCS layer on the austenitic SAE201 stainless steel surface, which is different from the rough products on the mild steel. Obviously, microstructure difference of steel will significantly affect the structure of IMCS layer.

In recent years, a lot of stainless steels have been used for aluminized plating to meet growing high-temperature serving demands. However, inevitable recrystallization occurs in austenitic or martensitic steel during hot-dipping process[

18–19]. This matrix structural transition can release the stored energy generated in the deformation process, affecting interfacial Fe-Al phase nucleation and growth. Since few of researches report this structural rearrangement issue, it is urgent to clarify the influence of recrystallization on interfacial IMCS layer formation.

In this study, two commercial stainless steels were researched, including martensitic 410SS and austenitic 304SS. These commercial stainless steels are widely used in corrosive and high temperature environment. To simulate hot-dipping condition, these steels were immerged in pure aluminum at 700 °C. This temperature is high enough for these steels to recrystallize. The evolution of interfacial Fe-Al phase was also researched.

1 Experiment

1.1 Annealing experiment

The composition of austenitic stainless steel (304SS) and martensitic stainless steel (410SS) is shown in Table 1. All the annealed samples were prepared with the dimensions of 12 mm×10 mm×2 mm. In order to reveal the structure evolution, these samples were heated to 700 °C for 3, 5, 10, 30 and 60 min, and then air cooled.

Table1  Composition of 304SS and 410SS samples (wt%)
SampleFeCrNiMn
304SS 72 18 8 2
410SS 86.5 13 0 0.5

1.2 Dipping experiment

For hot-dipping Al test, steel samples were polished with 800# metallographic sandpaper to remove surface impurities and oxides. Then, original size and mass of samples were measured.

During each test, 50 g industrial pure aluminum (99.9%) and five parallel samples were set in corundum crucible. Then, they were heated to 700 °C in high temperature furnace. After test, these dipped samples were taken out and air cooled.

1.3 Characterization

In order to calculate the mass change of these samples, dipped samples were immersed in 20wt% NaOH solution to remove adhered aluminum. Then, they were cleaned, dried and weighed. These samples for cross-sectional observation were polished with adhered Al.

The Rockwell hardness tester and optical microscope were used to measure the hardness and macrostructure evolution after annealing. Scanning electron microscope (SEM) was used to observe the micromorphology of the corrosion products. X-ray diffraction analysis (XRD) was used for phase identification. Microhardness and Young's elastic modulus of the IMCS layer and matrix were researched by nanoinden-tation technology. The EBSD technique was taken to detect the grain orientation of the alloy matrix.

2 Results and Discussion

2.1 Recrystallization texture

Fig.1 shows the microhardness and microstructure of the 304SS and 410SS after annealing at 700 °C for different time. After heat treatment for 60 min, the microhardness of the 304SS reduces sharply from 228×9.8 MPa to 180×9.8 MPa, especially at the initial 10 min. This change also matches with its microstructure evolution, i.e. these elongated grains have transformed to equiaxed crystals after annealing for 10 min. For the 410SS, its microhardness gradual decreases from initial 168×9.8 MPa to 140×9.8 MPa. Various equiaxed crystals grow slightly in this steel. Obviously, recrystallization occurs in these steels, especially the 304SS.

Fig.1  Tempering transformation of microhardness and micro-structure of 304SS and 410SS

2.2 Corrosion kinetics and phases

Fig.2 shows the corrosion kinetics of the 304SS and 410SS samples in liquid aluminum at 700 °C. For the 304SS, its mass change curve fits with linear laws roughly. After corrosion for 60 min, its mass loss reaches to 0.48 mg/mm2. Much lower corrosion rate is observed for the 410SS and an obvious incubation period occurs at the initial period. Thereafter, its corrosion rate increases gradually. Different corrosion kinetics is related with their recrystallization behavior. Referring to their kinetics, these samples are researched at three periods: 3, 10 and 60 min.

Fig.2  Corrosion kinetics curves of 304SS (a) and 410SS samples (b)

Fig.3 shows the XRD results of IMCS on the 304SS and 410SS after corrosion for different time. Both of Fe2Al5 and Fe4Al13 phases are detected on these two steels but in different ratios. For the 304SS, strong diffraction peaks of Fe2Al5 and Fe4Al13 phases are detected just after 3 min. Thereafter, most Fe4Al13 peaks are missing, leaving Fe2Al5 as the main phase after 60 min corrosion. For the 410SS, Fe2Al5 is the major product during the whole test, and weak peak of the Fe4Al13 phase even cannot be detected. Obviously, recrystallization accounts for different Fe4Al13/Fe2Al5 ratios on these steels.

Fig.3  XRD patterns of 304SS (a) and 410SS (b) after corrosion in molten aluminum at 700 °C for different time

2.3 Intermetallic compounds structure

In order to clarify the effect of recrystallization on IMCS growth, their microstructure, elements distribution and mechanical properties were researched in detail.

The microstructures of the IMCS on the 304SS are shown in Fig.4. After immerging in liquid-Al for 3 min, the 304SS surface turns to loose and porous. Spongy and acicular phases completely cover its surface (Fig.4a). Thereafter, many large-size holes are observed on this IMCS surface, as shown in Fig.4b. These deep holes provide rapid diffusion channels for liquid Al to erode the matrix continuously. Then, loose IMCS layer spalls off the 304SS easily, leading to its pitting (Fig.4c).

Fig.4  Microstructures of the IMCS on 304SS after corrosion for different time: (a, a1) 3 min, (b, b1) 10 min, and (c, c1) 60 min

Different from the case of the 304SS, much flatter IMCS surface morphology is observed on the 410SS, as shown in Fig.5. Uniformly distributed rod-like phases cover the 410SS surface just after 3 min test. Extending corrosion to 10 min, the morphology of IMCS shows no obvious change. Only some microcracks are initiated on surface, as shown in Fig.5b. The crack density on the IMCS further increases after 60 min corrosion, while no deep corrosion hole is found, as shown in Fig.5c.

Fig.5  IMCS surface morphologies on 410SS after corrosion for different time: (a, a1) 3 min, (b, b1) 10 min, and (c, c1) 60 min

The cross-sectional microstructure of the IMCS on the 304SS is shown in Fig.6. At the initial stage, an IMCS layer with loose outer sublayer and dense inner sublayer is observed on this steel (Fig.6a). Referring to XRD patterns and other researches[

8–11], the IMCS layer consists of outer Fe4Al13 and inner Fe2Al5. This IMCS layer is unstable and many island-like Fe4Al13 phases peel off its surface just after 3 min corrosion. Various microcracks are initiated in the porous Fe4Al13 layer, accelerating its detachment. Then, quite a lot of flocculent phases are observed on the IMCS surface (Fig.6b). Since this interfacial layer fails to prevent element interdiffusion, the outer Fe4Al13 layer breaks down after corrosion for 60 min. Significant recrystallization of this steel accounts for unprotective products and linear kinetic curve of the 304SS (Fig.2).

Fig.6  Cross-sectional microstructures of the IMCS on 304SS after corrosion for different time: (a, a1) 3 min, (b, b1) 10 min, and (c, c1) 60 min

Much thicker IMCS layer is observed on the 410SS surface in Fig.7. The dense Fe2Al5 layer, about 25 μm in thickness, shows typical tongue-like characteristics. This rough interface is in favor of the IMCS stability and it maintains the same thickness during initial 10 min (Fig.7b). This phenomenon also explains its incubation period in kinetics. Further extending corrosion to 60 min, the IMCS layer still keeps compact. Quite a few Fe4Al13 particles dissolve into liquid-Al as shown in Fig.7c. The high stability of this layer is also related with slight substrate recrystallization.

Fig.7  Cross-sectional microstructures of 410SS after corrosion for different time: (a, a1) 3 min, (b, b1) 10 min, and (c, c1) 60 min

In order to clarify the difference of IMCS on these two steels, its EDS elemental line scanning was used. As the product on the 304SS begins to degenerate after 10 min dipping test, elemental distribution on these two steels was analyzed at this case. These alloying elements in two stainless steels gradually dissolve into the melt. But quite different distribution tendency of Al is detected along the Fe-Al layer on these steels in Fig.8. For the 304SS, the content of Al changes sharply at the Fe4Al13/Al and Fe2Al5/substrate interfaces. However, the content of Al increases gradually from the matrix to Al side on the 410SS. Different element transformation mechanisms in Fe-Al layer accounts for elemental distribution discrepancy.

Fig.8  EDS elemental distribution of compound layer on 304SS (a) and 410SS (b) after 10 min corrosion

Nanoindentation curves of the steel matrix and IMCS layer after 10 min corrosion are shown in Fig.9[

20]. The peak load (green curve) of 304SS is higher than that of 410SS, matching with the Rockwell hardness result. The harder 304SS shows lower modulus than the 410SS in Table 2. For the IMCS layer, both of the hardness and modulus (red curve) of the 304SS are lower than those of 410SS. Referring to Ref.[21–22], the ratio of H/E also can be used to evaluate the IMCS layer's tough-ness. For the 304SS, this value (0.047) is lower than that of 410SS (0.050), indicating the better toughness of the IMCS on the later alloy. This result also matches with the high spallation tendency of Fe-Al layer from the 304SS matrix.

Fig.9  Nanoindentation load-displacement curves of compound layer and matrix on 304SS (a) and 410SS (b) after corrosion for 10 min

Table 2  Mechanical properties of IMCS layer and matrix
SampleHardness/GPaModulus/GPa
304SS 3.7 205.1
304SS IMCS 8.53 179.7
410SS 2.92 216.6
410SS IMCS 9.79 194.4

Above results reveal that recrystallization of these steels indeed affects the stability of IMCS layer in liquid Al. In order to illustrate the influence of this grain orientation rearrange-ment, the subsurface microstructures of two steels were analy-zed by EBSD after 10 min corrosion. As shown in Fig.10, strong (001) oriented structure is observed in the RD of the 304SS, matching with the silk texture observed in Fig.1. Along the ND, (111) and other oriented structure are detected. For the 410SS, strong (001) and (111) oriented structures are observed in the TD and ND, respectively. As the major corro-sion of steel plate sample is along ND, the grain orientation in this direction will affect its corrosion behavior significantly.

Fig.10  Reverse pole diagrams of matrix 304SS (a) and 410SS (b) grain orientation after corrosion for 10 min

For the 304SS, recrystallization happens at 700 °C. This structure transformation accelerates the degeneration of porous IMCS on this steel, leading to its linear corrosion kinetics. For the 410SS, slight corrosion occurs in this alloy due to its high structure stability. Dense interfacial Fe2Al5 layer formed on this steel delays the substrate dissolution. In the following part, the effect of steel recrystallization on the evolution of IMCS is discussed from corrosion dynamics, thermodynamics and microstructure.

2.4 Corrosion dynamics

Different corrosion dynamics is observed for the 304SS and 410SS after dipping in molten Al for 60 min. The kinetic curve of the 304SS follows linear laws, while an obvious incubation period is observed on the 410SS. During hot-dipping in liquid-Al, the total mass change of the matrix ΔM is determined by the amount of inward diffusion of Al atoms ΔAl and outward dissolution of Fe atoms ΔFe, as shown in Fig.11[

23–24].

Δm=ΔAl+ΔFe (1)

Fig.11  Fe/IMCS/Al interface diffusion diagram

As shown in Fig.11, when ΔAl>ΔFe, the growth of Fe2Al5 layer toward steel substrate is faster than that of Fe4Al13 layer dissolving into liquid aluminum. Then, the IMCS layer becomes thicker. When ΔAl<ΔFe, the thinner IMCS layer is related to the fast dissolution of Fe4Al13 layer. Under the case of ΔAlFe, the thickness of IMCS layer remains stable. This difference of IMCS thickness is related with their mass transportation discrepancy, caused by different recrystalliza-tion behavior.

In this study, the thickness of IMCS layer on the 304SS roughly remains a constant, matching ΔAlFe condition. Loose Fe-Al phase cannot prevent the outward dissolution of alloying elements. A great amount of holes and cracks in the Fe4Al13 layer provide preferential channels for the rapid diffusion of Al, as shown in Fig.4 and Fig.6. Moreover, significant variation of Al is observed at substrate/IMCS interface in Fig.8. This result reveals that element dissolution rate through these defects determines its corrosion.

For the 410SS, the IMCS layer becomes thicker after 10 min, confirming ΔAl>ΔFe within this steel. Elemental interdiffusion guides the growth of IMCS layer on this steel, as confirmed by an incubation period in its kinetic curve (in Fig.2). Under the protection of thick and dense IMCS layer, preferential diffusion of Al through the c-axis of Fe2Al5 lattice accounts for its tongue-like interfacial structure in Fig.7. This mass transportation pattern delays the failure of the 410SS[

25–27].

Different recrystallization degree accounts for the diversity of ΔAlFe on these steels. Deformation stress is released during this structural rearrangement, accompanied with matrix softening. The microhardness of the 304SS has decreased by 20.7% after 60 min annealing, while this decrement is 13.7% for the 410SS. Such high recrystallization degree of the 304SS increases its stress concentration at the IMCS/steel interface. Due to the poor deformation capacity of the intermetallic Fe-Al layer, various cracks are initiated at the IMCS/304SS interface, as shown in Fig.6. These defects build up fast channels for mass transformation through the IMCS layer and the amount of ΔAl is equal to ΔFe on the 304SS. On the contrary, the lower stress concentration ensures the stability of IMCS on the 410SS. The growth of IMCS is guided by the lower ΔFe on this steel.

2.5 Corrosion thermodynamics

For these two stainless steels, two Fe-Al phases are detected during test. In thermodynamics, the Gibbs free energy for the Fe2Al5 and Fe4Al13 formation at 973 K is calculated as follows[

7, 28].

2Al+5Fe=Fe2Al5, ΔG=-19.6 kJ/mol (2)
13Al+4Fe=Fe4Al13, ΔG=-22.9 kJ/mol (3)

Similar reaction energy of Eq.(2) and Eq.(3) indicates that the two reactions occur spontaneously during this test. However, different ratios of Fe2Al5/Fe4Al13 are observed for these two steels. This discrepancy is related with their recrystallization degree, which can offer stored deformation energy for reactions. For the 304SS, its original strip structure gradually transforms to equiaxed-grain just after annealing at 700 °C for 10 min, as shown in Fig.1. Released stain energy is in favor of Fe-Al phase nucleating on this steel surface, especially for the Fe4Al13. That is, the thickness of Fe4Al13 layer is about 1.5 times larger than that of the Fe2Al5 on the 304SS. This phenomenon is quite different from much thinner Fe4Al13 layer observed on the 410SS, which suffers slight recrystallization. So the formation of Fe4Al13 on steel is closely related with this matrix structural rearrangement.

2.6 Surface texture

Different surface textures are detected on the two stainless steels after their recrystallization. As the ND is parallel to diffusion path of Al ion, the orientation of matrix structure along ND affects steel’s corrosion significantly. In Fig.10, the two steels display the preferred [111] orientation along the ND.

For the austenite 304SS, [111] direction is perpendicular to (111) plane, which is the close-packed plane in fcc structure. The crystal face spacing of the (111) plane (0.2073 nm) is an order of magnitude larger than the gap of two adjacent iron atoms on this plane (about 0.02 nm). So aluminum atoms (radius=0.163 nm) tend to enter into the gap of two close-packed (111) planes of the 304SS. This result also explains loose Fe4Al13 layer and its lamellar detachment in Fig.6.

For the martensitic 410SS, its matrix phase transforms to sorbite (bcc structure) during annealing, as shown in Fig.1. The [111] orientation is the close-packed direction in bcc structure. Along this direction, the crystal face spacing of (111) is 0.1666 nm, which is close to aluminum atom radius (0.163 nm). It is difficult for Al ion to enter into the gap of (111) face on the 410SS after recrystallization.

Apart for the influence on Al ion diffusion, surface texture also causes different degrees of stress concentration at IMCS/steel interface[

29–30]. This stress mismatch (Ce) can be calculated by the following formula:

Ce=EIMCS/EFe (4)

where EIMCS is the IMCS layer modulus, EFe is the matrix modulus. Taking the data in Table 2, Ce≈0.876 for 304SS, and Ce≈0.897 for 410SS. It can be concluded that the modulus matching of 304SS is worse than that of 410SS. When the 304SS suffers thermal shock, the higher internal stress at IMCS/steel interface will weaken its boundary.

3 Conclusions

1) Loose and non-protective intermetallic compounds layer on the 304SS leads to linear corrosion kinetics, while thick and dense layer on the 410SS delays its failure.

2) Notable recrystallization on the 304SS reduces the energy for Fe-Al phases nucleation and growth, especially for Fe4Al13.

3) Slight structural rearrangement of the 410SS along its close-packed direction prevents the fast diffusion of Al.

4) Matchable stress of the Fe-Al layer and matrix is in favor of its well-adhesion.

References

1

Frutos E, Álvarez D, Fernandez L et al. Journal of Alloys and Compounds[J], 2014, 617: 646 [Baidu Scholar] 

2

Lemmens B, Springer H, Peeters M et al. Materials Science and Engineering A[J], 2021,710: 385 [Baidu Scholar] 

3

Chen G, Xue L, Wang J et al. Corrosion Science[J], 2020, 174: 108 836 [Baidu Scholar] 

4

Xu G, Wang K, Dong X et al. Journal of Materials Science and Technology[J], 2021,71: 12 [Baidu Scholar] 

5

Shi L, Kang J, Qian C et al. Materials Science and Engineering A[J], 2022, 831: 142 233 [Baidu Scholar] 

6

Deng Z, Xiao H, Yu C. Materials Design[J], 2022, 222: 111 106 [Baidu Scholar] 

7

Prasanthi T N, Sudha C, Rebby P. Surface and Coatings Technology[J], 2022, 440: 128 456 [Baidu Scholar] 

8

Takata N, Nishimoto M, Kobayashi S et al. Intermetallics[J], 2015, 67: 1 [Baidu Scholar] 

9

Cheng W, Wang C. Surface and Coatings Technology[J], 2009, 204: 824 [Baidu Scholar] 

10

Luo Q, Liu W, Li W et al. Journal of Materials Research and Technology[J], 2022, 17: 452 [Baidu Scholar] 

11

Awan G H, Hasan F. Materials Science and Engineering A[J], 2008, 472: 157 [Baidu Scholar] 

12

Dangi B, Brown T W, Kulkarni K N. Journal of Alloys and Compounds[J], 2018, 769: 777 [Baidu Scholar] 

13

Huilgol P, Udupa K, Bhat K U. Surface and Coatings Techno-logy[J], 2018, 348: 22 [Baidu Scholar] 

14

Liu B, Yang Q, Wang Y. Results in Physics[J], 2019, 12: 514 [Baidu Scholar] 

15

Cheng W J, Wang C J. Applied Surface Science[J], 2013, 277: 139 [Baidu Scholar] 

16

Chen S, Yang D, Zhang M et al. Metallurgical and Materials Transactions A[J], 2016, 47: 5089 [Baidu Scholar] 

17

Li Y, Jia Q, Zhu Z et al. Surface Review and Letters[J], 2017, 24: 1 750 046 [Baidu Scholar] 

18

Derazkola H A, Garcia E, Marrodán A M et al. Journal of Materials Research and Technology[J], 2022, 18: 2293 [Baidu Scholar] 

19

Huang K, Logé R. Materials Design[J], 2016, 111: 548 [Baidu Scholar] 

20

Oliver W, Pharr G. Journal of Materials Research[J], 1992, 7: 1564 [Baidu Scholar] 

21

Chen X, Du Y, Chung Y. Thin Solid Films[J], 2019, 688: 137 265 [Baidu Scholar] 

22

Musil J, Kunc F, Zeman H et al. Surface and Coatings Technology[J], 2002, 154: 304 [Baidu Scholar] 

23

Zhang J, Hosemann P, Maloy S. Journal of Nuclear Mater- [Baidu Scholar] 

ials[J], 2010, 404: 82 [Baidu Scholar] 

24

Dybkov V. Journal of Materials Science[J], 1990, 25: 3615 [Baidu Scholar] 

25

Azimaee H, Sarfaraz M, Mirjalili M et al. Surface and Coatings Technology[J], 2019, 357: 483 [Baidu Scholar] 

26

Jeshvaghani R, Emami M, Shafiee O et al. Surface and Coatings Technology[J], 2014, 240: 365 [Baidu Scholar] 

27

Cheng W J, Wang C J. Applied Surface Science[J], 2013, 277: 139 [Baidu Scholar] 

28

Rong J, Kang Z, Chen S et al. Materials Characterization[J], 2017, 132: 413 [Baidu Scholar] 

29

Patra L, Pandey R. Materials Today Communications[J], 2022, 31: 103 626 [Baidu Scholar] 

30

Surya R, Siddhartha P. Acta Materialia[J], 2008, 56: 3523 [Baidu Scholar]