Effect of Cold Rolling before Hydrogen Reduction on Reduction Behavior and Morphologies of Oxide Scale on Hot-rolled Low-carbon Steel

Abstract

A new ‘acid-free picking’ method that adding cold rolling prior to hydrogen-reduction descaling was put forward in this study. The purpose of cold rolling was to break the completeness of oxide scale and increase reduction efficiency. Subsequently, the high-temperature reduction behavior of oxide scale on the surface of hot-rolled low carbon steel, which reacted with hydrogen after 10% deformation of cold rolling in 20%H₂-Ar at the temperature range of 400–850°C, was investigated to validate the feasibility. The mass loss of specimens in the non-isothermal and isothermal reduction experiment was clearly measured by thermogravimetric analysis (TGA). Scanning electron microscopy (SEM) was used to observe surface morphology and cross-section microstructure of reduction products. The phase composition of reduction products was identified by X-ray diffraction (XRD). The experimental results indicated that the reduction reaction degree can be controlled by three main factors: reduction temperature, time and cold rolling. After 10% deformation of cold rolling, the completeness of original oxide scale was broken and then a large amount of micro-cracking appeared in oxide scale and the gap formed at oxide/substrate interface. The reduction rate of broken oxide scale was significantly faster than that of original oxide scale when the reduction temperature achieved more than 700°C. The effect of micro-cracking and gap on the mechanism of reduction process were clarified based on reaction of gas (H₂) – solid (iron oxides) interface.

1. Introduction

In the production process of hot-rolled steel strips, the oxide scale formed on the surface of steels due to high temperature and oxygen-rich environment.^1,2) In order to obtain a desirable surface, the oxide scale was usually removed before the deep processing of galvanization or cold rolling.³⁾ Generally, the descaling method was the usage of chloric acid or sulfonic acid in most factories. However, the duration and intensive usage of acid easily caused air pollution and groundwater pollution.

In view of the above-mentioned facts, instead of traditional acid picking method for removing the oxide scale on hot-rolled steel strips, the acid-free decaling process with gaseous (H₂, CO) reduction was proposed. Therefore, the gaseous reduction procedure was extensively studied in recent years due to the advantages of environmental protection.^4,5,6,7) In the research, Shi et al.⁵⁾ studied CO reduction of oxide scale on low-carbon hot-rolled steel and revealed that the optimum reduction temperature was 750°C for 3 min with a CO flow rate of 3 L/min. However, the reduction procedure with carbonic oxide was very difficult to be widely promoted, which attributed to the influence of reactant (CO) and the product (CO₂) on the individuals and the environment, respectively.⁸⁾ He et al.⁹⁾ investigated the reduction of oxide scale on hot-rolled low carbon steel strips in 10% H₂-Ar at the temperature range of 400–800°C. Nevertheless, the reduction reaction efficiency was slow owing to the low concentration of hydrogen. Guan et al.¹⁰⁾ experimentally evaluated the reduction of oxide scale on hot-rolled steel strips by 20 and 50% H₂ with N₂ balance at 550, 700 and 800°C for 4 min, and suggested that the rate of reduction became fast with the increase of reduction temperatures and hydrogen concentration. The results of previous researches indicated hydrogen has a strong edge on energy efficiency and environmental protection as the reactant compared with carbon monoxide. For this reason, the technology of ‘acid-free picking’ with the reactant (H₂) was gradually applied to the laboratory studies of galvanized sheet^11,12,13) and the high surface quality of cold rolling steel strips.^14,15) During the practical application, the procedure of reduction reaction is implemented through the hydrogen annealing furnace. If the reduction rate was slow, it needs to extend the length of hydrogen annealing furnace to obtain a far longer reaction time. However, the longer furnace will cause the problem of safety and cost. In light of these limitations, the new method how to improve reduction efficiency is urgent to be put forward to adapt to the needs of continuous production.

In this study, a new method is proposed that adding cold rolling before the hydrogen reduction process of oxide scale. The purpose of cold rolling is to break the completeness of oxide scale and obtain large amounts of micro-cracking in oxide scale and the gap at oxide/substrate interface. The micro-cracking and gap can provide diffusion paths of the reactant (H₂) and the product (H₂O) to accelerate reduction reaction. Subsequently, in order to identify the function of cold rolling on the high-temperature reduction behavior and microstructure evolution of the original and broken oxide scale in 20%H₂-Ar at the temperature range of 400–850°C are investigated.

2. Experimental

2.1. Materials

The hot-rolled low carbon steel strips with thickness 1.8 mm were used in this study and the chemical composition of specimens was 0.18C - 0.2Mn - 0.15Si - 0.004S - 0.002P - 0.4Cr (wt.%). The steel strips were compressed 10% with the lubricant by the four-high cold mill, and the process of cold rolling was exhibited in Fig. 1. The diameter of back-up roll and work roll of four-high mill were 350 mm and 150 mm, respectively. The rolling speed was 0.04 m/s and the mineral oil as the lubricant was used in the process of cold rolling. The specimens of the reduction experiment were cut into the size of 6 mm × 15 mm from steel strips before and after cold rolling deformation. Meanwhile, a hole of 1.5 mm diameter was drilled on the edge of specimens in order to suspend the specimen in the furnace chamber of continuous thermogravimetric analysis (TGA). Moreover, all specimens were cleaned by acetone prior to reduction experiments. For convenience, the oxide scale on the surface of hot-rolled steel strips was named as original oxide scale, and the oxide scale after 10% deformation of cold rolling was named as broken oxide scale.

Fig. 1.

Simplified sketch of cold rolling process.

2.2. Non-isothermal Reduction Experiment

The non-isothermal reduction experiment was carried out by Setsys Evolution 1750, SETARAM thermal-gravimetric analyzer (TGA) with a sensibility of 30 μg. As shown in Fig. 2(a), the specimen was placed inside the furnace and the furnace chamber was filled with Ar after evacuation, and then heated from room-temperature to 400°C at a rate of 30°C/min. After researching 400°C, the atmosphere of furnace was immediately replaced by the synthetic atmosphere 20%H₂-Ar with the gas flow rate of 200 mL/min. The specimen was heated in the reduction atmosphere from 400°C to 1000°C at a rate of 10°C/min. At last, the specimen in Ar reached room-temperature at a cooling rate of 60°C/min.

Fig. 2.

Schematic illustration of experimental procedure. (a) non-isothermal reduction of oxide scale from 400°C to 1000°C; (b) isothermal reduction of oxide scale at 400, 500, 600, 700, 800 and 850°C.

2.3. Isothermal Reduction Experiment

The isothermal reduction experiment was also conducted in TGA. As shown in Fig. 2(b), the specimen was placed inside the furnace and the furnace chamber was filled with Ar after evacuation, and then heated from room-temperature to reduction temperature (400, 500, 600, 700, 800 and 850°C) at a rate of 30°C/min. After researching the target temperature, the atmosphere of furnace was immediately replaced by the synthetic atmosphere 20%H₂-Ar with the gas flow rate of 200 mL/min. The specimen was isothermally reduced in the reduction atmosphere for 5 min and the weight variation as a function of time was recorded by a data acquisition system of precision balance. Subsequently, the specimen in Ar was cooled to room-temperature at a rate of 60°C/min.

2.4. Characterization

The mass loss of reduction experiments can be measured by thermal-gravimetric analyzer (TGA) and the reduction degree of oxide scale was calculated by mass loss per unit area. The scanning electron microscope (SEM) was used to observed surface and cross-section (longitudinal section) microstructure of reduction products. Oxygen element evolution of reduction products was exhibited by energy dispersive spectrometer (EDS). Phase composition of reduction products was identified by X-ray diffraction (XRD).

3. Results

3.1. Microstructure and Phase Composition of Oxide Scale before Hydrogen Reduction

The results of XRD in Figs. 3(a) and 3(b) can be observed that the phase composition of oxide scale which mainly contained magnetite and iron was not affected by the deformation of cold rolling. Figs. 3(c) and 3(d) shows surface morphology of oxide scale before and after cold rolling, respectively. A large amount of micro-cracking appeared on the surface of oxide scale after 10% deformation of cold rolling, importantly, there was not oxides breakaway from surface. The cross-section microstructure of original and broken oxide scale were presented in Figs. 3(e) and 3(f). It can be found that oxide scale mostly included the outward of magnetite layer and the lamellar of iron-magnetite eutectoid structure closed to substrate. In the lamellar structure, the white layers were metallic iron and the dark grey regions were magnetite.^16,17,18) Meanwhile, a large amount of micro-cracking in the oxide scale and the gap at oxide/substrate interface were observed after 10% deformation. The width of micro-cracking and gap were 0.62–1.43 μm and 0.1–1 μm, respectively.

Fig. 3.

X-ray diffraction spectra of (a) original and (b) broken oxide scale; Surface morphology of (c) original and (d) broken oxide scale; cross-section morphology of (e) original and (f) broken oxide scale.

3.2. Non-isothermal Hydrogen Reduction

The non-isothermal reduction curves of the original and broken oxide scale were shown in Fig. 4, which reflected the mass loss of oxide scale with increasing temperature. Compared Fig. 4(a) with Fig. 4(b), the reduction rates of oxide scale before and after deformation presented obvious discrepancy. For the original oxide scale, the onset, peak and slowdown temperature of reduction process can be observed at 440, 478 and 513°C, respectively. However, compared to the original oxide scale, the onset, peak and slowdown values of reduction temperature of the broken oxide scale were declined to 426, 462 and 501°C, respectively. The results indicated that the reducibility of original oxide scale was slower than that of broken oxide scale.

Fig. 4.

TGA and DTA curves for isothermal reduction of the oxide scale in synthetic atmosphere of 20% H₂-Ar, (a) original oxide scale; (b) broken oxide scale.

3.3. Isothermal Hydrogen Reduction

The effects of reduction temperature and time on the reduction degree of original and broken oxide scale in 20% H₂-Ar were investigated in this work. The reduction degree α was calculated by:   

$$\alpha =\frac{\mathrm{\Delta }{\text{W}}_t}{\mathrm{\Delta }{\text{W}}_{\alpha }}\times 100\%$$(1)

where ΔW_t was the mass loss of oxide scale at certain time t; ΔW_α was defined the mass loss during the reduction process that the total oxygen content in oxide scale was removed by hydrogen.^7,9) Figure 5 shows the isothermal reduction α-t curves of original and broken oxide scale at 400–850°C. It can be found that the mass loss of specimens increased with the time in all temperatures. Compared Fig. 5(a) with Fig. 5(b), the reduction time, temperature and deformation of cold rolling presented the significant influence on the reducibility of oxide scale. The experimental time of 1 min was spent on the replacement of protect atmosphere (Ar) with the reactant (H₂), and the next 4 min was used for the reduction reaction of gas-solid interface. As shown in Fig. 5(a), the reduction degree of broken oxide scale after 5 min hydrogen reduction was 40.7%, 30.2% and 19.6% at 850, 800 and 700°C, respectively, which was 1.66, 3.43 and 2.68 times higher than that of original oxide scale. However, the reduction degree of broken oxide scale was below 15% when the reduction temperature decreased to 400–600°C. Interestingly, it can be seen that the reduction degree of original oxide scale at 500°C was higher than that in the temperature range of 600–800°C. This is because the reaction of magnetite and hydrogen directly generate iron and water vapor without solid phase transition of oxide scale (Fe₃O₄→ Fe_(1−x)O + Fe),^8,19,20) the mechanism will be further discussed in Section 4.

Fig. 5.

α-t curves for isothermal reduction of original and broken oxide scale in synthetic atmosphere of 20% H₂-Ar, (a) 850, 800 and 700°C; (b) 600, 500 and 400°C.

3.4. Surface Morphology and Oxygen Element Analysis

The typical surface morphology of reduction products at 500, 700 and 850°C was presented in Fig. 6. The reduction products on the surface between original and broken oxide scale exhibited almost no difference in the same temperature except micro-cracking. Therefore, it is also of concern that the surface morphology of reduction products was dependent on reduction temperatures. The evolution of surface morphology was basically observed from porous iron to dense iron with the reduction temperature from 500°C and 700°C up to 850°C. Additionally, the main difference from porous iron was the pore size between 500°C and 700°C, which owing to the increasing of diffusion rate of the reactant (H₂) and the product (H₂O) and the reaction rate of gas-solid with the increasing of temperature. At 850°C, the partial oxygen in the oxide scale was removed by the reduction product (H₂O) vaporized, subsequently, the residual metallic iron started to nucleate, crystalize and grow on the surface of reduction products and formed dense iron layer finally.²¹⁾

Fig. 6.

Surface morphologies of reduction products in 20% H₂-Ar, original oxide scale reduction at (a) 500°C; (b) 700°C; (c) 850°C; broken oxide scale reduction at (d) 500°C; (e) 700°C; (f) 850°C.

To further illustrate the relationship between reduction temperatures and surface morphology, the oxygen content on the surface of reduction products was measured by EDS and presented in Fig. 7. It can be observed that the oxygen content on the surface of original and broken oxide scale exhibited a similar trend in all temperatures, which suggested the micro-cracking after cold rolling would not affect the microstructure of reduction products. In the case of reduction at 500 and 850°C, the residual oxygen content after gas-solid reaction was lower than that of the other temperatures. It is worth noting that the oxygen content analysis agreed well with the above mentioned reduction degree in Fig. 5.

Fig. 7.

EDS atomic percent of oxygen content vs temperature of the surface of reduction products in synthetic atmosphere of 20% H₂-Ar.

3.5. Phase Structure Evolution and Cross-section Microstructure

The X-ray diffraction by using Cu–Kα target has been performed for phase identification of original and broken oxide scale after reducing at the temperature range of 400–850°C and the diffraction peaks were presented in Fig. 8. On the basis of XRD analysis results, with the increasing of temperatures, magnetite diffraction intensity of peaks significantly decreased, and the intensity of wüstite and iron diffraction peaks was accordingly increased. When the reduction temperatures were at 400, 500 and 600°C, the phase structure of reduction products was composed of magnetite and iron. The magnetite diffraction intensity of peaks in original oxide scale was slightly lower than broken oxide scale at 500°C, and it was just opposite at 400 and 600°C. Furthermore, when the temperature increased to 700°C, the diffraction peaks of wüstite can be seen in the reduction products, and then the magnetite was entirely replaced by wüstite and iron when the temperature further increased to 800 and 850°C. The intensity of iron diffraction peaks for broken oxide scale was higher than that of original oxide scale at 800 and 850°C.

Fig. 8.

X-ray diffraction spectra of reduction products, (a) original oxide scale; (b) broken oxide scale.

The typical cross-section microstructure of original and broken oxide scale of reduction products at 500 and 850°C was presented in Fig. 9. Combined with results of XRD analysis, it can be found that the main phase of reduction products was magnetite and porous iron at 500°C. The reduction products consisted of wüstite, dense iron and porous iron closed to substrate at 850°C, and the wüstite was surrounded by dense iron. The size of porous iron at 500°C was clearly lower than that at 850°C. Additionally, it is worth noting that the process of cold rolling before hydrogen reduction experiment was helpful to increase the number of reduction products. The causes of this phenomenon could be divided into two aspects: on the one hand, hydrogen molecules easily absorbed on both sides of the micro-cracking due to the high roughness; on the other hand, the micro-cracking and the gap provided the diffusion and volatilization paths for the reactant (H₂) and the product (H₂O).

Fig. 9.

Cross-section morphologies of reduction products in 20% H₂-Ar, original oxide scale reduction at (a) 500°C; (b) 850°C; broken oxide scale reduction at (c) 500°C; (d) 850°C.

4. Discussion

The reduction process of oxide scale on the surface of hot-rolled steel strips in hydrogen includes the adsorption and inward diffusion of the reactant (H₂), the chemical reaction of gas-solid interface and the outward volatile of product (H₂O). The mechanism of hydrogen reduction is presented in Fig. 10. For original oxide scale, the reduction reaction depends on the surface adsorption and the grain boundary diffusion of reactant (H₂), which attributes to the completeness and dense of oxide scale. However, a large amount of micro-cracking appears in oxide scale and the gap forms at oxide/substrate interface after deformation of cold rolling owing to the compressive and tensile stress. The micro-cracking and the gap provide the diffusion paths for hydrogen and enlarges the contact area between gas phase and solid phase, and the formation of high roughness on the both sides of the micro-cracking increases absorbability of hydrogen atoms. This is one of the key reasons why the reduction degree of broken oxide scale was higher than that of original oxide scale.

Fig. 10.

Schematic illustration of mechanism for cold rolling procedure before the reduction process of oxide scale in hydrogen.

Except the chemical reaction between hydrogen and oxide scale, phase transitions of oxide scale also have significant influence on reduction behavior at high temperature. Based on the Fe–O equilibrium phase diagram, it can be noted that wüstite is thermodynamically stable above 570°C. During heating, the eutectoid structure of magnetite-iron in oxide scale transforms into wüstite through following reactions:^9,22)   

$${\text{Fe}}_{\text{3}}{\text{O}}_{\text{4}}\text{+}\mathrm{\hspace{0.17em} }\text{(1}-\text{4}y\text{)Fe}\to {\text{4Fe}}_{1-y}\text{O}$$(2)

  

$${\text{Fe}}_{\text{3}}{\text{O}}_{\text{4}}\text{+}\mathrm{\hspace{0.17em} }\text{(1}-\text{4}y{\text{)Fe}}^{\text{2+}}{\text{+2e}}^{-}\to {\text{4Fe}}_{1-y}\text{O}$$(3)

Therefore, the difference from reduction degree in Fig. 5 and reduction products’ phase composition in Fig. 8 at different temperatures can be explained by following reactions:   

$${\text{Fe}}_{\text{3}}{\text{O}}_{\text{4}}\text{+} {\text{H}}_{\text{2}}\to \text{Fe} \text{+} {\text{H}}_{\text{2}}\text{O}\left(\mathrm{B}\mathrm{l}\mathrm{o}\mathrm{w}\mathrm{\hspace{0.17em} }570\mathrm{\hspace{0.17em} }\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{e}\mathrm{e}\mathrm{\hspace{0.17em} }\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{e}\right)$$(4)

  

$$\begin{array}{l}{\text{Fe}}_{\text{3}}{\text{O}}_{\text{4}}\text{+}\mathrm{\hspace{0.17em} }{\text{H}}_{\text{2}}\to \text{FeO}\mathrm{\hspace{0.17em} }\text{+}\mathrm{\hspace{0.17em} }{\text{H}}_{\text{2}}\text{O}\mathrm{\hspace{0.17em} }\text{+}\mathrm{\hspace{0.17em} }{\text{H}}_{\text{2}}\to \text{Fe}\mathrm{\hspace{0.17em} }\text{+}\mathrm{\hspace{0.17em} }{\text{H}}_{\text{2}}\text{O}\\ \left(\mathrm{A}\mathrm{b}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{\hspace{0.17em} }570\mathrm{\hspace{0.17em} }\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{e}\mathrm{e}\mathrm{\hspace{0.17em} }\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{e}\right)\end{array}$$(5)

Consequently, if the new method that adding cold rolling before hydrogen reduction process of oxide scale can be developed in engineering practice, the pre-condition is the reduction temperature above 700°C.

5. Conclusion

(1) After 10% deformation of cold rolling, a large amount of micro-cracking appeared in oxide scale and the gap formed at oxide/substrate interface. The experimental result indicates that the micro-cracking and gap not only provided diffusion paths of reactant (H₂) and product (H₂O) but also enlarged the region of gas-solid phase reaction interface.

(2) The reduction temperature played an important role in determining the morphology and phase composition of reduction products. The porous iron and dense iron can be observed blew and above 700°C, respectively, when the oxide scale was reduced in 20% H₂-Ar for 5 min. Meanwhile, at the temperature range of 400–600°C, the phase composition of reduction products was magnetite and iron. At 700°C, magnetite, wüstite and iron existed simultaneously. At 800 and 850°C, magnetite was fully reduced to wüstite and iron.

(3) The new method that adding cold rolling before the procedure of hydrogen reduction was verified a feasible way to efficiently remove the oxide scale on the surface of hot-rolled steel strips under the optimum reduction temperatures (above 700°C).

Acknowledgements

The authors would like to thank the National Nature Science Foundation of China (51204047 and U1660117) for the financial support.

References

• 1)   R. Y.  Chen and  W. Y. D.  Yuen: ISIJ Int., 45 (2005), 52.
• 2)   L.  Suárez,  Y.  Houbaert,  X. V.  Eynde and  R.  Colás: Corros. Sci., 51 (2009), 309.
• 3)   Y.  Wang and  J.  Zeng: Mater. Des., 69 (2015), 64.
• 4)   S. K.  Kuila,  R.  Chatterjee and  D.  Ghosh: Int. J. Hydrog. Energ., 41 (2016), 9256.
• 5)   J.  Shi,  D. R.  Wang,  Y. D.  He,  H. B.  Qi and  W.  Gao: Mater. Lett., 62 (2008), 3500.
• 6)   E. A.  Mousa,  M.  Bahgat and  A. A.  El-Geassy: Ironmaking Steelmaking, 40 (2013), 452.
• 7)   C.  Guan,  J.  Li,  N.  Tan,  Y. Q.  He and  S. G.  Zhang: Int. J. Hydrog. Energ., 39 (2014), 15116.
• 8)   A.  Pineau,  N.  Kanari and  I.  Gaballah: Thermochim. Acta, 447 (2006), 89.
• 9)   Y. Q.  He,  T.  Jia,  Z. F.  Li,  G. M.  Cao,  Z. Y.  Liu and  J.  Li: Metall. Mater. Trans. A, 47 (2016), 4845.
• 10)   C.  Guan,  J.  Li,  N.  Tan and  S. G.  Zhang: Ironmaking Steelmaking, 43 (2016), 739.
• 11)   C.  Guan,  J.  Li,  N.  Tan,  S. G.  Zhang and  W. Y.  Zhang: Surf. Coat. Technol., 279 (2015), 142.
• 12)   Y. Q.  He,  T.  Jia,  X. J.  Liu,  G. M.  Cao,  Z. Y.  Liu and  J.  Li: J. Iron Steel Res. Int., 21 (2014), 222.
• 13)   N.  Tan,  J.  Li and  C.  Guan: Ironmaking Steelmaking, 40 (2013), 578.
• 14)   Y. A.  Jing,  X.  Zang,  Q.  Shang,  Y.  Qin,  Y.  Li and  B.  Song: J. Mater. Process. Technol., 219 (2015), 303.
• 15)   Y. A.  Jing,  Q.  Shang,  X.  Zang,  L.  Zhang,  X.  Peng and  P.  Jia: Metall. Mater. Trans. A, 47 (2015), 424.
• 16)   R. Y.  Chen and  W. Y. D.  Yuen: Oxid. Met., 53 (2000), 539.
• 17)   Y.  Shizukawa,  S.  Hayashi,  S.  Yoneda,  Y.  Kondo,  H.  Tanei and  S.  Ukai: Oxid. Met., 86 (2016), 315.
• 18)   H.  Tanei and  Y.  Kondo: ISIJ Int., 52 (2012), 105.
• 19)   A.  Pineau,  N.  Kanari and  I.  Gaballah: Thermochim. Acta, 456 (2007), 75.
• 20)   M. C.  Bagatini,  V.  Zymla,  E.  Osório and  A. C. F.  Vilela: ISIJ Int., 51 (2011), 1072.
• 21)   M. C.  Bagatini,  V.  Zymla,  E.  Osório and  A. C. F.  Vilela: ISIJ Int., 54 (2014), 2687.
• 22)   R. Y.  Chen and  W. Y. D.  Yuen: Oxid. Met., 56 (2001), 89.