Evaluation for Carbonization Rate of Porous Iron Whisker with CO

Abstract

Currently, research and developments are underway for steel production using hydrogen-direct reduced iron and large-scale electric arc furnace (EAF). Fe₃C has attracted attention as a carburizing agent and clean source of iron for EAF steel production to lower the concentration of impurities. However, production capacity of cementite is low since the carbonization reaction rate of reduced iron pellet is limited by gas diffusion inside micropores in the pellet.

In this study, the carbonization reaction rate of porous iron whisker with approximately 95% of porosity was examined. The porous iron whisker can be produced through carbothermal reduction of fine iron ore and biochar mixture. The carbonization reaction rate of the porous iron whisker was approximately three times faster than that of fine iron particles. The porous iron whisker has advantageous for rapid cementite production compared to fine iron particles since the effective surface area is larger in the porous iron.

1. Introduction

In the ironmaking process, an innovative technology to drastically reduce carbon dioxide emissions has been required. Currently, research and developments are underway for primary steel production using hydrogen-direct reduced iron (H-DRI)^1,2,3) and large-scale electric arc furnace (EAF).^4,5) In Sweden, the HYBRIT project has already begun the operation of a pilot plant for H-DRI production using green hydrogen in August 2020,⁶⁾ the aim is to create a demonstration plant by 2025. The H-DRI contains little carbon unlike the DRI produced by natural gas. The dissolved carbon in iron is, however, essential for primary steelmaking process since it lowers the melting point of metallic iron and protects from re-oxidization during refining process. Since utilization of fossil fuel-derived carbon leads to emitting CO₂ in the atmosphere, Fe₃C has attracted attention as a carburizing agent.⁷⁾ In the EAF steelmaking, on the other hand, the contamination of Cu and Sn derived from steel scrap is a problem. Such elements called tramp elements cause the surface cracking of the steel during hot working.⁸⁾ The permissible concentration of Sn in steel has been estimated at 0.04 mass%.⁸⁾ However, tramp elements cannot be removed by oxidation refining. Thus, clean source of iron required for the EAF steelmaking to lower the concentration of impurities. Especially, Fe₃C is considered to be suitable for supplement in EAF since it contains carbon as a chemical energy source.⁹⁾ In 1990s, the first commercial plant to produce Fe₃C was built in Trinidad by Nucor.¹⁰⁾ The operation only lasted for a few years since the production capacity was lower than expected.¹¹⁾

Our group has proposed a new carbon recycling ironmaking process, CRIP-D.¹²⁾ In the CRIP-D process, solid carbon is recovered by reforming exhaust gas as iron carbide and free carbon. Hot metal can be produced by using the solid carbon as reducing and carburizing agents. Iron carbide and free carbon are generated in the surface of iron by CO gas according to the reactions described as Eqs. (1) and (2).

  

$$2\text{CO}=\text{C}+{\text{CO}}_2$$(1)

  

$$3\text{Fe}+\text{C}={\text{Fe}}_3\text{C}$$(2)

Numerous studies on iron carbonization have been conducted. Hayashi et al.¹³⁾ has investigated the carbonization behavior of reduced metallic iron pellet with CO–H₂ gas mixture. The iron pellet is carbonized topochemically indicating the carbonization rate is limited by gas diffusion. The carbonization rate of reduced iron particles, on the other hand, is increased with increasing the surface area of the iron while is decreased by submicron pore diffusion resistance.¹⁴⁾ Thus, further acceleration of the carbonization rate is expected by using iron with large pore size and surface area. Such porous iron whisker can be produced via a carbothermic reduction of carbon and iron oxide composite at 1000°C.¹⁵⁾ The porous iron whisker exhibits approximately 95% porosity and open-celled structure with large void diameter by intertwining the iron whisker. However, there is no data of its carbonization rate. Furthermore, the effect of gas diffusion resistance in pores on iron carbonization have not been investigated quantitatively.

The objective of this study is to investigate the carbonization rate using the porous iron whisker as a substrate for the carbonization reaction and the effect of gas diffusion in the pores on the carbonization.

2. Experimental

Hematite reagent (1 μm size, purity 99%) and biomass char (53–150 μm size) were well mixed to be that the molar ratio of fixed carbon in the char to oxygen in the hematite, C/O ratio, was 0.75. The hematite reagent contains 0.11 mass% of sulfur. The used biomass char contains 95.2 mass% of fixed carbon, 3.9 mass% of volatile matter and 0.9 mass% of ash in dry basis. The mixture was filled in the crucible with 26 mm in diameter. The porous iron whisker was prepared by heating the mixture up to 1000°C from room temperature at the heating rate of 10°C/min and holding for 90 min in N₂ atmosphere. The porous iron whisker has approximately 250 ppm of sulfur measured by infrared absorption method after combustion. This sulfur concentration in iron is at same level as the iron particle reduced by H₂–H₂S mixed gas.^16,17) The appearance and SEM image of the porous iron whisker are shown in Figs. 1 and 2, respectively. It shows high porosity of 94.6% due to its fibrous structure. The specific surface area of the produced iron was evaluated as 0.92 m²/g by BET method. The sample for carbonization experiment was prepared by cutting the porous iron whisker with 5 mm in thick, and was set in an experimental apparatus shown in Fig. 3. The porous iron whisker piece was put on between mullite tubes with 20 mm in height. The underside tube was laid a bottom with 0.5 mm diameter Pt wire woven into mesh. The Pt wire, thermocouple tips and mullite tubes were coated by applying ceramic slurry to avoid proceeding the carbon deposition reaction on them. A gas pre-heater zone was prepared by filling the alumina ball with 3 mm diameter in the tube with 10 mm in thick. The gap between the sample side and the tube was filled with clayey inorganic adhesive. In addition, the surface of the adhesive was coated with ceramic wool to protect the reaction tube. After setting the sample holder inside a fused silica reaction tube, the tube was sealed. Ar-10%H₂ gas was introduced from the bottom of the tube at atmospheric pressure after vacuuming the air using a pump. The gas flow rate was kept constant at 500 mL/min controlled by mass flow controllers. The sample was heated by an infrared image furnace and maintained at 250°C for 3.6 ks to cure the inorganic adhesive. The heating temperature was increased to 800°C for 3.6 ks for pre-reduction of the sample. Subsequently, the heating temperature was set at 600, 650, 700, 750, and 800°C. Carbonization reaction proceeded for 3.6 ks by introducing CO gas after Ar gas was purged for 1.8 ks to remove H₂ remaining inside the reaction tube. After that, the flowing gas was switched to Ar gas to terminate the carbonization reaction. Simultaneously, the furnace turned off and cooled to room temperature. The CO₂ partial pressure of the outlet gas, $P_{{\text{CO}}_2}^{\text{out}}$, during the experiment was quantitatively analyzed by gas chromatography. $P_{{\text{CO}}_2}^{\text{out}}$ is expressed by Eq. (3).

  

$$P_{{\text{CO}}_2}^{\text{out}}=\frac{M_{{\text{CO}}_2}^{\text{out}}}{M_{\text{CO}}^{\text{in}}-M_{\text{C}}}$$(3)

where, $M_{{\text{CO}}_2}^{\text{out}}$ and $M_{\text{CO}}^{\text{in}}$ represent the mol amount of CO₂ in the outlet gas and CO in the inlet gas, respectively. M_C is the mol amount of deposited carbon and same amount with $M_{{\text{CO}}_2}^{\text{out}}$ according to Eq. (1). Hence, Eq. (3) can be transformed to Eq. (4).

  

$$M_{\text{C}}=\frac{P_{{\text{CO}}_2}^{\text{out}}}{1+P_{{\text{CO}}_2}^{\text{out}}}\times M_{\text{CO}}^{\text{in}}$$(4)

where $P_{{\text{CO}}_2}^{\text{out}}$ and $M_{\text{CO}}^{\text{in}}$ represent the partial pressure of CO₂ in the outlet gas and mol amount of CO in the inlet gas. The carbonization degree, f_θ, is defined as shown in Eq. (5).

  

$$f_{\theta }=\frac{M_{\text{C}}}{M_{\text{C}}^{\theta }}$$(5)

where $M_{\text{C}}^{\theta }$ is the ideal mol amount of combined carbon as the porous iron whisker sample was totally carbonized to cementite (Fe₃C). The reaction products were pulverized into powder using an alumina mortar. The constituents of the sample powder were identified by XRD (tube voltage 40 kV, tube current 30 mA, Fe Kα (0.19373 nm)). The interruption experiments for 180, 300, and 420 s were also conducted for carbonization at 700°C. Pulverized reaction products were filled with resin, cut, and polished. The exposed cross sections were subjected to etching treatment for iron carbide using sodium picrate.^16,18)

Fig. 1. Appearance of the porous iron whisker. (Online version in color.)

Fig. 2. SEM image of the porous iron whisker.

Fig. 3. Schematic diagram of an apparatus for carbonization experiment using an image furnace. (Online version in color.)

3. Results and Discussion

Figure 4 shows the changes in the carbonization degree of porous iron whisker for each temperature with reaction time. The degree at 600°C increases with time almost linearly and reaches 0.91 for 3.6 ks of reaction time. The slope of the degree increases with increasing the temperature up to 750°C. Under the conditions of 650°C and 700°C, the degree also increases linearly in the early stage of the reaction, however, the slope of the carbonization curve decreases as the degree approaches 1.0. In the case of 800°C, on the other hand, the carbonization degree continues to increase at a constant rate even above 1.0. Especially at 800°C, the carbon deposition reaction proceeds significantly even when the carbonization degree exceeds 1.0.

Fig. 4. Changes in carbonization degree of porous iron whisker for each temperature with reaction time. (Online version in color.)

Figure 5 shows XRD profiles obtained for the iron samples carbonized each temperature for 3.6 ks. The carbonized porous iron whisker exhibited such brittleness that it was readily pulverized to micron-sized powder using an alumina mortar. The Fe₃C peaks are observed in all samples. Especially, only Fe₃C peaks are detected at 650°C and 700°C. This indicates the stable generation of Fe₃C. Sawai et al.¹⁷⁾ reported that the sulfur atoms absorbed on iron blocks carbon precipitation on the surface leading stable formation of Fe₃C. In this study, the porous iron whisker and hematite reagent has high sulfur concentration. Hence, it is possible that the porous iron whisker is stably carbonized to Fe₃C due to existence of sulfur. Previous report indicates that carbon precipitation is preceded by the formation of cementite due to carbonization of iron.¹⁹⁾ The Fe peaks are also observed in the sample carbonized at 600°C since uncarbonized metallic iron remained. The samples carbonized at 750°C and 800°C, on the other hand, the C peaks are recognized along with the Fe peaks. The peaks of Fe₄C are also shown. DeCristofaro et al.²⁰⁾ reported that the carbon in martensite is rearranged to form Fe₄C by the aging effect at room temperature. In this study, Fe₄C was observed under conditions where the carbonization reaction proceeded above eutectoid temperature (723°C). This might be because relatively high amount of carbon can dissolve in iron.

Fig. 5. XRD profiles of the porous iron whisker carbonized at each temperature for 3.6 ks by CO gas. (Online version in color.)

The magnified appearance of porous iron whisker sample carbonized at each temperature are shown in Fig. 6. All samples maintain a fibrous structure. The carbon particles are significantly observed on the surface of the sample carbonized at 800°C. The particle exhibits a size in the submicron range.

Fig. 6. SEM images of the porous iron whisker carbonized at each temperature for 3.6 ks by CO gas.

To examine the carbonization behavior of the porous iron whisker, interruption tests at 180 s, 300 s and 420 s in the carbon deposition experiment at 700°C were conducted. Figure 7 shows cross-section of the sample particles etched by alkaline solution of sodium picrate with the carbonization degree. Speckled Fe₃C grains grow over carbonization time. Since this microstructural change suggests the carbonization reaction of the porous iron whisker is dominated by the nucleation-growth of Fe₃C same as in the case of iron particles,¹⁶⁾ the Johnson-Mehl-Avrami equation shown in Eq. (6) was used to determine the carbonization rate, R_C.

  

$$f_{\theta }=1-exp(-{(R_{\text{C}}t)}^n)$$(6)

where, n is avrami constant. Equation (6) can be transformed as in Eq. (7). Figure 8 shows the relation between carbonization degree and reaction time. The reaction rate can be obtained from the slope and intercept of the graph plotting the carbonization degree with the vertical axis as ln(−ln(1−f_θ) and the horizonal axis as lnt.

  

$$ln(-ln(1-f_{\theta })=nlnt+nln{R}_{\text{C}}$$(7)

For each temperature condition, the plot varies almost linearly from the start of carbonization to the point where the carbonization degree reaches 0.8. Thus, the carbonization reaction rate was determined from the approximate line of the plot in this section. Avrami constant, n, can be obtained from the slope of the approximate line. Table 1 shows the Avrami constant of the carbonization rate. The Avrami constant exhibits nearly 1 at each temperature. This indicates that nucleation of Fe₃C readily initiates and Eq. (6) can be regarded as a first order reaction equation.¹⁶⁾ Thus, the changes in carbonization degree with time do not exhibit sigmoid curve. Arrhenius plots of the carbonization rate are shown in Fig. 9. The reported carbonization reaction rate of iron particles reduced by H₂–H₂S mixed gas (Size: 150–210 μm)¹⁶⁾ are also shown. The porous iron whisker is carbonized approximately three times faster than iron particles. The carbonization reaction rate has a peak at around 750°C in relation to the reaction temperature. The carbonization reaction of iron is an overall reaction of Eqs. (1) and (2). The carbon deposition reaction shown in Eq. (1) can be divided into elementary processes as follows;²¹⁾

  

$$\text{CO}={\text{O}}^{*}+\text{C}$$(8)

  

$${\text{O}}^{*}+\text{CO}={\text{CO}}_2$$(9)

where, O^* indicates the oxygen atom on the surface of the iron. Equation (9) is rate-determining reaction.²¹⁾ The reaction rate of Eq. (9), k₂, increases with rising the reaction temperature, whereas the equilibrium constant of Eq. (8), K₁, decreases. Therefore, the carbonization rate reaches its maximum at a certain temperature.²²⁾ In this experiment, the maximum carbonization reaction rate is observed at 750°C. However, the crystal structure of iron affects the carbonization behavior as shown in Fig. 5. Hence, the carbonization behavior is investigated in the temperature range of 600°C to 700°C. The activation energy of carbonization reaction by CO gas was evaluated as 90.3 kJ/mol. Sawai et al.¹⁶⁾ has also estimated the activation energy as 81.6–88.0 kJ/mol.

Fig. 7. Cross-section of porous iron whisker particles etched by alkaline solution of sodium picrate after carbonization at 700°C for 180, 300 and 420 s. (Online version in color.)

Fig. 8. Relation between carbonization degree and reaction time, ln(−ln(1−f_θ)) vs. ln t, in CO gas at each temperature. (Online version in color.)

Table 1. Avrami constant of the carbonization rate of porous iron whisker at each temperature.

T (°C)	600	650	700	750	800
n	1.40	1.29	1.21	1.17	1.08

Fig. 9. Arrhenius plot of reaction rate of porous iron whisker and iron particles¹⁶⁾ in CO gas at each temperature. (Online version in color.)

The obtained result that porous iron whisker is carbonized much faster than iron particles is discussed in terms of gas diffusion in the pores and voids. Molecular diffusion is a form of gas diffusion which the molecular collides each other. Its diffusion coefficient, D_m can be calculated as shown in Eq. (10).²¹⁾ In the case of Knudsen diffusion, on the other hand, the gas molecular collides to pore wall caused by short mean free pass. The Knudsen diffusion coefficient, D_k is described in Eq. (11).²³⁾

  

$$D_m=\frac{2kT}{3\pi P{\sigma }^2}\sqrt{\frac{RT}{\pi M}}$$(10)

  

$$D_k=\frac{2d}{3}\sqrt{\frac{2RT}{\pi M}}$$(11)

where, k and R are Boltzmann’s constant and gas constant, respectively. σ and M are diameter and weight of the molecular. d is pore diameter. The gas diffusion in pores is expressed by serial model of molecular and Knudsen diffusions. The effective diffusion coefficient is given by Eq. (12).^23,24)

  

$$\frac{1}{D}=\frac{1}{D_m}+\frac{1}{D_k}$$(12)

Figure 10 shows the relation between effective diffusion coefficient of CO gas and pore diameter at each temperature. The diffusion coefficient is relatively large when the pore size is larger than 1 μm. Furthermore, temperature dependence is significantly observed. This indicates the gas diffusion is dominated by molecular diffusion. Since Knudsen diffusion gets dominance of the gas diffusion as the pore size decreases, both the diffusion coefficient and temperature dependence become smaller. The dominant factor of gas diffusion can be determined by the coefficient ratio of the Knudsen diffusion to effective diffusion, $R_K=\frac{D}{D_k}$. Molecular diffusion dominates when R_K < 0.5 and Knudsen diffusion dominates when R_K > 0.5.²³⁾ Therefore, in this study, assuming that CO gas diffusing by molecular diffusion contributes to the carbonization reaction, the effective pore diameter, d^* is defined as the pore diameter at which R_K = 0.5. The relation between R_K and pore diameter at each temperature using CO gas is shown in Fig. 11. The temperature dependence of R_K is almost negligible. The effective pore diameter is approximately 0.2 μm at all temperatures. In the case of the porous iron whisker, the void size is much larger than 0.2 μm as shown in Fig. 2. Thus, whole specific surface area of the iron measured by the BET method can be considered as the effective surface area. Kondo et al.²⁵⁾ reported the accumulated pore volume distribution of iron particles (Size: 180–250 μm) reduced by hydrogen at 650°C. They pointed the volume distribution was concentrated in the pores on the order of 0.1 μm in diameter. In this study, the distribution of pore surface area, S is calculated from the pore volume, V²⁵⁾ according to Eq. (13) assuming a pore shape as an elongated cylinder.

  

$$S=\frac{4V}{d}$$(13)

Fig. 10. Relation between effective diffusion coefficient and pore diameter at each temperature using CO gas. (Online version in color.)

Fig. 11. Relation between the coefficient ratio of the Knudsen diffusion to effective diffusion and pore diameter at each temperature using CO gas. (Online version in color.)

Figure 12 shows the cumulative distribution of pore surface area with pore diameter.²⁵⁾ The specific surface area of whole pores is accumulated to be 1.35 m²/g. This value is close to the specific surface area of hydrogen-reduced iron particles measured by BET method reported by Inoue et al.²⁶⁾ The most of the specific surface area of the iron particles is that of pores less than 0.2 μm in diameter. Consequently, the effective surface area of iron particles for the carbonization is small, approximately 0.3 m²/g. Arrhenius plots of the carbonization rate per effective surface area, R_C^* are shown in Fig. 13. Independently of the type of substrate iron, the carbonization rate per effective surface area is almost same. This suggests that the porous iron whisker which has large effective surface area is advantageous for rapid cementite production compared to iron particles. The large void size enables the CO gas to diffuse through molecular diffusion even inside the voids, and hence the porous iron whisker is carbonized much faster.

Fig. 12. Cumulative under specific surface area distribution with pore radius of reduced iron particle with hydrogen.²⁵⁾ (Online version in color.)

Fig. 13. Arrhenius plot of reaction rate per unit effective surface area of porous iron whisker and iron particles¹⁶⁾ in CO gas at each temperature. (Online version in color.)

4. Conclusions

In this study, the objective is investigating the carbonization rate using the porous iron whisker as a substrate for carbonization reaction by CO gas. The factors for accelerating of the carbonization rate were also examined. The following results were obtained.

• The carbonization reaction of porous iron whisker with 5 mm thick is dominated by the nucleation-growth of Fe₃C same as in the case of iron particles. The rate, however, is approximately three times faster in the porous iron whisker than iron particles.

• The void diameter of the porous iron whisker is large enough so that gas molecules diffuse inside the voids by molecular diffusion.

• The porous iron whisker has advantageous for rapid cementite production compared to fine iron particles since the effective surface area is larger in the porous iron whisker.

Declaration of Competing Interest

The authors declare that they have no known competing financial interest or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgment

This work was supported by JSPS Grant-in-Aid for JSPS Fellows Grant Number 22KJ0282 and Steel Foundation for Environmental Protection Technology.

References

• 1)   J.  Huss,  A.  Vickerfält and  N.  Kojola: Steel Res. Int., 95 (2024), 2300325. https://doi.org/10.1002/srin.202300325
• 2)   P.  Cavaliere,  A.  Perrone,  A.  Silvello,  P.  Stagnoli and  P.  Duarte: Metals, 12 (2022), 203. https://doi.org/10.3390/met12020203
• 3)   A.  Pfeiffer,  K.  Thiele,  H.  Zheng,  H.  Mali,  G.  Wimmer and  J.  Schenk: Steel Res. Int., (2024), 2300268. https://doi.org/10.1002/srin.202300268
• 4)   D.  Mombelli,  G.  Dall’Osto,  C.  Mapelli,  A.  Gruttadauria and  S.  Barella: Steel Res. Int., 92 (2021), 2000580. https://doi.org/10.1002/srin.202000580
• 5)   M.  Sahoo,  S.  Sarkar,  A. C. R.  Das,  G. G.  Roy and  P. K.  Sen: Steel Res. Int., 90 (2019), 1900034. https://doi.org/10.1002/srin.201900034
• 6)  HYBRIT (Hydrogen Breakthrough Ironmaking Technology): SSAB, LKAB and Vattenfall first in the world with hydrogen-reduced sponge iron, (2021), https://www.hybritdevelopment.se/en/hybrit-ssab-lkab-and-vattenfall-first-in-the-world-with-hydrogen-reduced-sponge-iron, (accessed 2024-04-09).
• 7)   H.  Ahmed,  T. K. S.  Kumar,  J.  Alatalo and  B.  Björkman: J. Mater. Res. Technol., 21 (2022), 1760. https://doi.org/10.1016/j.jmrt.2022.10.019
• 8)   K. E.  Daehn,  A. C.  Serrenho and  J. M.  Allwood: Environ. Sci. Technol., 51 (2017), 6599. https://doi.org/10.1016/j.jmrt.2022.10.019
• 9)   D. Y.  Kim,  Y. U.  Heo and  Y.  Sasaki: ISIJ Int., 53 (2013), 950. https://doi.org/10.2355/isijinternational.53.950
• 10)   R.  Garraway: Conference: 79. Steelmaking and 55th Ironmaking Conference, Vol. 55, (1996), 807. https://www.osti.gov/biblio/460653, (accessed 2024-04-09).
• 11)   Y.  Kapelyushin,  Y.  Sasaki,  J.  Zhang and  O.  Ostrovski: Extraction 2018. The Minerals, Metals & Materials Series, (2018), 683. https://doi.org/10.1007/978-3-319-95022-8_54
• 12)   R.  Higashi,  D.  Maruoka,  Y.  Iwami and  T.  Murakami: ISIJ Int., 64 (2024), 1486. https://doi.org/10.2355/isijinternational.ISIJINT-2024-073
• 13)   S.  Hayashi,  O.  Asai,  S.  Sawai and  Y.  Iguchi: Tetsu-to-Hagané, 85 (1999), 208 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.85.3_208
• 14)   K.  Horii,  Y.  Iguchi,  S.  Hayashi and  T.  Moriya: Tetsu-to-Hagané, 89 (2003), 1011 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.89.10_1011
• 15)   T.  Murakami,  T.  Takahashi,  S.  Fuji,  D.  Maruoka and  E.  Kasai: Mater. Trans., 58 (2017), 1742. https://doi.org/10.2320/matertrans.M2017232
• 16)   S.  Sawai,  Y.  Iguchi and  S.  Hayashi: Tetsu-to-Hagané, 85 (1999), 20 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.85.1_20
• 17)   S.  Hayashi and  Y.  Iguchi: ISIJ Int., 37 (1997), 16. https://doi.org/10.2355/isijinternational.37.16
• 18)   S.  Sawai,  Y.  Iguchi and  S.  Hayashi: Tetsu-to-Hagané, 85 (1999), 6 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.85.1_6
• 19)   R.  Higashi,  D.  Maruoka,  Y.  Iwami and  T.  Murakami: Proceedings of METEC and 6th ESTAD (European Steel Technology and Application Days), Düsseldorf, (2023).
• 20)   N.  DeCristofaro and  R.  Kaplow: Metall. Trans. A., 8 (1977), 35. https://doi.org/10.1007/BF02677261
• 21)   S.  Ando,  T.  Shimoo and  K.  Moriguchi: J. Jpn. Inst. Met., 49 (1985), 45 (in Japanese). https://doi.org/10.2320/jinstmet1952.49.1_45
• 22)   X.  Jiang,  R.  Takahashi,  T.  Akiyama and  J.  Yagi: Tetsu-to-Hagané, 83 (1997), 299 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.83.5_299
• 23)   N.  Wakao: Kagaku Kōgaku, 28 (1964), 561 (in Japanese). https://doi.org/10.1252/kakoronbunshu1953.28.561
• 24)   S.  McCormack,  H.  Cao,  J. P.  Martins,  P. J.  Withers and  P.  Xiao: Corros. Sci., 219 (2023), 111238. https://doi.org/10.1016/j.corsci.2023.111238
• 25)   S.  Kondo,  Y.  Hara,  M.  Sugata and  M.  Tutiya: Tetsu-to-Hagané, 53 (1967), 724 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.53.7_724
• 26)   M.  Inouye and  Y.  Iguchi: Tetsu-to-Hagané, 53 (1967), 1181 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.53.11_1181