Materials Transactions, Vol. 45, No. 6 (2004) pp. 1907 to 1910 #2004 The Japan Institute of Metals
Reduction of Iron Oxides by Nano-Sized Graphite Particles Observed in Pre-Oxidized Iron Carbide at Temperatures around 873 K Masaaki Hisa1 , Atsushi Tsutsumi2 and Tomohiro Akiyama1; * 1 2
Department of Chemical Engineering, Graduate School of Engineering, Osaka Prefecture University, Sakai 599-8531, Japan Department of Chemical System Engineering, School of Engineering, The University of Tokyo, Tokyo 113-8656, Japan
Iron oxides containing nano-sized graphite particles were prepared from porous iron carbide and reaction characteristics of the oxides/ graphite composite mixture were investigated by thermogravimetry in combination with microstructure analysis using X-ray diﬀractometry and optical microscopy. A drastic mass reduction was observed in a temperature range between 833 K and 1023 K, and the loss of mass reached 2.8%. The mass reduction was found to correspond to a growth of Fe and FeO at the expense of Fe3 O4 and Fe2 O3 in the composite mixture, suggesting that the oxides were reduced by graphite. The oxides-graphite reaction is presumably indirect reduction caused by CO and H2 that are generated through the water gas reaction, C + H2 O ! CO + H2 . (Received March 2, 2004; Accepted April 19, 2004) Keywords: iron carbide, iron oxide, graphite, nano-structure, indirect reduction, thermogravimetry, microscopy, X-ray diﬀractometry
examined by ex situ microstructure characterization using X-ray diﬀractometry and optical microscopy.
It has previously been reported by Akiyama et al. and Ito et al. that iron carbide can be recognized as a possible hydrogen source available at temperatures around 573 K when reacting with steam.1,2) Hisa et al. have thereafter revealed using materialographic techniques that hydrogen generation at such low temperatures is attributed to oxidation of iron carbide by steam, which is represented by the following equation:3) Fe3 C þ 4H2 O ! Fe3 O4 þ 4H2 þ C:
Carbon formed through the reaction was found to remain in the iron carbide in the form of nano-sized graphite particles that are dispersed uniformly in the oxide matrix. For the microstructure of this kind, iron oxides are potentially reduced by graphite when exposed to high temperatures. The reduction kinetics of iron oxides by ﬁne graphite particles was investigated in detail over wide temperature ranges with mechanically milled hematite/ graphite mixtures by several researchers.4,5) It was revealed that microstructure reﬁnement by milling greatly accelerated the reduction process and, at the same time, decreased the initiation temperature of the reaction. In addition, Khaki et al. reported that the eﬀect of structure reﬁnement was more noticeable for graphite than for hematite.5) Thus, it would be expected that the graphite in the oxidized iron carbide may exhibit good reaction characteristics in the reduction of iron oxides because of its nano-sized nature, and in addition to hydrogen, metallic iron can possibly be collected from iron carbide at low temperatures. In the present study, porous iron carbide produced directly from iron ore was pre-oxidized in an argon atmosphere containing water to form an iron oxide/graphite composite structure, and the reaction characteristics of the composite was subsequently analyzed by thermogravimetry. A possibility of iron formation through the reaction between iron oxides and nano-sized graphite in the composite was *Corresponding
author, E-mail: [email protected]
The iron carbide used in the present study was prepared by a two-step ﬂuidized bed method.6) Chemical analysis revealed that the iron carbide contained around 90 mass% of Fe3 C, as well as metallic iron and iron oxides (mainly Fe3 O4 ) as impurities. After sieved mechanically with a 300 mm stainless steel sieve, about 150 mg of iron carbide powder was weighed out and put in a cylindrical sample pan made of alumina for heat treatment and thermogravimetry (TG) using a simultaneous thermal analyzer. Following drying treatment at 393 K for 3.6 ks, two consecutive heat treatments were applied to the iron carbide powder. Firstly the powder was heated up to 673 K at a constant rate of 10 Kmin1 and kept at the temperature for 86.4 ks for oxidation treatment. Then, immediately after furnace-cooled down to 298 K, the oxidation-treated specimen was continuously heated up to 1073 K at 10 Kmin1 for TG analysis. All procedures were conducted under a ﬂowing argon atmosphere containing 4.6 ppm of water, and the ﬂow rate of argon gas was set at 50 cm3 min1 . The heat treated specimens were subjected to powder X-ray diﬀractometry (XRD) for phase identiﬁcation with an X-ray powder diﬀractometer equipped with a rotational Cu target (tube voltage = 50 kV, tube current = 300 mA) and a graphite monochromator. The microstructure of the iron carbide specimens was investigated by cross-sectional microscopy. After molded in polyester resin, they were polished mechanically with a conventional polishing technique and examined under an optical microscope. 3.
Results and Discussion
Figure 1 shows the time evolution of sample mass during the oxidation treatment at 673 K. While the heat treatment proceeded, the mass of the iron carbide specimen increased
M. Hisa, A. Tsutsumi and T. Akiyama
Fig. 1 Time evolution of the relative mass and temperature of the iron carbide during an oxidation treatment in a ﬂowing argon atmosphere containing 4.6 ppm of water. The specimen was heated up to 673 K at a constant rate of 10 Kmin1 and then kept at the temperature for 86.4 ks, followed by furnace-cooling down to 298 K.
monotonically up to around 28.8 ks, beyond which the rate of mass increase gradually decreased with reaction time. Since the mass increase at 673 K has been found to result from oxidation of iron carbide by steam,3) two consecutive oxidation mechanisms were likely to operate in the iron carbide-steam reaction at this temperature. Cross-sectional microscopy conﬁrmed that oxidation proceeded to a large extent in the iron carbide specimen during the prolonged oxidation treatment. As shown in Fig. 2(b), the microstructure of the oxidation-treated iron carbide particle was almost entirely occupied by a dark-toned phase, which was identiﬁed as iron oxides containing nano-sized particles of graphite. Unreacted Fe3 C, appearing bright in the micrograph, was dispersively present, but its volume fraction was small compared to that of the oxide phase. Furthermore, comparison between Figs. 2(a) and (b) revealed that the oxidation treatment made the microstructure of the iron carbide particles less porous; ﬁne cracks and pores contained in raw particles were scarcely observed in oxidation-treated particles. Such structural densiﬁcation presumably stems from a volume expansion due to the phase transformation from Fe3 C to Fe3 O4 and might correlate with the deceleration of oxidation seen in Fig. 1. The result of TG analysis for the oxidation-treated iron carbide specimen is presented in Fig. 3(a). The TG curve was almost ﬂat up to 723 K, suggesting that further oxidation rarely took place in the pre-oxidized specimen. Then, after taking a small increase (0:1%) perhaps due to the oxidation of residual Fe3 C, the relative mass of the specimen started to decrease drastically at around 833 K. This temperature was found to be approximately 20 K higher than the initiation temperature for Stage II in the iron carbide-steam reaction.3) The mass decrease continued until the temperature reached 1023 K, and afterward the sample mass turned increasing again. The amount of mass reduction occurred between 833 K and 1023 K was measured to be 2.8%. Figure 3(b) indicates a characteristic microstructure of the iron carbide specimen after furnace-cooled down to 298 K, following the
50 µm Fig. 2 Optical micrographs showing typical cross-sectional microstructures of the iron carbide specimens: (a) raw, and (b) heat treated isothermally at 673 K for 86.4 ks (oxidation treatment).
TG analysis. It was found that a bright phase with a coarse granular microstructure was present inside the particles although its distribution varied in each particle. The bright phase showed a metallic gloss under an optical microscope and was thought to be a reduction product of the graphitecontaining iron oxide. Phase analysis by XRD supported the results of crosssectional microscopy. As Fig. 4(b) indicates, Fe3 O4 and Fe2 O3 were the major phases in the oxidation-treated specimen, suggesting that Fe3 C contained in the raw specimen (Fig. 4(a)) was largely oxidized by steam. After a continuous heating to 1073 K was applied on the preoxidized specimen, reﬂections from Fe3 O4 and Fe2 O3 were weakened and, in contrast, diﬀraction peaks that belong to Fe and FeO increased their intensity, as shown in Fig. 4(c). The result suggests that the reduction of Fe3 O4 and Fe2 O3 led to the formation of Fe and FeO in the specimen. The mass reduction occurred at 833 K was, therefore, primarily attributed to this reaction and perhaps included a contribution from the decomposition of the residual Fe3 C that brought on the co-production of Fe and H2 .1) Based on the above-mentioned results, the reduction process were assumed to be described possibly by the following equations:7) (x/4)Fe3 O4 þ C ! (3x/4)Fe þ COx ; xFe3 O4 þ C ! (3x)FeO þ COx
ðx ¼ 1 or 2Þ:
Reduction of Iron Oxides by Nano-Sized Graphite in Oxidized Iron Carbide at 873 K
Fe FeO Fe3O4
Temperature, T / K
Scattering angle, 2 θ
Fig. 3 (a) Thermogravimetry curve of the pre-oxidized iron carbide specimen obtained through continuous heating up to 1073 K at a constant rate of 10 Kmin1 , and (b) optical micrograph showing a corresponding microstructure of the specimen after furnace-cooled down to 298 K. The bright phase at the center of the particle is likely to be a reduction product of the pre-oxidized iron carbide, and the surrounding dark phase seems to be formed as a result of re-oxidization of the bright phase. The specimen was oxidation-treated at 673 K for 86.4 ks prior to the continuous heating experiment.
However, it was calculated that, starting with 1 mol of Fe3 C (refer to the reaction (1)), any of those reactions should accompany more than 9% of mass reduction at completion, which is much higher than the experimentally measured value, 2.8%. This suggests that, despite the fact that graphite was still present in the specimen (Fig. 4(c)), the reduction of the iron oxides was somehow terminated without completion before the re-oxidation (i.e. mass increase) of the specimen became dominant at 1023 K. The reason for the incomplete termination remains unclear, but one possibility is probably that a relative amount of carbon was consumed by reacting directly with water in the argon gas to form carbon monoxide; i.e. C þ H2 O ! CO þ H2 :
Normalized Intenisity (Arbitrary Unit)
Relative Mass (%)
Because this is a gas-solid reaction, it should only be induced by graphite particles included in areas adjacent to the external and internal surfaces of the pre-oxidized iron carbide particles. Therefore, if the formation of Fe and FeO from Fe3 O4 is supposed to occur as a result of indirect reduction of Fe3 O4 by CO generated through the reaction (4), then a limited amount of graphite can only contribute to the reaction, resulting in a low reduction yield of Fe3 O4 . The
Fig. 4 XRD spectra of the iron carbide specimens: (a) raw, (b) heat treated at 673 K for 86.4 ks (oxidation treatment), and (c) heated continuously up to 1073 K at a constant rate of 10 Kmin1 , following the oxidation treatment. The intensity of X-ray is normalized by the maximum intensity in each spectrum.
eqs. (2) and (3) were thus modiﬁed so as to include H2 O as a reactant, and the reaction equations for the reduction taking place in the pre-oxidized iron carbide were accordingly proposed as follows. (1/3)Fe3 O4 þ C þ (2/3)H2 O ! Fe þ (2/3)H2 þ CO2 ;
Fe3 O4 þ C þ H2 O ! 3FeO þ H2 þ CO2 :
It is also anticipated that hydrogen generated through the reaction (4) should contribute to the reduction of iron oxides: for example, (1/4)Fe3 O4 þ H2 ! (3/4)Fe þ H2 O:
If H2 O in the above reaction successively causes the shift reaction with CO, namely CO þ H2 O ! CO2 þ H2 ;
then the overall reaction becomes identical to the reaction (5). Consequently, the eqs. (5) and (6) can be applied to the iron oxide/hydrogen reaction. Further clariﬁcation of the behavior of carbon in the iron carbide would be a key to a full understanding of the mechanism of iron formation from the graphite-containing iron oxides derived from the porous iron carbide. To achieve the goal, it is essential to trace carbon precisely throughout the reactions by means of composition analysis (e.g. gas
M. Hisa, A. Tsutsumi and T. Akiyama
chromatography and X-ray spectroscopy) in combination with thermal analysis and microscopy.
(1/3)Fe3 O4 þ C þ (2/3)H2 O
Fe3 O4 þ C þ H2 O ! 3FeO þ H2 þ CO2 :
The reaction characteristics of iron oxides containing nano-sized graphite particles were analyzed by thermogravimetry at a constant heating rate of 10 Kmin1 . The iron oxides were derived from porous iron carbide by isothermal heat treatment (oxidation treatment) at 673 K for 86.4 ks in a ﬂowing argon atmosphere containing 4.6 ppm of water. Xray diﬀractometry and optical microscopy were employed to examine the phase and microstructure evolutions brought into the specimens by the heating processes. The results are summarized as follows. (1) During the oxidation treatment sample mass increased monotonically up to 28.8 ks, and then mass increase decelerated, suggesting that the oxidation of the porous iron carbide is controlled by two consecutive mechanisms. (2) A drastic mass reduction occurred in a temperature range between 833 K and 1023 K in the thermogravimetry curve for the pre-oxidized iron carbide. The amount of mass reduction was measured to be 2.8%. (3) The mass reduction was conﬁrmed to occur as a consequence of the reduction of Fe3 O4 and Fe2 O3 to Fe and FeO, and the reaction was likely to be indirect reduction caused by nano-sized graphite with a help of water contained in the argon gas. (4) The equations that describe the reduction process were proposed as follows:
! Fe þ (2/3)H2 þ CO2 ;
Acknowledgements The authors wish to thank Professor Munetake Satoh of Osaka Prefecture University for his assistance in specimen preparation for microscopy. New Energy and Industrial Technology Development Organization of Japan is appreciated for the provision of a grant aid ‘‘The project of Fundamental Technology Development for Energy Conservation’’ that supported a part of the present investigation. REFERENCES 1) T. Akiyama, A. Miyazaki, H. Nakanishi, M. Hisa and A. Tsutsumi: Int. J. Hydrogen Energy 29 (2004) 721–724. 2) T. Ito, A. Tsutsumi and T. Akiyama: J. Chem. Eng. Jpn. 36 (2003) 881– 886. 3) M. Hisa, A. Tsutsumi and T. Akiyama: Mater. Trans. 45 (2004) 1911– 1914. 4) E. Kasai, K. Mae and F. Saito: ISIJ Int. 35 (1995) 1444–1451. 5) J. V. Khaki, Y. Kashiwaya, K. Ishii and H. Suzuki: ISIJ Int. 42 (2002) 12–22. 6) E. Inoue, Y. Uchiyama, T. Miyashita and J. Nakatani: Japanese Patent No. 9040414 (1997). 7) Although the oxidation-treated iron carbide specimen contained Fe2 O3 as well, the argument here will be limited to Fe3 O4 for simplicity.