Rate Constants of the Initial Reduction of a Single Iron Ore Pellet by CO-H₂ Gas Mixture

Abstract

As the global community moves toward carbon neutrality by 2050, the iron and steel industries are facing significant pressure to reduce CO₂ emissions. Direct reduction processes using hydrogen-rich gases are emerging as a critical alternative to traditional blast furnace methods. This study investigates the reduction behavior of single hematite pellets using varying H₂-CO gas mixtures. Thermogravimetric analysis (TGA) was employed to measure reduction rates, while exhaust gas analysis helped elucidate the reaction mechanism. The results indicate that while higher temperatures and H₂ concentrations accelerate reduction, carbon deposition presents a significant challenge for real-time TGA measurements at lower temperatures. By focusing on the initial stage of reduction, this study effectively excluded the influence of carbon formation on the weight change. Under these controlled conditions, a topochemical receding interface model was found to be the most appropriate for determining rate constants and activation energies under different gas mixing ratios. The derived kinetic parameters and the understanding of H₂-CO reduction behavior provide essential fundamental data for optimizing the operating conditions of gas-based direct reduction processes.

1. Introduction

Over the past few decades, the global climate crisis has been highlighted more than ever before. Based on a series of globalized discussions, many actions to limit global temperatures have been carried out. Since the iron- and steel-making sector is one of the sources most responsible for greenhouse gas emissions, a drastic process transformation is urgently required. In the conventional blast furnace process, a vast amount of CO₂ is inevitably produced during the reduction of iron ore by coke, which is difficult to replace entirely within the existing framework.

The most viable strategy to effectively mitigate CO₂ emissions in iron and steelmaking processes involves transitioning from coal-based reductants to hydrogen-rich gas mixtures. Specifically, iron ore direct reduction (DR) processes—whether utilizing shaft furnaces, fluidized bed reactors, or rotary kilns—have significant potential to contribute to the mitigation of CO₂ emissions by utilizing these hydrogen-rich mixtures. The transition from coal-based reductants to hydrogen-rich gas mixtures as a primary reductant may cause significant changes in the kinetics and thermodynamics of iron ore reduction. Since the 1970s, shaft type DR processes such as Midrex and HYL III have been developed to use syngas produced by reforming natural gas [1,2]. Following the commercialization of gas-based direct reduction processes, the effect of gas composition on the reduction behavior of iron ore pellets has been extensively investigated at both laboratory and industrial scales. For instance, Towhidi and Szekely [3] demonstrated that a higher hydrogen fraction in the reducing gas significantly accelerates reduction kinetics owing to its superior diffusivity. Beyond gas composition, other operational parameters—such as inlet temperature and gas-to-solid ratio—have also been considered in comprehensive mathematical models to simulate heat and mass transfer within shaft furnaces [4]. Additionally, the macroscopic heat balance during industrial-scale reduction has been thoroughly analyzed under various gas pressures and compositions [5,6].

There have been a large number of studies to clarify the mechanisms and kinetics of gas-based reduction in single iron ore pellets. According to the reduction experiments of iron ore pellets by Li and Zhang, the reduction reaction is initially governed by a gas–solid chemical reaction, whereas gas diffusion through the reacted dense layer becomes the rate-controlling step during the later stages of the reduction [7]. To account for this transition, several mathematical models incorporating the unreacted core model and the effective diffusivity of each gas component have been proposed [8,9].

Another critical issue in the gas-based reduction of iron ore by CO-rich gas is the concurrent carbon deposition onto the pellets. Since most experimental kinetic studies rely on thermogravimetric analysis (TGA), the weight gain from carbon deposition can cause significant inaccuracies in measuring the net weight loss associated with reduction. From an industrial perspective, however, the carburization of reduced iron must also be carefully considered, as it provides valuable chemical energy for the subsequent melting process. Generally, it has been found that the addition of hydrogen enhances not only the reduction efficiency but also carbon deposition, up to a certain hydrogen concentration at relatively low temperatures [10,11,12]. Furthermore, the influence of carbon deposition under CO-rich gas conditions on the microstructure and the physical properties of reduced pellets has also been investigated [13].

However, despite these experimental investigations and extensive modeling, a better understanding of the micro-kinetics of a single pellet, particularly regarding complex side reactions like carbon deposition in a highly concentrated CO atmosphere, remains necessary. Therefore, to enable accurate numerical predictions and optimization of large-scale shaft furnace processes, it is essential to rigorously determine the kinetic constants at the single-pellet level. In the present study, reduction experiments of single hematite pellets using thermogravimetric analysis (TGA) were carried out. Specifically, the explicit objectives of this research are to (1) evaluate the reduction behavior under varying H₂-CO mixture gases while excluding carbon deposition interference, (2) identify the appropriate kinetic model for the initial reduction, and (3) determine rate constants and the apparent activation energies.

2. Theoretical Background

The reduction of a single iron ore pellet is a complex physicochemical process comprising several sequential steps: external mass transfer of the reducing gas, inward gas diffusion through the reacted product (ash) layer, interfacial chemical reactions, phase nucleation and growth, and the outward diffusion of product gases. Understanding these micro-kinetics at the single-pellet level is a fundamental requirement for optimizing large-scale direct reduction processes and enhancing the accuracy of macroscopic simulation models. This is particularly critical in H₂-CO mixed gas atmospheres, where the distinct diffusivities, reducing potentials, and reaction enthalpy changes of each component cause highly complex reduction behaviors compared to pure gas conditions. Consequently, precisely determining the reduction rate constants across various temperatures and H₂-CO ratios is essential for robust process design.

To quantitatively evaluate these kinetics, the unreacted core topochemical model is widely employed. The overall reduction process can be broken down into several sequential steps: (1) mass transfer of reducing gas to the pellet surface, (2) inward diffusion of gas through the reacted layer, (3) chemical reaction at the interface between the unreacted core and the product layer, (4) nucleation and growth of the newly produced phase, (5) outward diffusion of the product gas (CO₂ or H₂O). The rate-controlling step is identified from the specific integral kinetic function, $g\left(x\right)$, which exhibits the best linearity with the reaction time, t. These specific integral kinetic functions have been well established according to pre-existing kinetic models. If the overall process is controlled by the chemical reaction at the interface, for instance, the function is expressed as $g\left(x\right)=1-{\left(1-x\right)}^{1/3}$ On the other hand, if intraparticle gas diffusion through the reacted layer dictates the reduction rate, the model follows $g\left(x\right)=3-2x-3{\left(1-x\right)}^{2/3}$. Once the appropriate kinetic model is selected, the reaction rate can be expressed by the product of the kinetic energy function $f\left(x\right)$ and the rate constant $k$, and its temperature dependence is evaluated using the Arrhenius equation to calculate the activation energy of the reduction reaction of iron ore.

3. Experimental

The overall methodology framework of this study, from material preparation to the final kinetic modeling, is illustrated in Figure 1. Commercial hematite pellets were used as the raw material for the reduction experiments. The pellets had an average diameter of 15.2 mm with a standard deviation of 1.37 mm, and their porosity was measured to be about 28% by mercury porosimetry. As shown in

Table 1. Chemical composition of hematite pellets.

	Fe2O3	FeO	SiO2	Al2O3	CaO	MgO	T-Fe	O
wt%	94.1	0.3	2.6	0.4	2.5	0.1	65.6	28.26

, the specimens exhibited a typical chemical composition for hematite pellets, with an oxygen concentration of 28.26 wt%. For the determination of the kinetic rate constants, the apparent density of the pellets was calculated to be approximately 2400 kg/m³.

The reduction experiments were carried out using a thermogravimetric analysis (TGA) setup, as described in Figure 2. An electronic balance was installed above a vertical electrical resistance furnace. A Kanthal suspension wire was attached to the bottom of the balance, with its lower end positioned in the isothermal zone of the furnace. To minimize unwanted vibration and signal noise during weight measurement, the wire was shielded by an alumina protective tube. At the end of the wire, a single pellet was suspended in a spiral-shaped holder, which was specifically designed to prevent the blockage of the upward gas stream from the bottom of the furnace. The reducing gas mixture was prepared by controlling the flow rates of individual pure gases and blending them through a static mixer. The total flow rate of the reducing gas was maintained above 1.8 L/min to eliminate the effect of external mass transfer on the overall reaction rate. Furthermore, this sufficient flow rate facilitated the continuous sampling of the exhaust gas by a flue gas analyzer (NOVAprime syngas, MRU GmbH, Neckarsulm, Germany). In addition to the main reducing gas stream, a minimal flow of helium gas was introduced into the balance chamber to protect the microbalance from any adverse effects of the hydrogen-containing atmosphere.

For each experimental run, once the pellet was loaded into the holder and the furnace was sealed, the system was heated to the target reduction temperature (1073, 1123, 1173, 1223 or 1273 K) at a heating rate of 7 K/min under an argon (Ar) atmosphere. It was confirmed that any weight loss during this heating period was negligible. Upon reaching the desired isothermal temperature, the H₂-CO reducing gas mixture was introduced to initiate the reduction of iron ore. A systematic experimental design was applied where the temperature and gas mixing ratio were selected as variable parameters to evaluate their kinetic effects, while the total gas flow rate and initial pellet size were strictly fixed to eliminate external mass transfer resistance and geometric variations. The real-time weight change of the pellet was continuously recorded by the electronic balance. After the pre-determined reduction time had elapsed, the gas supply was switched back to the inert Ar gas, and the sample was allowed to cool within the furnace to room temperature. Each measurement was repeated at least twice to ensure reproducibility. Repeated measurements confirmed excellent reproducibility with mostly identical results; a maximum error interval of ~5.7% observed in specific cases was attributed to inherent pellet heterogeneities.

4. Results and Discussion

4.1. Influences of Temperature and Gas Composition

Isothermal reduction experiments of a single hematite pellet were conducted at various temperatures and gas compositions. The reduction rate was calculated using Equation (1), and the results of the measurements are summarized in Figure 3.

$$\left(\mathrm{r}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e},\%\right)={\displaystyle \frac{removed O by reduction,\mathrm{g}}{amount of O in a pellet,\mathrm{g}}}\times 100$$

(1)

As shown in Figure 3a, the reduction efficiency significantly improves with higher temperatures and increased H₂ content. Specifically, under a pure H₂ atmosphere (fraction = 1.0), the reduction reached approximately 100% within one hour, regardless of the experimental temperature. Similar dependencies on temperature and gas mixing ratio are also evident in Figure 3b.

It is widely recognized that H₂ gas shows superior reduction efficiency for iron ore compared to CO gas, primarily because of its higher diffusivity—approximately four times that of CO. In general, gas diffusion within porous media depends on pore size as well as the mean free path of the diffusing gas. Bulk diffusion is the dominant mechanism only when the pore size is sufficiently larger than the mean free path. Conversely, if the pore size is on a smaller or comparable scale to the mean free path, gas molecules collide more frequently with the pore walls than with each other, thereby restricting diffusion. Therefore, it is essential to compare the initial pore size in the iron ore pellets with the mean free paths of H₂ and CO gases under the given experimental conditions. Figure 4 shows the mercury porosimetry results of a hematite pellet, revealing that the majority of the pores in the pellet are concentrated in the size range of 1.0~10 μm (10⁻⁶~10⁻⁵ m). Meanwhile, the mean free path of a specific gas can be theoretically computed using Equation (2):

$$\lambda ={\displaystyle \frac{k_{B}T}{\sqrt{2}\pi d^{2}P}}$$

(2)

where $k_B$ and $d$ represent the Boltzmann constant and the collision diameter of the gas molecule, respectively. In addition, $T$ and $P$ denote the absolute temperature and the pressure, specifically 1 atm for the present estimation. The calculated mean free paths for H₂ and CO correspond to 0.47 μm and 0.28 μm, respectively, which are smaller than the pore size in the hematite pellets. Consequently, bulk diffusion can be applied for both H₂ and CO gases within the hematite pellets. Since the pores in the pellets will grow due to the removal of oxygen and vacancy formation, the condition for bulk diffusion can be maintained during the initial reduction stage, before a dense iron phase is sufficiently formed at the outer surface.

Beyond these intrinsic diffusion characteristics, an apparent decreasing tendency in the reduction rate was observed during the reaction. This tendency was more noticeable at lower temperatures and higher CO fractions, although the presence of H₂ still significantly influenced the overall behavior. This reversal phenomenon in the reduction curve actually indicates a net weight gain of the sample during the experiment, which is attributed to carbon deposition. In fact, this was confirmed by the observation of powdery carbon on the surface of the pellet and around the holder after the experiments. Furthermore, subsequent compositional analysis revealed that the carbon content of the pellet reduced at 1073 K exceeded 4.0 wt%.

4.2. Influences of Carbon Deposition

Previous studies have also demonstrated that carbon deposition from CO gas can be accelerated by the catalytic effects of metallic iron, the syngas atmosphere or both [14,15]. This catalytic carburization phenomenon has been widely utilized in various industrial applications, such as blast furnace processes and carbon black production. Considering the inverse Boudouard reaction for solid carbon formation, as shown in Equation (3), the equilibrium activity of solid carbon, $a_{\mathrm{C}}$ at a specific temperature can be estimated using the partial pressure ratio of the off-gas.

$$2\mathrm{C}\mathrm{O}\left(\mathrm{g}\right)=\mathrm{C}{\mathrm{O}}_2\left(\mathrm{g}\right)+\mathrm{C}\left(\mathrm{s}\right)$$

(3)

As an example, the typical changes in the ${p_{\mathrm{C}\mathrm{O}}}^2/p_{\mathrm{C}{\mathrm{O}}_2}$ ratio, calculated from the analyzed gas composition, are presented in Figure 5. As the reduction takes place, the relative amount of CO over CO₂ rapidly increases. Eventually, it crosses the threshold of $a_{\mathrm{C}}=1$ indicated by the red dashed line, creating a highly favorable condition for solid carbon formation. This thermodynamic shift is significantly more predominant at a higher CO concentration, where the equilibrium activity of solid carbon ratio rises much faster and reaches a higher value compared to the 50:50 condition. Furthermore, this carbon deposition is particularly favored at lower temperatures because the inverse Boudouard reaction is highly exothermic.

Despite these thermodynamic considerations, it is practically impossible to identify the exact onset of carbon formation during the experiments. Consequently, separating the weight gain caused by carbon deposition from the net weight loss to obtain the reduction rate of iron ore is highly challenging. This concurrent nature remains the primary obstacle in accurately measuring the true reduction kinetics and determining the rate-controlling step for hematite pellets.

4.3. Rate-Controlling Step of Initial Reduction in a Hematite Pellet

Detecting the exact onset of carbon deposition in real time is fundamentally limited, which could introduce uncertainty into kinetic calculations. To exclude the interfering effect of carbon deposition, the kinetic analysis was selectively focused on the early stage of the reduction (~1000 s). During this initial period, no carbon formation was observed, and the carbon content in the pellets reduced for 1000 s was found to be no more than 0.025 wt%. Therefore, it was reasonably deduced that any weight gain caused by carbon deposition was negligible, ensuring sufficient reliability of the kinetic analysis during this early stage. Based on this initial reduction rate, which directly corresponds to the net weight loss, the rate-controlling step could be determined using the unreacted core model.

As mentioned in the previous section, the rate-controlling step can be determined by identifying the suitable specific integral kinetic function, $g\left(x\right)$ of this reduction process. According to Heidari et al. [16], most previous studies utilizing the unreacted core model have established that the overall reduction rate is primarily governed by either the chemical reaction at the interface or gas diffusion through the product layer [17,18,19,20,21,22,23,24,25]. In the present study, therefore, these two potential rate-controlling mechanisms were considered along with the case controlled by both two rate-controlling steps, i.e., mixed control. The specific integral kinetic function, $g\left(x\right)$ for each rate-controlling step as a determinator can be expressed as follows.

$$\mathbf{Topochemical} \mathbf{reaction} \mathbf{at} \mathbf{the} \mathbf{interface}\mathbf{:} g\left(x\right)=1-{\left(1-x\right)}^{1/3}$$

$$\mathbf{Gas} \mathbf{diffusion} \mathbf{through} \mathbf{the} \mathbf{product} \mathbf{layer}\mathbf{:} g\left(x\right)=3-2x-3{\left(1-x\right)}^{2/3}$$

$$\mathbf{Mixed} \mathbf{control}\mathbf{:}g\left(x\right)=\left\{1-{\left(1-x\right)}^{1/3}\right\}+{\displaystyle \frac{A{r}_0}{C_0-C_{eq.}}}\left\{3-2x-3{\left(1-x\right)}^{2/3}\right\}$$

Here, $C_0$ and $C_{eq.}$ stand for the concentration of reducing gas at the outer surface of a pellet and that at the reaction interface, respectively, while $r_0$ is designated to be the initial diameter of a pellet, and $A$ is a constant, whose value was adopted from the reference [8].

Based on the experimental results obtained under various temperatures and gas mixing ratios, the specific integral kinetic functions, $g\left(x\right)$ of each rate-controlling mechanism were plotted as a function of time, as shown in Figure 6. The coefficients of determination (R²) are tabulated in

Table 2. Coefficients of determination (R²) of the specific integral kinetic functions for the evaluated kinetic models (The highest R² values are presented in bold).

Temperature	H2:CO Ratio	Kinetic Models
		Topochemical	Gas Diffusion	Mixed Control
1073 K	100:0	99.86%	97.17%	89.08%
	75:25	99.98%	96.31%	97.55%
	50:50	99.92%	96.63%	97.99%
	25:75	99.91%	96.29%	98.68%
	0:100	99.74%	96.91%	98.92%
1173 K	100:0	99.90%	97.60%	96.18%
	75:25	99.97%	96.78%	97.04%
	50:50	99.89%	96.82%	97.99%
	25:75	99.60%	97.92%	97.53%
	0:100	99.29%	98.19%	97.71%
1273 K	100:0	98.36%	99.50%	93.01%
	75:25	99.26%	99.15%	94.33%
	50:50	99.53%	98.45%	96.22%
	25:75	99.51%	98.48%	96.25%
	0:100	99.42%	98.01%	97.90%

. It is evident that the function corresponding to the topochemical reaction at the interface exhibits the best linearity and the highest R² value compared to other kinetic mechanisms. This verifies that the early stage of hematite pellet reduction in most experimental conditions is governed by the topochemical reaction at the receding interface. However, in certain cases, the kinetic plot deviates from linearity after approximately 600 s, and this tendency is more pronounced at higher reducing temperatures and increased hydrogen fractions in the gas mixture. Considering that the reduction efficiency is enhanced under these highly reducing conditions, the inward growth of dense reduced iron from the outer surface likely promotes the transition in the rate-controlling step from the topochemical reaction to another kinetic mechanism, which should be gas diffusion through the product layer. The formation of this dense product layer was confirmed through SEM analysis of a pellet reduced for 1000 s, as shown in Figure 7.

To visually confirm that the initial reduction of the hematite pellets is primarily controlled by the topochemical reaction rate at the interface, the reduction kinetics were investigated through interrupted reduction experiments. The pellets were reduced in a 50 vol% H₂–50 vol% CO gas mixture at 1073 K and interrupted at pre-determined times (180, 240, 300, 360, and 420 s). At each specified time, the reducing gas was switched to an inert gas, and the pellet was rapidly withdrawn and quenched under an Ar flow. The partially reduced pellets were then diametrically cross-sectioned for macroscopic observation. Under a light optical microscope, the outer reacted layer and the inner unreacted core were clearly distinguishable by their color and texture, which is a distinct characteristic of the topochemical reaction model at a receding interface. As shown in the cross-sectional macrographs in Figure 8, the area of the unreacted core outlined by the white dashed lines gradually shrinks as the reduction proceeds. Further analysis for phase identification showed that the product layer is wüstite, while the inner core remained unreacted hematite.

4.4. Rate Constant of the Reduction in a Hematite Pellet and Its Temperature Dependence

Since it was proven that the initial stage of reduction of the hematite pellets is primarily controlled by the topochemical reaction rate at the interface, the rate constant k can be obtained from the slope of the $g\left(x\right)$ versus $t$ plot. Mckewan suggested a mathematical kinetic model, assuming that the reduction rate depends on the interfacial area of the product layer and the unreacted core, as expressed in Equation (4) [26].

$$r_0{d}_0\left\{1-{\left(1-x\right)}^{1/3}\right\}=kt$$

(4)

Here, $r_0$ and $d_0$ denote the radius of a pellet before the reduction and the density of oxygen (kg/m³), respectively. Consequently, the rate constant k can be calculated from the values of the slope in Figure 6 and the measurable properties of the pellets. The evaluation results of the rate constant of reduction are tabulated in

Table 3. Rate constant (kg/m²∙s) of the initial reduction in a hematite pellet at various temperatures and gas mixing ratio.

H2:CO Ratio	1073 K	1123 K	1173 K	1223 K	1273 K
100:0	11.9 × 10−4	13.1 × 10−4	14.3 × 10−4	16.0 × 10−4	18.1 × 10−4
75:25	10.7 × 10−4	11.3 × 10−4	12.1 × 10−4	12.9 × 10−4	16.0 × 10−4
50:50	8.21 × 10−4	9.12 × 10−4	9.27 × 10−4	9.71 × 10−4	12.1 × 10−4
25:75	6.17 × 10−4	6.77 × 10−4	7.07 × 10−4	8.27 × 10−4	9.59 × 10−4
0:100	4.69 × 10−4	5.03 × 10−4	5.72 × 10−4	6.50 × 10−4	6.71 × 10−4

. Here, it should be noted that the degree of reduction was at least 36%, and no trace of a residual hematite or magnetite core was observed in the pellets, even under the least favorable conditions after 1000 s of reduction. Since the present reduction conditions are more than sufficient to complete the reduction step of Fe₃O₄ → FeO, the present rate constant corresponds to the apparent rate constant for the overall reduction, which includes the sequential reduction steps of Fe₂O₃ → Fe₃O₄ and Fe₃O₄ → FeO. The magnitude of the apparent rate constant for the gaseous reduction in the hematite pellet is found to be in the range of 0.5~2.0 × 10⁻³ (kg/m²∙s), which agrees well with previously reported values under similar experimental conditions [27,28,29]. As expected from the previous discussion, it is clear that the reduction rate constant increases as the temperature rises and the hydrogen concentration in the gas mixture increases.

Based on the temperature dependence of the rate constant, the apparent activation energy can be derived using the Arrhenius equation. Figure 9 presents not only the relationships of the rate constant but also the Arrhenius plot, illustrating the relationship between the natural logarithm of the rate constant and the reciprocal of the absolute temperature. Mathematically, the slope of this linear plot corresponds to $-E_a/R$, where $E_a$ is the activation energy and $R$ is the universal gas constant. The apparent activation energies evaluated for the reduction in the hematite pellets are summarized in

Table 4. Apparent activation energies for the reduction in the hematite pellets.

H2:CO Ratio	Activation Energy (J/mol)
100:0	23,631.7
75:25	21,146.8
50:50	20,095.5
25:75	24,358.2
0:100	22,201.0

. It should be noted that the activation energy does not vary significantly with the gas mixing ratio, suggesting that the fundamental mechanism of the chemical reaction remains similar regardless of the specific H₂:CO proportion. Moreover, the activation energies obtained in the present study are in the range of 20~24 kJ/mol, which are relatively lower compared to the previously reported data summarized in

Table 5. Comparative analysis of kinetics and activation energy for the reduction of iron oxides under H₂- and CO-based atmospheres.

Raw Materials	Temperature (K)	Reducing Gas	Particle Size (µm)	Activation Energy (kJ/mol)	Rate Controlling Step	Final Product Phase	Source
Hematite (Fe2O3)	1073~1473	Pure CO	9000	64	Chemical reaction at metal-oxide interface	Metallic iron (Fe)	[28]
Calcium Ferrites	1173~1373	Pure CO	10,000 ~10,800	61.1 ~69.5
Hematite (Dense)	973~1373	100~0% H2 + CO 0~100%	9800	31.6 ~53.6	Chemically controlled & Mixed control	α-Fe, Fe3C	[18]
Hematite (Porous)	973~1373	100~0% H2 + CO 0~100%	10,800	9.5 ~21.5	Gaseous diffusion controlled
Hematite (sintered pellets)	573~773	H2 (0.65 atm)	13,000	30.1	Surface chemisorption	Fe3O4	[30]
Hematite (Fe2O3) briquette	973~1373	Pure CO	d8000 h8000	15.0 ~28.9	Mixed control (gaseous diffusion/interfacial reaction)	Metallic iron (Fe)	[31]
Hematite (Fe2O3) briquette	1073~1223	Pure CO	17,800	19.8	Mixed control (interfacial reaction→ gaseous diffusion)	Metallic iron (Fe)	[32]
Hematite (Fe2O3) briquette	1073~1223	Pure H2	17,800	42.1
Hematite (Fe2O3)	1073~1173	Pure CO	1.2, 58.6	10.0 ~14.7	Reaction controlled kinetics	Metallic iron (Fe)	[33]
Hematite (α-Fe2O3)	493~953	Pure H2	1~2	75.9	Nucleation and Growth (2- and 3D) & Phase-boundary-controlled reaction	Fe3O4	[12]
Hematite (α-Fe2O3)	493~953	10% H2 + 90% N2	1~2	94.8
Hematite (α-Fe2O3)	493~953	Pure CO	1~2	114.1
Magnetite (Fe3O4)	493~953	Pure H2	1~2	39 ~88	Nucleation and Growth (2- and 3D) & Phase-boundary-controlled reaction	Metallic iron (α-Fe)	[12]
Magnetite (Fe3O4)	493~953	10% H2 +90% N2	1~2	35.9 ~103
Magnetite (Fe3O4)	493~953	Pure CO	1~2	40.3 ~114.2		α-Fe, Fe3C
Magnetite (Ore fines)	973~1173	H2 (0.5~1 atm)	75~180	11~33	Pore diffusion	Metallic iron (Fe)	[34]
Magnetite (MSPS *)	1073~1273	Pure H2	63~75	47.2	Phase boundary controlled	Wüstite (FeO)	[35]
Wüstite (MSPS *)	1073~1273	Pure H2	63~75	29.7		Metallic iron (Fe)	[35]
Wüstite (FeO)	1173~1373	0~100% H2 + CO	50~150	53.8 ~134.0	Interfacial chemical reaction	Metallic iron (Fe)	[36]
Wüstite (SiO2-doped)	1173~1373	0~100% H2 + CO	50~150	28.3 ~58.4
Hematite (Fe2O3)	823~1573	Pure H2 /Pure CO	0.74 (mean)	13.5 (H2), 2.3 (CO)	Phase-boundary-controlled reaction	Metallic iron (Fe)	[37]
Hematite (iron ore pellets)	1033~1273	75% H2 + 25% N2	10,000 ~12,500	41.0	Interfacial chemical reaction	Metallic iron (Fe)	[38]
Magnetite (Fe3O4)	773~823	5~20% H2 + Ar	0.03	40.1 ~50.6	1-dimensional nucleation and growth	Metallic iron (Fe)	[39]
Wüstite (FeO)	973~1123	5~20% H2 + Ar	0.03	23.9 ~29.0
Hematite (Fe2O3)	1123 ~ 1273	H2 + CO +10% Ar	9900~10,100	47.7	Spherical Shrinking Model	Metallic iron (Fe)	[40]

*: Magnetically stabilized porous structure.

.

Evidently, the reported activation energies for the reduction of iron oxide vary widely—from a few kJ/mol to over 100 kJ/mol—depending on the initial properties of the iron oxide, the reducing atmosphere, and the rate-controlling step considered. Some studies focusing on the stepwise reduction of iron oxide have shown that the Fe₂O₃ → Fe₃O₄ reduction step has higher activation energy than the Fe₃O₄ → FeO reduction step. As sufficiently explained in the previous section, however, the activation energy obtained in the present study is an apparent value including these sequential steps. Further examination of Table 5 indicates that the activation energy tends to be higher when the particle size is smaller and the rate-controlling step is the chemical reaction at the receding interface, rather than gas diffusion through the product layer.

For a more appropriate comparison with the present findings for hematite pellets, the following analysis focuses solely on previously reported data for iron oxide pellets or briquettes larger than 5 mm. In these cases, the activation energy for the reduction rate constant is substantially lower than that for the reduction in powdery iron oxide, mainly because gas diffusion through the product layer becomes more dominant in controlling the overall rate compared to the chemical reaction at the interface [37]. In fact, many studies utilizing large particles have proposed mixed control involving the gas diffusion through the product layer and the chemical reaction at the interface as the rate-controlling step [18,30,31,32]. This mixed control can be understood as a transition of the rate-controlling step from the chemical reaction at the interface to the gas diffusion through the product layer, as the outer product layer, which is considerably dense in many cases, grows inward during the reduction process [31,32]. Therefore, the present results, which suggested that the reduction rate of hematite pellets at the initial stage (~1000 s) is controlled by the chemical reactions at the interface are well consistent with those previous studies. As is well known, H₂ has far higher diffusivity and transports even in the reduction environment much faster than CO gas. This difference results in a higher reduction rate for H₂ gas compared to CO gas reduction. When the overall rate is controlled by the interfacial chemical reaction, however, the influence of the gas composition on the reduction rate and the activation energy should be less significant, which perfectly aligns with the independence of the activation energy from the gas atmosphere observed in the present study.

Furthermore, El-Geassy et al. investigated the effect of porosity on the reduction kinetics and the activation energy [11]. It was demonstrated that the activation energy is lower when the iron oxide is more porous and less dense, which enables easier gas transport. It is also noteworthy that fired briquettes before the reduction in the study by Moon et al. had approximately 22% porosity, and the activation energy for the reduction was no more than 30 kJ/mol [31]. Considering the comparatively high porosity of the hematite pellets in the present study, the correspondingly low value of the activation energy is highly comprehensible.

5. Summary

In this study, the initial reduction kinetics of single hematite pellets using H₂-CO gas mixtures were systematically investigated through thermogravimetric analysis (TGA) at 1073–1273 K. Due to the concurrent catalytic carbon deposition under high CO concentration and low temperatures, measuring the precise net weight loss for reduction was highly challenging. To prevent this interference, the kinetic evaluation was focused on the initial reduction stage (within 1000 s), yielding the following principal findings:

(1)

The macroscopic reduction rate significantly accelerated with elevated temperatures and increased H₂ fractions, inherently benefiting from the superior diffusivity of H₂. Based on the mean free path evaluations, bulk diffusion was confirmed as the dominant gas transport mechanism within the porous pellets.

(2)

Kinetic modeling coupled with cross-sectional macroscopic observations verified that the initial reduction stage is primarily governed by the topochemical reaction at the receding interface. This mechanism effectively incorporates the sequential intermediate phase transformations from hematite to wüstite.

(3)

The apparent rate constants, $k$ for the overall initial reduction were derived in the range of 0.47~1.81 × 10⁻³ kg/m²·s. Accordingly, the apparent activation energies ($E_a$) were calculated to range from 20.0 to 24.3 kJ/mol.

(4)

Notably, the apparent activation energy remained largely independent of the H₂:CO gas mixing ratio, strongly indicating that the fundamental interfacial chemical reaction dictates the overall rate regardless of the specific gas atmosphere. Furthermore, the relatively low activation energies are physically well-justified by the high initial porosity of the hematite pellets, which facilitates easier intraparticle gas transport.

Practically, these quantitative kinetic parameters and mechanistic insights provide essential knowledge for optimizing gas compositions and operational conditions in the transition toward hydrogen-based direct reduction processes for iron ore. Since this study only focuses on the initial reduction stage to avoid carbon formation, future research should explore methods to evaluate the later-stage reduction kinetics and to independently determine the kinetic parameters for the FeO → Fe reduction step.