Reduction and Nitridation of Iron/Vanadium Oxides by Ammonia Gas: Mechanism and Preparation of FeV45N Alloy

Abstract

The steel micro-alloyed with ferrovanadium nitride has extremely superior properties that make it widely utilized in structural components, construction and aircraft. The conventional methods for synthesizing ferrovanadium nitride include nitridation of pure ferrovanadium alloy or carbothermal nitridation of metallic oxides, using nitrogen or ammonia gas as nitrogen sources. In this study, ferrovanadium nitride (FeV45N) was prepared by direct reduction and nitridation of the corresponding metal oxides with ammonia as the reductant and nitrogen source. This method avoids the introduction of other impurity elements, except the negligible trace elements accompanied with the raw materials. The thermodynamics of the reduction and nitridation process were initially analyzed. During the subsequent ammonia reduction process, the FeV45N powders were successfully obtained at 1273 K for 6 h. The obtained powders were pressed into cylindrical briquettes by hot pressing (HP) at 1473 K for 1 h in vacuum. In the investigation, the X-ray diffraction and morphological analysis of the products was also carried out, and the reaction mechanisms were discussed in detail. The nitrogen content of the final product can reach 11.85 wt. %, and the residual oxygen content can be reduced to 0.25 wt. %. By sintering, the density of the alloy can reach 5.92 g/cm³.

1. Introduction

Recently, high-strength low-alloy (HSLA) steel has attracted much attention in industry due to its superior properties and economic benefits [1,2,3]. The superior performance of HSLA steel mainly accounts for microalloying with carbide- and nitride-forming metals (V, Nb, Al, Ti, etc.). The advantages of nitrogen are related to grain refinement, austenite stabilization, formation of nitrides and carbonitrides with extremely high structural hardness and significant increasing of strength, without restriction of ductility [3,4]. Among these microalloying additives, HSLA steel microalloyed with vanadium and nitrogen exhibits particular economic superiority.

During steel making, vanadium can be added by means of V-bearing compounds—ferrovanadium (FeV), vanadium carbide (VC), vanadium silicide (V₃Si), etc. However, the drawback of these additives is the absence of nitrogen. Although nitrogen is capable of being introduced into the melt by blowing N-bearing gases, the final composition of the melt is generally corrected by adding N–bearing alloys. Up to now, various alloys, including vanadium nitride (VN), vanadium carbonitride (V(C, N)) and ferrovanadium nitride (FeVN), have been proposed for the combined introduction of vanadium and nitrogen into the molten steel [5,6,7,8,9]. Vanadium carbonitride has been researched for many years; however, its practical use is inhibited by its low briquette density (≈3 g/cm³) and high melting point (>2673 K). Compared with vanadium nitride, ferrovanadium nitride has better affinity and higher density (>5 g/cm³), which makes it more easily absorbed by molten steel and can increase the recovery of vanadium and nitrogen, thereby reducing the production cost and improving the mechanical properties of the steel [10,11].

Traditionally, ferrovanadium nitride could be synthesized by direct nitridation of ferrovanadium alloy, carbothermal reduction–nitridation or hydrogen-based reduction-nitridation of corresponding metallic oxides, using nitrogen or ammonia gas as nitrogen sources. The current methods are classified as solid nitridation, liquid nitridation, sintering processes, plasma smelting and carbothermal reduction–nitridation [1,12,13,14,15,16,17,18,19,20,21]. Grishchenko et al. have synthesized ferrovanadium nitride using ferrovanadium as raw material via roasting at 1327 K in a vacuum furnace for about 17 h [12]. Nevertheless, the nitrogen content of the final product was 5.0–6.6% and was characterized as heterogeneous. Franke et al. conducted similar experiments, whose results also proved that the N content could not meet the requirements of industrial production [13]. The study by Krastev indicated that the nitrogen content could be improved with the increase of temperature and reaction time, and by using a finer particle size [14]. Preparation of ferrovanadium nitride by liquid nitridation greatly shortened the reaction time [15]. However, the nitridation of liquid alloys has not been industrially adopted due to the high viscosity of nitrides-bearing ferrovanadium melts, which resulted in low nitrogen content in the melts. Ziatdinov et al. also synthesized ferrovanadium nitride from ferrovanadium powders by the self-propagation high-temperature synthesis (SHS) method [16]. Ferrovanadium nitride can also be prepared by the plasma smelting method using ferrovanadium alloys under a nitrogen atmosphere [17]. With respect to the above four methods, ferrovanadium was used as the raw material, which not only contained impurities (Al, Si) but also consumed relatively more money and energy. Carbothermal reduction–nitridation of iron and vanadium oxides has also been investigated to prepare ferrovanadium nitride [19,20]. During the process, although Al, Si impurities were not introduced, there was still some residual carbon in the final product, and the roasting temperature was very high (more than 1823 K). Excessive reducing agent was always added to enhance the deoxidation process, which inevitably increased the residual carbon content and correspondingly decreased the nitrogen content in the product.

In recent decades, transition metal nitrides (TiN, VN, CrN, etc.) have been prepared from corresponding oxides using ammonia as nitrogen source and reducing agent [22,23,24,25,26,27,28,29]. Ammonia gas is proved to be more effective in the nitridation process than high-purity N₂ gas [30]. In view of that superiority, ferrovanadium nitride was also expected to be synthesized by reduction and nitridation of a mixture of vanadium and iron oxides [31]. In the present study, ferrovanadium nitride (FeV45N) was prepared by the specified process with ammonia gas. A systematic study on the thermodynamics of the NH₃-VO₂-Fe₂O₃ system was performed first. The phase transformation of vanadium and iron dioxides was studied using X-ray diffraction. Subsequently, the obtained powders were sintered into bulk ferrovanadium nitride by hot pressing (HP). Morphological analysis of the product was also carried out, allowing the study to be concluded with a discussion of the applicable reaction mechanisms. Compared with the carbothermal reduction nitridation method, this method, using ammonia as the reductant and nitrogen source, avoided the introduction of carbon. Furthermore, the final product contained a relatively high nitrogen content and a low residual oxygen content.

2. Experimental

2.1. Experimental Procedure

Vanadium dioxide powder (Tiantian Chemical Reagent Company, Jiangsu, China, VO₂, >99.7 wt. %) and Fe₂O₃ (Aladdin Industrial Corporation, Shanghai, China, >99.7 wt. %) were used as raw materials to synthesis FeV45N. The particle sizes of these two materials were less than 74 μm. Theoretically, the iron and vanadium contents in FeV45N are about 45 wt. % and 45 wt. %, respectively. Thus, in order to synthesize FeV45N, 200 mg of Fe₂O₃ and 220 mg of VO₂ were uniformly mixed and loaded onto an ultra-fine stainless steel mesh, which was supported by a corundum crucible. The steel mesh was used to obtain a porous structure, to ensure that ammonia gas was able to flow through the bottom of the anatase powder. Before the experiment, argon gas (99.999 vol.%) flow of 150 mL/min was introduced into an electric tube furnace to expel oxygen and moisture. This gas flow was maintained until the furnace was heated to the required temperature, through steps of 10 K/min. The crucible loading with the mixture was then placed in the roasting zone of the horizontal furnace tube, and the gas was switched to NH₃ (99.999 vol.%), which was also maintained at 150 mL/min. The sample was reduced and nitrided isothermally at temperatures of 873–1273 K for 1–6 h. Once roasting was completed, the furnace was allowed to cool down under flowing argon gas. After that, the obtained powders were pressed into a cylindrical briquette (10 mm in diameter, 4 mm in thickness) by hot pressing (HP) at 1473 K for 1 h in vacuum.

2.2. Analytical Method

Thermodynamic analysis was conducted by using FactSage 6.2 based on the pure substances database [32,33]. After reaction, the cooled powders were examined by X-ray powder diffraction (Rigaku D/max 2500, Rigaku, Tokyo, Japan) under the conditions as follows: Cu Kα, tube current and voltage: 250 mA, 40 kV, scanning range: 10–90° (2θ), step size: 0.02° (2θ) and scanning speed: 8°/min. The oxygen and nitrogen contents of powders was tested by the German Eltra On–900 (German Eltra, Shanghai, China) N/O elemental analyzer. The sample was heated and fused in an ink crucible. The oxygen in the sample reacted with carbon to form carbon monoxide, and nitrogen was released in the elemental state. Scanning electron microscopy (JSM–7800F, Tokyo, Japan) was carried out to monitor overall powder morphology, and the microstructure of bulk ferrovanadium nitride. The density of the FeV45N briquette was measured using the Archimedes method (GB/T1423–1996).

3. Thermodynamic Analysis

In this work, NH₃ contributed simultaneously as the reduction and the nitridation agent to react with the mixture of Fe₂O₃ and VO₂. To predict the predominant field of various products, the equilibrium calculations of the Fe₂O₃–NH₃ system and the VO₂–NH₃ system were performed, respectively. Figure 1 shows the solid phase fields calculated in the temperature range of 673–1573 K at a pressure of 1 bar. The input NH₃ mole fraction was given as n_NH3/(n_NH3+n_Fe2O3) and n_NH3/(n_NH3+n_VO2), where n_NH3, n_Fe2O3 and n_VO2 represent the input mole numbers of NH₃, Fe₂O₃ and VO₂, respectively.

As shown in Figure 1a, the solid phases, including Fe₂O₃, Fe₃O₄, FeO and Fe, were mainly dependent on temperature and input NH₃ fraction. The condition for the occurrence of Fe was that the temperature was about 823 K and the input NH₃ fraction was about 0.5. Thermodynamically, Fe₂O₃ can be easily reduced even at very low temperature and NH₃ fraction. The reaction path of Fe₂O₃ varied as Fe₂O₃→Fe₃O₄→FeO→Fe or Fe₂O₃→Fe₃O₄→Fe, which depended on the temperature. It is worthy to note that iron nitride was not observed in Figure 1a under the calculated conditions. As shown in Figure 1b, VO₂ can be easily reduced to V₂O₃ at a temperature above 973 K, regardless of NH₃ fraction. When the temperature was lower than 973 K, some V₃O₅ would have appeared and then been reduced to V₂O₃. It meant the reaction path of VO₂ followed VO₂→V₂O₃→VN or VO₂→V₃O₅→V₂O₃→VN at a different temperature range. However, the NH₃ fraction required for reduction and nitridation of V₂O₃ was relatively higher, which should be greater than 0.98. Thus, compared with Figure 1a, when Fe₂O₃ was converted to Fe, VO₂ oxide was converted to VN. At the same time, the generated iron can promote the decomposition of ammonia and affect the formation of vanadium nitride [34].

In order to predict the phase composition of NH₃–Fe₂O₃–VO₂ system, the solid phase composition as a function of input NH₃ mole fraction at 1273 K was calculated. The calculation was carried out at a pressure of 1 bar by fixing Fe₂O₃ at 1 mol and VO₂ at 2.2 mol. The input NH₃ mole fraction was given as n_NH3/(n_NH3+n_Solid), where n_Solid represents the input mole numbers of Fe₂O₃ and VO₂. The results in Figure 2 show the absolute mole numbers of solid phases during the reaction process. It can be seen from Figure 2 that Fe₂O₃ undergoes the transformation process of forming Fe step by step. The NH₃ fraction required for generating Fe is around 0.78. However, the reduction and nitridation of VO₂ to form VN required NH₃ fractions to reach above 1, which was much higher than the formation of Fe. It can be seen from Equations (1) and (2) that, at 1273 K, the ΔG^Θ of Fe generated by FeO reduction was far less than the ΔG^Θ needed for VN formation. From the perspective of thermodynamics, the reaction condition for VN generation was extremely harsh.

2NH₃ + FeO = Fe + H₂O + N₂ + 2H₂   ΔG^Θ = 123.42 − 0.2405T (KJ/mol)

(1)

2NH₃ + V₂O₃ = 2VN + 3H₂O   ΔG^Θ = 144.00 − 0.1452T (KJ/mol)

(2)

4. Results and Discussion

4.1. Phase Evolution during Reduction with Ammonia

The phase compositions of the reduced and nitrided products were investigated, and the results are shown in Figure 3. Reduction roasting was conducted as a function of temperature (4 h) and duration (1273 K).

Figure 3a shows the phase composition with temperature in the range of 873–1273 K for 4 h, respectively. As can be seen from the Figure 3a, three different phases, including VN, Fe and Fe₄N, existed in the roasted samples. It indicated that VO₂ and Fe₂O₃ have already been completely reduced or nitrided after 4 h. When the reaction temperature was 873 K, VN, Fe and Fe₄N were present in the sample. The diffraction peaks were still assigned to VN, Fe and Fe₄N with an increase of reaction temperature to 1073 K. However, when the reaction temperature increased to 1173 K, the diffraction peak of Fe₄N disappeared. This is because Fe₄N is unstable at high temperature and would decompose to Fe and N₂. As can be seen from Equations (3) and (4), under standard conditions, the initial temperatures for reactions (3) and (4) are 595 K and 429 K, respectively. It can be deduced that Fe₄N is easily converted to Fe under the experimental conditions. Meanwhile, as can be seen from Figure 3a, the relative strength of the diffraction peak of Fe₄N was relatively low, which also indicated that Fe₄N was unstable in ammonia atmosphere.

2NH₃ + Fe₂O₃ = 1/2Fe₄N + 3H₂O + 3/4N₂    ΔG^Θ = 172.01 − 0.2890T (KJ/mol)

(3)

2NH₃ + Fe₄N = 4Fe + 3/2N₂ + 3H₂   ΔG^Θ = 123.68 − 0.2884T (KJ/mol)

(4)

The phase transformation of VO₂ and Fe₂O₃ as a function of duration is also compared. Figure 3b shows the X-ray diffraction (XRD) patterns of samples roasted for different hours at 1273 K. It was found that metallic Fe had already been formed after reacting for 1 h. At the same time, VO₂ has also been reduced. Both V₂O₃ and VN were observed after roasting for 1 h, indicating VO₂ was not completely converted to VN. This is because Fe₂O₃ was easier to be reduced than VO₂ under an ammonia atmosphere. The diffraction peak of Fe₄N was not present because the formed iron nitride was unstable at high temperature. As the duration increased to 4 h and 6 h, only Fe and VN were observed. The results of phase transformation were in agreement with the thermodynamic analysis to some extent. As can be seen from Figure 2, Fe₂O₃ was initially reduced and metallic iron can coexist with V₂O₃. The phase of V₂O₃ will be finally reduced and nitride into VN.

4.2. Morphological Analysis

The micromorphologies of the products obtained under different reaction conditions were investigated by scanning electron microscopy (SEM), and the results are plotted in Figure 4. It can be seen from Figure 4a that when reducing at 1073 K for 4 h, the powder particles of the reduced samples were still dispersive. The bright phase and the gray phase were mainly metallic Fe and VN, respectively, and a sintering phenomenon was negligible at 1073 K. The morphology feature of VN was more obvious due to its face centered cubic crystal structure. By reducing at 1273 K for 6 h, the powder particles of the samples accumulated to a larger size and showed an obvious sintering phenomenon, as shown in Figure 4b and the energy dispersive spectrometer (EDS) results of points 3 and 4.

4.3. Characterization of the Products

In order to determine whether a prepared sample meets commercial application standards, the sample needs to be tested for oxygen and nitrogen contents.

Table 1. O and N contents of FeV45N.

Temperature (K)	Reaction time (h)	O content (wt. %)	N content (wt. %)
1073	6	1.16	11.74
1173	4	1.11	11.75
1273	4	0.74	11.79
1273	6	0.25	11.85

shows the experimental results of nitrogen and oxygen measurements of FeV45N prepared under different conditions. It can be seen from Table 1 that as the reaction time and the reaction temperature increased, the O content of the product decreased and the nitrogen content increased, correspondingly. After reacting at 1273 K for 6 h, the O content was as low as 0.25%, and the N content was increased to 11.85%, which meets the requirement of FeV45N that the elemental content of N is in the range of 9–12% (National Standard of China, GB/T 30896-2014, Beijing, China). Meanwhile, no impurities are introduced into the powder, such as C, Al or S. Additionally, compared to the products obtained by solid or liquid nitridation methods, whose N contents were less than 10% and whose density was 4.0–4.5 g/cm³, the N content and density in the present were obviously higher [14,15]. The FeVN alloy obtained by high-temperature synthesis method also had high density (5.5–6.5 g/cm³), while the N content was in the range of 9–11% [16].

4.4. Microstructure of the Alloy after Sintering

In order to increase the density of ferrovanadium nitride alloy, the obtained powders after reduction and nitridation were pressed into a bulk alloy (cylindrical briquette, 10 mm in diameter, 4 mm in thickness) by hot pressing (HP) at 1473 K for 1 h in vacuum. The density of the FeV45N alloy after sintering was determined as 5.92 g/cm³, which was beneficial to increasing its utilization efficiency during the steel making process. The SEM micrograph and its corresponding area scanning results of the bulk alloy are presented in Figure 5. It can be seen that there were two main phases in the alloy, named metallic iron (Fe) and vanadium nitride (VN), respectively. There was an obvious boundary between the metallic iron and VN phases. The area scanning results also demonstrated that vanadium mainly existed in the form of VN, because the region of vanadium overlapped with that of nitrogen.

4.5. Reaction Mechanism

Although synthesizing vanadium nitride by ammonia reduction has been investigated [30,35]. The study by Vaidhyanathan et al. proved that high-purity N₂ gas was less effective in the nitridation process than ammonia gas [30]. It is difficult to form vanadium nitride by reacting vanadium oxide with N₂ and H₂. Thermodynamically, NH₃ would easily decompose to form N₂ and H₂ upon heating. However, the behavior of NH₃ was rarely studied, and the components decomposed from ammonia gas were the actual reaction agents which would contact the raw materials. The actual H₂ concentration in off-gas was detected by gas analyzer and the results are shown in Figure 6. It can be observed that the H₂ concentration increased with an increase in temperature, and it remained almost unchanged after 1373 K. The equilibrium H₂ concentration increased from 12.5 vol.% to 53.5 vol.%, which revealed that the pyrolysis of ammonia gas was enhanced by increasing the temperature. However, the experimental pyrolysis ratio was still less than the theoretical value, which can be reflected by lnK. The theoretical and experimental values of lnK can be calculated according to Equations (6) and (7), respectively. The difference in the pyrolysis ratio may be attributed to the kinetic factors. Based on the above results, it can be deduced that ammonia gas was not completely decomposed, even at temperatures above 1373 K, leaving the possibility of reacting by NH₃. A wide range of molecular or radical N-containing intermediates of ammonia pyrolysis, such as N, NH, NH₂ and N₂H_x (x = 1–4) may also be formed for nitridation process [27].

2NH₃(g) → N₂(g) + 3H₂(g)   ΔG^Θ = 150340 − 230.6T (J/mol)

(5)

$$\ln K=–\frac{\mathsf{\Delta }{G}^{\mathsf{\Theta }}}{RT}=–\frac{18082.75}{T}+27.74$$

(6)

$$\ln K=\ln \frac{P_{N_2}^{}\times P_{H_2}^3}{P_{N{H}_3}^2}$$

(7)

where K is the equilibrium constant; P_N2, P_H2 and P_NH3 are the partial pressures of N₂, H₂ and NH₃, respectively.

Thus, the reaction mechanisms can be described by the schematic diagram illustrated in Figure 7. NH₃ may also react with VO₂ and Fe₂O₃, even though it was easily decomposed during heating. When the temperature was lower than 1073 K, NH₃ reacted with Fe₂O₃ to form Fe and Fe₄N directly. The phase transformation during the entire reaction process proceeded in the following order: Fe₂O₃→Fe→Fe₄N. However, when the temperature was higher than 1073 K, the transition path of Fe₂O₃ was Fe₂O₃→Fe. Fe₄N disappeared at higher temperatures. The difference may be ascribed to the kinetic factors. During the reaction process, the phase transformation of VO₂ follows the order: VO₂→V₂O₃→VN. According to the above results, it can be concluded that ammonia gas is an effective reduction and nitridation agent for preparing FeVN alloy with iron and vanadium oxides as the raw material. Although FeV45N alloy has been prepared and the reaction mechanisms were preliminarily revealed in this paper, there are still lots of things that need to be studied further, such as the effect of precursor, gas composition, etc., on the preparation of FeVN alloy. Moreover, scaling up experiments are still on the way for future industrial production on the basis of the studied process.

5. Conclusions

This study describes the synthesis of FeV45N alloy by the reduction and nitridation of iron, and vanadium oxides mixture with ammonia gas. Thermodynamically, it was feasible to form metallic iron and vanadium nitride using ammonia gas as the reduction and nitridation agent. The experimental results indicated that the reaction temperature and time had significant effects on the reduction and nitridation process. With the increase of reaction temperature and time, the oxygen content was decreased gradually in the product. The maximum nitrogen and minimum oxygen content of FeV45N reached 11.85% and 0.25%, respectively, by roasting at 1273 K for 6 h. During the reaction process, the phase transformation of VO₂ follows the order: VO₂→V₂O₃→VN. When the temperature was lower than 1073 K, NH₃ reacted with Fe₂O₃ to form Fe and Fe₄N directly. Fe could not form Fe₄N when the temperature was over 1073 K. After sintering the composite powders at 1473 K for 1 h, it is possible to obtain a bulk of FeV45N alloy with a density of up to 5.92 g/cm³.