Fate of Nitrogen and Sulfur during Reduction Process of Carbon-containing Pellet Prepared by Vapor Deposition of Gaseous-Tar and the Influences of the Hetero Elements on the Reduction Behavior and Crushing Strength

Abstract

Influences of the vapor deposition (VD) atmosphere on the nitrogen/sulfur contents in carbon-containing pellet and the crushing strength during preparation by VD method of coke oven gas (COG) tar against cold-bonded pellet (CP) are first investigated using a flow-type quartz made fixed-bed reactor and a tensile and compression crushing machine. Although N in NH₃ that is fed with simulated-COG components is not transferred into the prepared VD sample, some part of the S in H₂S moves the VD sample; the carbonaceous materials derived from COG tar fill the pores of the CP. However, these elements do not affect the crushing strengths of the prepared VD samples. The N and S forms in the VD sample are then investigated using XRD and XPS, and the results show that these elements mainly exist as organic-N and -S in the VD samples. The fates of N and S in the VD sample during the reduction process are examined using a flow-type fixed-bed reactor under inert (He) and reduction (55%H₂/He) atmospheres. The N species in the samples mainly evolve as NH₃ and N₂ at 400–1000°C, and the cumulative amount of N₂ that evolves is greater than that of NH₃. The H₂S evolution begins at 400°C, and the profile provides the main peak at approximately 800°C. The amount of evolved H₂S in 55%H₂/He is greater than that in He. Although the reduction of the VD sample starts at approximately 400°C and stops at 1000°C, N and S species in the sample do not affect the reduction rate. In addition, the N and S in the VD samples do not influence the crushing strengths during heat treatment.

1. Introduction

In Japan, the iron-steel making industry has three main problems related to energy, resources, and the environment: high energy consumption, depleting resources and/or appreciating resource prices, and a large amount of CO₂ emission, respectively. Thus, technologies that can simultaneously address these three problems are in demand. So far, reducing the thermal reserve zone temperature is expected to reduce CO₂ emissions and/or energy consumption from blast furnaces.^1,2,3) Therefore, the technical development of iron-making materials that are rapidly reducible is very important to solve these problems. However, iron materials with a low reduction disintegration index (RDI) are required for use in blast furnace because the reduction disintegration occurs in the upper region of the blast furnace by the reduction of hematite to magnetite in lump ore/sinters with a volumetric expansion, which causes a loss of permeability in the blast furnace.⁴⁾ In general, lump/sinter iron ores of low RDI value and high crushing strength are believed to show a low reduction reactivity.⁴⁾ Thus, it is very important to develop iron materials with low RDI values and high reduction reactivities for the blast furnace in the iron-making process.

Our research group has been developing an iron/carbon composite whereby the carbonaceous materials derived from tar completely fill the pores of the limonite/CP by VD method of tar or using an impregnation method.^5,6,7,8) In our previous work, we found that these two methods could be used to fill the mesopores in samples with carbonaceous materials derived from tar. In addition, these composites show high crushing strengths and can attain rapid reduction of iron oxides in the composites. Moreover, the strength of the prepared composites remains up to a 50% reduction rate (up to 800°C), which is the temperature range of the reduction disintegration observed at lump/sintered-ores.⁸⁾ If a composite with large amounts of C, a high crushing strength, high reduction reactivity, and low RDI can be prepared gaseous-tar, it may be applicable for blast furnaces in conventional or advanced iron-making processes (COURSE50).^9,10,11) For the above-mentioned composite preparation, it is considered that the utilization of the conventional process is effective. In other words, if COG containing gaseous-tar produced from the cokemaking process and its sensible heat can be utilized for composite preparation, the production can be completed in integrated iron steel plants. However, the fates of N and S were not determined during composite preparation using the VD method, and the N and S species in the composites did not influence the crushing strength and reduction rate of the prepared composite. In particular, it has been reported that sulfur affects the reduction rate of iron oxides during the reduction process.^12,13)

In this study, therefore, we first investigated the effect of the VD atmosphere on N and S contents in the VD samples and on the crushing strengths; then, the fate of N and S in the VD samples were examined during the reduction process. In addition, the influence of N and S against the reduction rate and crushing strength during the reduction process were investigated.

2. Experimental

2.1. Samples

CP and COG tar were supplied by a Japanese iron steel-making company. The details of CP preparation and the composition of the tar are provided elsewhere.^7,8) The composition of the as-received CP with a particle size fraction from 2 to 3.5 mm is as follows: total-Fe; 46, FeO; 0.5, SiO₂; 5.5, Al₂O₃; 2.7, CaO; 3.7, and MgO; 0.2 wt%-dry. The analysis results for the tar are as follows: C; 91, H; 4.7, N; 1.1, and S; 0.4 wt%-daf. The specific Brunauer-Emmet-Teller (BET) surface area (S_BET) and Barrett-Joyner-Halenda (BJH) pore volume (V_BJH) of CP were 20 m²/g and 0.06 cm³/g, respectively.⁷⁾

2.2. Preparation of Carbon-containing CP

Carbon-containing CP was prepared using VD method. The details of this method are reported elsewhere.^8,14) The VD treatment was carried out using a cylindrical flow-type quartz fixed-bed reactor consisting of tar-pyrolysis and VD sections. A mixture of tar and toluene (50 wt%-tar) was first dropped onto the quartz wool placed in the tar-pyrolysis section of the reactor with 200 mL-STP/min-He, 55%H₂/He, simulated COG (55%H₂/30%CH₄/5%CO/3%CO₂/3%H₂O/He), 3000 ppmv-H₂S/COG or 1 vol%-NH₃/COG, and the mixture was pyrolyzed at 700°C. Here, the above-mentioned pyrolysis atmospheres are denoted as He, H₂, COG, H₂S/COG, and NH₃/COG, respectively. The feed rate of the mixture was 0.4 mL/min. The gaseous-tar-containing pyrolysis gases generated in the tar-pyrolysis section were then loaded into the VD section; this section was packed with 3.0 g of CP, which was placed at the bottom of the reactor. VD of tar into CP was carried out at 350°C for 240 min.^8,14) Here, the tar-pyrolysis and VD temperatures used in the present study were optimized in previous work.^8,14)

2.3. Reduction

The reduction-reactivity of the VD samples was investigated using a flow-type fixed-bed reactor.^5,6,7,8) The sample was heated at 10°C/min up to 1000°C for 60 min in He or 55% H₂/He. The latter atmosphere was used for simulated COG, which is amplified H₂ by COG modification and will be used in a hydrogen reduction blast furnace in the COURSE50 project.⁸⁾ The N₂, CO, CO₂, H₂O, HCN, and NH₃ that evolved during the above heat treatment were measured on-line using a micro-gas chromatograph (Inficon) and an auto-focusing multi gas monitor (Innova) at intervals of 2 and 3 min, respectively. H₂S in the outlet gas from the reactor was measured using a standard gas detector tube (Gastec) at intervals of 5 min. The extent of reduction was estimated based on the amount of O-containing gases that evolved and the amount of O in the feed sample.^5,6,7,8)

2.4. Characterization

Samples before and after VD or reduction were characterized using an HCN-coder (Yanaco), CS analyzer (Horiba), powder X-ray diffractometer (XRD, Shimadzu), a measurement system for N₂ adsorption (Quantachrome), scanning-electron microscopy equipped with energy-dispersive X-ray spectroscopy (SEM-EDS, Joel), a Raman spectrometer (Renishaw), and a tensile and compression testing machine (Minebea) following the Japanese Industrial Standard (JIS M8718) method. The details of these analyses are described in the literature.^5,6,7,8,13) The Fe2p_3/2, N1s, and S2p spectra were measured using a X-ray photoelectron spectroscopy (XPS, Joel) using a Mg-Kα X-ray source to clarify the chemical-forms of Fe, N, and S in VD samples. The prepared VD samples were made into fine powders before XPS analysis and mounted onto a sample holder with Ag paste. The binding energies of the Fe2p_3/2, N1s, and S2p spectra were referred to the Ag3d_5/2 peak at 367.9 eV. Least-squares curve fitting of N1s spectra was performed using Gaussian peak shapes.¹⁵⁾ Upon deconvolution, the binding energy of each peak and the full width at half-maximum were fixed within 0.1 eV and only the amplitude was varied to obtain the optimum curve resolution. The XPS measurement for each sample was repeated at least twice, and each spectrum was deconvoluted.

3. Results and Discussion

3.1. Changes in Pore Distribution in the Samples before and after VD Treatment

Figure 1(a) shows the N₂ isothermal curves of the as-received CP, CP heated at 350°C in He before VD-treatment (denoted as He/350), and VD sample prepared in COG. As-received CP and He/350 showed type-IV N₂ adsorption curves according to the IUPAC classification, and hysteresis was observed in both samples. This shape of the adsorption curve indicates the existence of mesopores in the samples.¹⁶⁾ In addition, the observed shape hysteresis is distinguished as H3-type (IUPAC classification) and suggests the existence of slit-shaped pores in the samples. Although the N₂ amount adsorbed on He/350 was greater than that on the as-received CP, the amount on the VD sample dramatically decreased and almost reached zero. This result suggests that the pore that developed by heat treatment of as-received CP is almost filled completely with carbonaceous materials derived from tar during VD the treatment. Similar tendencies have been observed for VD samples prepared in He, H₂, H₂S/COG, and NH₃/COG.

Fig. 1.

N₂ isothermal adsorption curves (a) and pore size distributions (b) of as-received CP, CP heated at 350°C in He and VD-sample prepared in COG.

Figure 1(b) presents the changes in the pore size distribution and pore volume calculated by the BJH method from the results in Fig. 1(a). Although the pore volume was calculated based on the result of Fig. 1(a), the result until 10 nm is shown in Fig. 1(b) to clarify the change in pore size distribution. The pore size distribution profile with a peak at approximately 2 nm observed for the as-received CP increased on He/350, and the pore volume also increased. However, mesopores almost disappeared completely for the VD sample. The S_BET and V_BJH values of the as-received CP, He/350, and VD sample were 20, 60, and < 1 m²/g and 0.06, 0.07, and < 0.01 cm³/g (pore volume calculated based on the results up to a 50 nm of pore size distribution), respectively. Similar results were observed for the pore size distribution, and decreasing S_BET and V_BJH values were observed for VD samples prepared under other atmospheres. When He/350 was treated in COG at 350°C for 180 min, the pore properties of He/350 were retained. Therefore, the COG-composition (CH₄ decomposition and/or C deposition derived from the CO disproportionation reaction) does not affect the changing pore properties under the present conditions. The above results show carbonaceous materials derived from tar deposits in the mesopores of the CP sample. According to our previous research, carbonaceous materials derived from gaseous tar produced by tar pyrolysis are deposited into the mesopores in CP via Knudsen diffusion.⁸⁾ In addition, it has been found that carbonaceous materials derived from gaseous tar exists inside in the VD samples with uniformity from the results of SEM-EDS against the cross section of particle.⁸⁾

3.2. Chemical Forms of Constituent Elements of the Prepare VD Samples

Figure 2 shows the XRD patterns of samples before and after the VD treatment (corresponding the results in Fig. 1). The peak attributed to α-FeOOH observed in the as-received CP disappeared in He/350, whereas diffraction signals of Fe₂O₃ were detected. This transformation occurs from the de-hydrate of α-FeOOH, and mesopores can develop in above the temperature range of 300–350°C.¹⁶⁾ On the other hand, Fe₂O₃ observed in He/350 was reduced to Fe₃O₄ in the VD samples. Similar results were observed for the VD samples prepared in H₂, NH₃/COG, and H₂S/COG. When the VD-sample was prepared in He, peaks attributed to Fe₂O₃ and Fe₃O₄ were observed. These results show that some or all of Fe₂O₃ or all of Fe₂O₃ in He/350 before the VD treatment was reduced by the pyrolysis gas or VD atmosphere (H₂, COG, NH₃/COG, and H₂S/COG). However, Fe–N or Fe–S compounds were not found in the XRD patterns of the VD samples prepared under any of the atmospheres.

Fig. 2.

XRD patterns of (a) as-received-CP, (b) He/350, and (c) VD-sample prepared in COG.

Table 1 lists the elemental analyses, Fe forms, and crushing strengths of the VD samples prepared under various atmospheres, and the N and S contents in Table 1 are compared in Fig. 3. The N content in the VD sample prepared in NH₃/COG was 0.4 wt%-dry, and similar values (0.34–0.38 wt%-dry) were observed in VD sample prepared under other atmospheres. On the other hand, the S content in the VD sample prepared in H₂S/COG was 0.56 wt%-dry, whereas that in He, H₂, COG, and NH₃/COG ranged from 0.18–0.22 wt%-dry. In other words, some H₂S was transferred into the solid phase during VD preparation in H₂S/COG, and the value was approximately three-times greater than that of the other VD samples. However, as mentioned above, the Fe–S compound was not detected in the VD sample prepared in H₂S/COG in the XRD measurement. This may be because the S content in the VD sample is very small, and the penetration depth of the X-rays and the minimum limit of the crystal phase detection by the XRD measurements are 0.01–20 μm and 0.5%, respectively. According to previous reports, char (carbonaceous material) produced from coal pyrolysis and/or carbonization reacts with H₂S, which exists under a pyrolysis atmosphere and/or is derived from pyrite decomposition during coal pyrolysis; the newly organic S is formed in the solid phase.¹⁷⁾ Therefore, increasing the S content in the VD sample may cause a secondary reaction (gas-solid reaction) between the H₂S fed and C deposition on the VD sample. However, it may be possible for the gas-gas reaction between gaseous tar and the H₂S fed to occur; the details of this process will be the subject of future work.

Table 1. Summaries of elemental analyses, iron forms and crushing strength of VD samples prepared in various atmosphere.

Preparation atmosphere of VD samples	Elemental analyses, wt%-dry					Iron formb	Crushing strength, daN
	C	H	N	S	O+Fea
In He	18.5	0.82	0.34	0.22	80.1	Fe2O3 (w), Fe3O4 (w)	10
In 50% H2/He	19.1	0.83	0.35	0.18	79.5	Fe3O4 (m)	10
In COG	19.2	0.76	0.36	0.19	79.5	Fe3O4 (m)	10
In 1 vol% NH3/COG	20.3	0.78	0.40	0.20	78.3	Fe3O4 (m)	10
In 3000 ppmv H2S/COG	19.7	0.75	0.38	0.56	78.6	Fe3O4 (m)	10

^a Estimated by difference. ^b XRD intensities designated by w (weak) and m (medium).

Fig. 3.

N (a) and S (b) contents in VD samples prepared at various atmospheres.

The crushing strength values of the prepared VD samples were almost similar with 10 daN (Table 1). This result shows that the preparation atmosphere and S content do not affect the strength of the composite prepared under the present conditions. When each VD sample was investigated with Raman analysis to clarify the C form, broad spectra with peaks at 1350 and 1600 cm⁻¹ were observed. According to previous work, the peaks at 1350 and 1600 cm⁻¹ are assigned to D-band of amorphous C and G-band of crystallized C, respectively.¹⁸⁾ As is well known, crystallized C forms amorphous C during heat treatment above 1000°C.¹⁹⁾ Carbonaceous materials derived from tar pyrolysis are unlikely to crystallize during the VD treatment because the tar pyrolysis and VD temperatures are 700 and 350°C, respectively. Therefore, the peak attributed to G band is likely partly crystallized C. A least-squares curve-fitting analysis of the Raman spectra was carried out using Gaussian peak shapes according to a previous report.¹⁸⁾ The proportions of amorphous C and partly crystallized C were estimated as 40 and 60%, respectively, and carbonaceous material in the VD samples almost thus exists as amorphous C.

Figure 4(a) presents Fe2p_3/2 spectra of the VD sample prepared in COG. The VD sample provided broad spectra with a main peak at 712 eV in the range from 700 to 722 eV. It is well-known that the main peak at around 712 eV is attributed to an Fe–O bond, such as that in Fe₂O₃, Fe₃O₄, or FeO.²⁰⁾ From the XRD results in Fig. 2, the main peak is distributed from Fe₃O₄ in the VD sample. On the other hand, according to previous work, although peaks attributable to α-Fe, Fe–N, FeS₂, and FeS have been observed at 707.5, 706.8, 707.4, and 710.4 eV, respectively, distinct peaks of the former three species were not detected in Fig. 4(a).^20,21,22,23) FeS may exist in the VD sample because the FeS peak was observed at approximately 710.4 eV. Our research group found that α-Fe is produced from limonite reduction and/or limonite reacting with NH₃ and H₂S at 300–800°C and nitrided as Fe_xN and sulfurized as FeS or Fe_1-xS.^24,25) According to this work, the peak attributable to Fe_xN was not found in the XPS spectra, as seen in Fig. 4(a), although Fe_1-xS and Fe_xN should be produced by the reaction between CP and NH₃ and H₂S during VD treatment at 350°C. This fact may show that NH₃ fed or derived from tar pyrolysis is unlikely to react with Fe species in CP. However, H₂S that is fed or derived from tar pyrolysis may react with the Fe species in CP to produce Fe–S species, but the content is very small because Fe_1-xS was not observed in the XRD measurement. Therefore, the main forms of N and S in the VD samples may be organic N and S derived from COG tar. The N1s and S2p XPS analyses were carried out to clarify the details of the N and S forms in the VD samples. Figure 4(b) shows the N1s spectrum of the VD sample prepared in COG. As shown by the solid line in Fig. 4(b), VD sample provided a wide spectrum with main and shoulder peaks at approximately 399 and 400 eV, respectively. Similar spectra were found for VD samples prepared under other atmospheres. According to previous work, pyridinic-, pyrrolic-, and quaternary-N gave N1s bonding energies at 398.7, 400.3, and 401.4 (± 0.1) eV, respectively.^{26,27,28,29,30,31)} Therefore, curve-fitting against the N1s spectrum observed was carried out to quantify the proportion of N forms. The results are showed in Fig. 4(b) by a broken line, and the resulting analysis method relatively well reproduced the spectrum measured. The proportions of N forms were calculated based on the areas and increase according to quaternary-N (27%) < pyrrolic-N (50%) < pyridinic-N (23%). In other words, a large portion of N in the VD sample exists in heterocyclic-structures. Similar tendencies have been observed for other VD samples and Argonne premium coals with 66–87 wt%-dry C.³¹⁾ In the S2p XPS spectrum, the VD sample gave a very weak, broad spectrum with a very small main peak at 164–165 eV in the range from 160–173 eV. It has been reported that FeS, FeS₂, alkly-, or thiophenic-S provides the main peak of S2p at 160.8–161.4, 162.9–163.7, 163.3, or 164.1 eV, respectively.^32,33,34) However, the Fe–S bond energy was not clearly observed at the above ranges because the intensities of the S2p spectra were much smaller than that of N1s and Fe2p_3/2. However, S in the VD samples mainly exists as aromatic S because a very small main peak of S2p XPS spectra was observed in the XPS spectra at 164–165 eV. It has been reported that S species in coal tar mainly exist as dibenzothiophene type.³⁵⁾ Thus, aromatic S may be mainly derived from tar.

Fig. 4.

Fe2p (a) and N1s (b) XPS spectra of VD sample prepared in COG.

3.3. Fate of Nitrogen and Sulfur during the Reduction Process

The reduction rates of the VD sample prepared in COG were first investigated during heat treatment in He and 55%H₂/He to clarify the fate of N and S during the reduction process, and the results are shown in Fig. 5. Here, CO, CO₂ and H₂O indicate the reduction rates derived from those gases, and the total reduction rate shows the reduction rate of the VD sample calculated based on the formation amounts of the above three gases species. The reduction of iron oxides in the VD sample in the He heat treatment started from 400°C and terminated at approximately 1000°C after the reduction rate profile gave the main peak at approximately 700°C. The reduction rate of the O-containing gases increases according to CO₂ < H₂O < CO, and direct reduction is dominated. This occurs from carbonaceous materials derived from tar and iron oxide that are in close contact with the VD sample. CO₂ formation occurs by indirect reduction. On the other hand, hydrogen species in the deposited carbonaceous materials could affect the reduction and resulting H₂O formed by hydrogen reduction. The reduction rate of VD sample under He was greater than that of the mixture of dehydrated-CP at 500°C and pulverized coke.⁸⁾ The 55%H₂/He reduction began at approximately 400°C and almost terminated up to 1000°C after the reduction rate provided the main peak at 800°C. The reduction rates of CO, CO₂ and H₂O increased according to CO₂ << CO < H₂O. However, hydrogen reduction was dominant. In addition, it was suggested that indirect reduction is inhibited because CO₂ formation decreases during the reduction process.

Fig. 5.

Reduction rates and extent of reduction of VD sample prepared in COG with temperature in He (a) and 55%H₂/He (b).

Figure 6 shows the evolution of N-containing gases from the composite prepared in the COG during the heat treatment in He and 55%H₂/He, and the N distribution at each temperature (corresponding with Fig. 6) is listed in Fig. 7. Figure 7 shows that the mass balance of N is within 100 ± 5% in the present work. When the composite was heated in He (Fig. 6(a)), HCN, NH₃, and N₂ evolved from approximately 400 and 600°C, respectively. The former two gave the small peaks at 600–800°C, and the latter provided the large main peak in 800–900°C. The gases evolve at 1000°C for 60 min in the following order: HCN (1.9 N-%) < NH₃ (3.0 N-%) << N₂ (54 N-%), as shown in Fig. 7(a). Figure 6(b) presents the result in 55%H₂/He. Although the evolution behavior of gaseous N species was similar to that in He (Fig. 6(a)), N₂ formation decreased under a reducing atmosphere. In contrast, the main clear peak for the NH₃ formation rate was observed at approximately 600–700°C comparing with that in He, whereas the HCN formation rate profile was the same as that in He. The cumulative amounts of gaseous N species in 55% H₂/He up to 1000°C for 60 min increase according to HCN (1.3 N-%) < NH₃ (4.6 N-%) << N₂ (36 N-%), as seen in Fig. 7(b). Similar tendencies were observed during the heat treatment in He and 55%H₂/He for the composite prepared in NH₃/COG. From these results, the N species in the VD-samples evolved mainly at 500–800°C. In addition, the temperature range for the evolution of gaseous N species corresponds with that of the iron oxide reduction in the composite (Fig. 5). In other words, the release behavior of the N-containing gases and reduction rate were almost similar. Our research group has been investigating the removal behavior of N species from coals during pyrolysis of Fe-loaded coals or the mixtures of limonite and coal.^{15,36,37,38,39)} According to our previous work, N₂ formation (removal extent of char-N) during pyrolysis increases by the addition of iron species, and the temperature range where denitrification is promoted almost corresponds with the temperature range in the present study.^{15,36,37,38,39)} Contact between iron and char-N is likely an important factor for denitrification. Therefore, the evolution amount of N₂ from the VD sample is large because Fe and the carbonaceous material containing heterocyclic-N species in the VD sample are in close contact. Therefore, the evolution of N species in VD samples during the reduction processes may proceed according to a similar mechanism as that in previous works.^{15,36,37,38,39)} According to previous work, hydrogen is present in the atmosphere in large quantities and could react with char-N, producing HCN, which could further react to form HCN or NH₃.⁴⁰⁾ Therefore, it may thus be possible that increasing NH₃ yield in 55%H₂/He occurs by between solid-N in VD sample and hydrogen.

Fig. 6.

Formation rates of N₂, HCN and NH₃ during heat treatment in (a) He and (b) 55%H₂/He from VD sample prepared in COG.

Fig. 7.

Distribution of nitrogen during heat-treatment in (a) He and (b) 55%H₂/He from VD samples prepared in COG.

Figure 8 presents the H₂S evolution from the VD sample prepared in COG (a) and H₂S/COG (b) during the heat treatment in He and 55%H₂/He, and the S distribution at each temperature is shown in Fig. 9. As seen in Fig. 9, mass balance of S was within 100 ± 5% in the present work. When the VD sample was heated in He (Fig. 8(a)), H₂S evolved at around 400°C and gave a weak main peak for the formation rate at 750–800°C. This type of H₂S evolution behavior in He was similar to that in 55%H₂/He (Fig. 8(b)), however, the peak intensity for the formation rate under a reducing atmosphere was greater than under an inert atmosphere. The H₂S yields up to 1000°C for 60 min in He and 55%H₂/He were 3 and 38 S-% (Fig. 9), respectively, and the cumulative amount of H₂S evolved was large under a reducing atmosphere. Similar results were observed for VD samples prepared in He and H₂. Figure 8(b) shows the results for the VD sample prepared in H₂S/COG. The H₂S formation rate in He was similar to that in Fig. 8(a). However, in 55%H₂/He, the peak intensity for the formation rate at around 800°C was greater than that in He, and it became five times greater than that in He. The H₂S yields up to 1000°C for 60 min in He and 55%H₂/He were 1.8 and 60 S-%, respectively. Similar tendencies that increases the amount of evolved H₂S under a H₂ atmosphere have been observed in coal pyrolysis.⁴¹⁾ Therefore, it is suggested that the source of H₂S during the reduction process of the VD sample in 55%H₂/He is organic-S. In addition, such a profile for the H₂S formation rate was similar to the reduction rate in shown in Fig. 5. It was thus found that H₂S evolution promotes a range of iron oxide reduction in the VD samples.

Fig. 8.

Formation rates of H₂S during heat treatment of VD sample prepared in COG (a) and H₂S/COG (b).

Fig. 9.

Distribution of sulfur during heat-treatment in (a) He and (b) 55%H₂/He from VD samples prepared in COG.

3.4. Influences of N and S Species on the Reduction Behavior and Crushing Strength of VD Samples

The temperature dependencies of reduction for the VD sample shown in Fig. 3 were investigated to clarify the effects of N and S on the reduction behaviors, which were calculated based on the amounts of gaseous-O produced during the heat treatment in He and 55%H₂/He, and the results are shown in Fig. 10. The reduction behavior of the VD sample prepared in NH₃/COG and H₂S/COG was almost the same as that of the samples prepared in He, H₂, and COG. Therefore, the N and S species do not affect the reduction rate of the VD sample prepared under this condition. Table 2 summarizes the changes in the iron forms in the VD samples prepared under COG during the heat treatment in He or 55%H₂/He. In He, α-Fe formation was observed above 800°C, and Fe₃C was also measured in the sample heated to 1000°C for 60 min. In contrast, formation of α-Fe occurred above 700°C in 55%H₂/He and Fe₃C was observed at 800°C. Similar results were observed for the VD samples prepared in NH₃/COG or H₂S/COG, and Fe–N or Fe–S compounds were not found in samples after the heat treatment under either atmosphere. The carbon contents in the samples recovered after heat treatment in He or 55%H₂/He were approximately 9 and 13 wt%-dry, respectively. In addition, as shown in Figs. 7 and 9, some of the N and S in prepared VD samples remained in the solid products after the heat treatment. These results suggest that N or S species may exist in the solid products after the heat treatment as thermally stable forms. According to previous work, Fe_xN derived from the reaction between NH₃ and α-Fe decomposes above 400°C.²⁴⁾ It is thus unlikely that the N species in the VD samples would exist as Fe–N species, and that N in solid phase after the heat treatment would be heterocyclic-N. However, it is possible that H₂S produced from carbonaceous materials in the VD samples during the reduction process could react with Fe species according to Eqs. (1), (2), (3):   

$$\begin{array}{l}\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4+3{\mathrm{H}}_2\mathrm{S}+{\mathrm{H}}_2\to 3\mathrm{F}\mathrm{e}\mathrm{S}+4{\mathrm{H}}_2\mathrm{O}\\ \mathrm{\Delta }{\mathrm{G}}_{400-1\mathrm{\hspace{0.17em} }000°\mathrm{C}}=-31\mathrm{\hspace{0.17em} }\mathrm{t}\mathrm{o}\mathrm{\hspace{0.17em} }-39\mathrm{\hspace{0.17em} }\mathrm{k}\mathrm{c}\mathrm{a}\mathrm{l}/\mathrm{m}\mathrm{o}\mathrm{l}\end{array}$$(1)

  

$$\begin{array}{l}\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4+3{\mathrm{H}}_2\mathrm{S}+\mathrm{C}\mathrm{O}\to 3\mathrm{F}\mathrm{e}\mathrm{S}+3{\mathrm{H}}_2\mathrm{O}+\mathrm{C}{\mathrm{O}}_2\\ \mathrm{\Delta }{\mathrm{G}}_{400-1\mathrm{\hspace{0.17em} }000°\mathrm{C}}=-34\mathrm{\hspace{0.17em} }\mathrm{t}\mathrm{o}\mathrm{\hspace{0.17em} }-37\mathrm{\hspace{0.17em} }\mathrm{k}\mathrm{c}\mathrm{a}\mathrm{l}/\mathrm{m}\mathrm{o}\mathrm{l}\end{array}$$(2)

  

$$\begin{array}{l}\mathrm{F}\mathrm{e}\mathrm{O}+{\mathrm{H}}_2\mathrm{S}\to \mathrm{F}\mathrm{e}\mathrm{S}+{\mathrm{H}}_2\mathrm{O}\\ \mathrm{\Delta }{\mathrm{G}}_{400-1\mathrm{\hspace{0.17em} }000°\mathrm{C}}=-13\mathrm{\hspace{0.17em} }\mathrm{t}\mathrm{o}\mathrm{\hspace{0.17em} }-12\mathrm{\hspace{0.17em} }\mathrm{k}\mathrm{c}\mathrm{a}\mathrm{l}/\mathrm{m}\mathrm{o}\mathrm{l}\end{array}$$(3)

Here, H₂ and CO were selected as the production gases during the reduction process. The former is generated from pyrolysis of carbonaceous materials because H₂O formation is derived from the reduction of iron oxide in the VD sample, as shown in Fig. 5. The latter gas is generated from direct reduction of iron oxide and carbonaceous materials in the VD samples, and Eqs. (1), (2), (3) are supported by thermodynamic calculations. However, Eqs. (4) and (5) are not thermodynamically favorable.   

$$\begin{array}{l}\mathrm{F}\mathrm{e}\mathrm{S}+{\mathrm{H}}_2\to \mathrm{F}\mathrm{e}+{\mathrm{H}}_2\mathrm{S}\mathrm{\hspace{0.17em} }\text{(}\mathrm{g}\text{)}\\ \mathrm{\Delta }{\mathrm{G}}_{400-1\mathrm{\hspace{0.17em} }000°\mathrm{C}}=14\mathrm{\hspace{0.17em} }\mathrm{t}\mathrm{o}\mathrm{\hspace{0.17em} }13\mathrm{\hspace{0.17em} }\mathrm{k}\mathrm{c}\mathrm{a}\mathrm{l}/\mathrm{m}\mathrm{o}\mathrm{l}\end{array}$$(4)

  

$$\begin{array}{l}\mathrm{F}\mathrm{e}\mathrm{S}+\mathrm{C}\mathrm{O}\to \mathrm{F}\mathrm{e}+\mathrm{C}\mathrm{O}\mathrm{S}\mathrm{\hspace{0.17em} }(\mathrm{g})\\ \mathrm{\Delta }{\mathrm{G}}_{400-1\mathrm{\hspace{0.17em} }000°\mathrm{C}}=18\mathrm{\hspace{0.17em} }\mathrm{t}\mathrm{o}\mathrm{\hspace{0.17em} }21\mathrm{\hspace{0.17em} }\mathrm{k}\mathrm{c}\mathrm{a}\mathrm{l}/\mathrm{m}\mathrm{o}\mathrm{l}\end{array}$$(5)

From these results, although S is present in the prepared and heat-treated VD samples, S will exist in the solid phase recovered after the heat treatment as iron sulfide and/or organic-S forms, and the details of this process will be subject of future work.

Fig. 10.

Reduction behavior of VD samples prepared in various atmosphere during heat treatment in He or 55%H₂/He.

Table 2. Summaries of XRD measurement for VD samples prepared in COG before and after heat treatment in He or 55%H₂/He atmosphere.

Temperature, °C	Atmosphere in heat treatment
	Hea	55%H2/Hea
As-prepared	Fe2O3 (m)	Fe2O3 (m)
500	Fe3O4 (m)	Fe3O4 (m)
600	Fe3O4 (m)	Fe3O4 (m), FeO (vw), Fe (vw)
700	Fe3O4 (m), FeO (vw)	FeO (vw), Fe (s)
800	FeO (w), Fe (s)	Fe (s), Fe3C (vw)
900	Fe (vs), Fe3C (vw)	Fe (vs), Fe3C (vw)
1000	Fe (vs), Fe3C (vw)	Fe (vs), Fe3C (vw)
1000 60 min	Fe (vs), Fe3C (vw)	Fe (vs), Fe3C (vw)

^a XRD intensities designated by vw (very weak), w (weak), m (medium) s (strong), and vs (very strong).

Figure 11 shows the results of examining the influence of N and S on the crushing strength of the VD sample during the heat treatment. The crushing strengths of the VD samples prepared in NH₃/COG or H₂S/COG do not change up to a reduction ratio of 40–50% at 800°C, and neither do those of the VD sample prepared in COG without NH₃ or H₂S. These results suggest that the N and S species incorporated in the VD sample almost do not affect the reduction of iron oxide and the change in the crushing strength in the VD sample.

Fig. 11.

Change in crushing strengths of VD samples prepared in various atmosphere during heat treatment in He or 55%H₂/He.

4. Conclusions

In this study, we first investigated the effect of the VD atmosphere on the N and S contents in VD samples and their crushing strengths; then, the fates of N and S in the VD samples during the reduction process were examined. In addition, the influence of N and S against the reduction rate and crushing strength during the reduction process were investigated, and the following conclusions were obtained.

(1) Although N in NH₃ that was fed with simulated COG components did not transfer to the prepared VD sample, some of the S in the H₂S in the feed moved into the VD sample.

(2) The amount of these elements did not affect the crushing strength of the prepared VD sample.

(3) The N and S forms in the VD sample mainly existed as organic-N or S in VD samples, based on the XRD and XPS analyses.

(4) The N species in the VD sample during the reduction process under inert (He) and reductive atmospheres (55%H₂/He) mainly evolved as NH₃ and N₂ at 400–1000°C, the cumulative amount of evolved N₂ evolved was greater than that of NH₃.

(5) The H₂S evolution began at 400°C, and the profile provided the main peak at approximately 800°C. The amount of evolved H₂S in 55%H₂/He was greater than that of He.

(6) Although the reduction of the VD sample started near 400°C and stopped at 1000°C, N and S species in the sample did not affect the reduction rate.

(7) In addition, N and S in the VD samples did not influence the crushing strength during the heat treatment.

Acknowledgements

This work was entrusted by the CO₂ Ultimate Reduction in Steelmaking process by innovative technology for cool Earth 50 (COURSE50) project, sponsored by the New Energy and Industrial Technology Development Organization (NEDO), Japan.

References

• 1)   M.  Naito,  A.  Okamoto,  K.  Yamaguchi,  T.  Yamaguchi and  Y.  Inoue: Tetsu-to-Hagané, 87 (2001), 357.
• 2)   S.  Nomura,  H.  Ayukawa,  H.  Kitaguchi,  T.  Tahara,  S.  Matsuzaki,  M.  Naito,  S.  Koizumi,  Y.  Ogata,  T.  Nakayama and  T.  Abe: ISIJ Int., 45 (2005), 316.
• 3)   A.  Kasai,  M.  Naito,  Y.  Matsui and  Y.  Tamagata: Tetsu-to-Hagané, 89 (2003), 1212.
• 4)   Y.  Iwanaga: Tetsu-to-Hagané, 68 (1982), 740.
• 5)   Y.  Mochizuki,  N.  Tsubouchi and  T.  Akiyama: Fuel Process. Technol., 138 (2015), 704.
• 6)   Y.  Mochizuki,  M.  Nishio,  N.  Tsubouchi and  T.  Akiyama: Fuel Process. Technol., 142 (2016), 287.
• 7)   Y.  Mochizuki,  M.  Nishio,  N.  Tsubouchi and  T.  Akiyama: Energy Fuel., 30 (2016), 2102.
• 8)   Y.  Mochizuki,  M.  Nishio,  N.  Tsubouchi and  T.  Akiyama: Energy Fuel., 31 (2017), 8877.
• 9)   S.  Tonomura: Energy Procedia, 37 (2013), 7160.
• 10)   K.  Nishioka,  Y.  Ujisawa,  S.  Tonomura,  N.  Ishiwata and  P.  Sikstrom: J. Sustain. Metall., 2 (2016), 200.
• 11)   S.  Watakabe,  K.  Miyagawa,  S.  Matsuzaki,  T.  Inada,  Y.  Tomita,  K.  Saito,  M.  Osame,  P.  Sikström,  L. S.  Ökcits and  J.-O.  Wikstrom: ISIJ Int., 53 (2013), 2065.
• 12)   R.  Kuwano,  T.  Oku and  Y.  Ono: Tetsu-to-Hagané, 66 (1980), 1622.
• 13)   S.  Hayashi,  Y.  Iguchi and  J.  Hirao: Tetsu-to-Hagané, 71 (1985), 281.
• 14)   Y.  Mochizuki,  M.  Nishio,  J.  Ma,  N.  Tsubouchi and  T.  Akiyama: Energy Fuel., 30 (2016), 6233.
• 15)   Y.  Ohshima,  N.  Tsubouchi and  Y.  Ohtsuka: Appl. Catal. B., 111 (2012), 614.
• 16)   H.  Naono and  R.  Fujiwara: J. Colloid Interface Sci., 73 (1980), 406.
• 17)   K.  Sugawara,  Y.  Enda,  T.  Kato,  T.  Sugawara and  M.  Shirai: Energy Fuel., 17 (2003), 204.
• 18)   X.  Li,  J.-i.  Hayashi and  C.-Z.  Li: Fuel, 85 (2006), 1700.
• 19)   J.-i.  Ozaki,  Y.  Uozumi,  A.  Oya,  K.  Saito and  S.  Aizawa: Tetsu-to-Hagané, 92 (2006), 157.
• 20)   J.  Donald,  C.  Xu,  H.  Hashimoto,  E.  Byambajav and  Y.  Ohtsuka: Appl. Catal. A, 375 (2010), 124.
• 21)   J.  Torres,  C. C.  Perry,  S. J.  Bransfield and  D. H.  Fairbrother: J. Phys. Chem. B, 107 (2003), 5558.
• 22)   N.  Tsubouchi,  A.  Ogawa and  Y.  Mochizuki: Appl. Catal. A, 499 (2015), 133.
• 23)   N.  Ikenaga,  H.  Taniguchi,  A.  Watanabe and  T.  Suzuki: Fuel, 79 (2000), 273.
• 24)   N.  Tsubouchi,  H.  Hashimoto and  Y.  Ohtsuka: Catal. Lett., 105 (2005), 203.
• 25)   Y.  Mochizuki and  N.  Tsubouchi: Energy Fuel., 31 (2017), 8087.
• 26)   D. L.  Cocke,  M.  Jurcik-Rajman and  S.  Veprek: J. Electrochem. Soc., 136 (1989), 3655.
• 27)   P. C. J.  Graat,  M. A. J.  Somers and  E.  Mittemeijer: Appl. Surf. Sci., 136 (1998), 238.
• 28)   W.  Diekmann,  G.  Panzner and  H.  Grabke: Surf. Sci., 218 (1989), 507.
• 29)   S. R.  Kelemen,  M. L.  Gorbaty and  P.  Kwiatek: Energy Fuel., 8 (1994), 896.
• 30)   J. R.  Pels,  F.  Kapteijn,  J. A.  Moulijn,  Q.  Zhu and  K. M.  Thomas: Carbon, 33 (1995), 1641.
• 31)   S. R.  Kelemen,  M. L.  Gorbaty,  S. N.  Vaughn and  P. J.  Kwiatek: Energy Fuel., 8 (1994), 896.
• 32)   S. R.  Kelemen,  G. N.  George and  M. L.  Gorbaty: Fuel, 69 (1990), 939.
• 33)   J. F.  Moulder,  W. F.  Stichle,  P. E.  Sobol and  K. D.  Bomben: Handbook of X-ray Photoelectron Spectroscopy, Perkin-Elmer, Eden Prairie, (1992).
• 34)   A. V.  Naumkin,  A.  Kraut-Vass,  S. W.  Gaarenstroom and  C. J.  Powell: NIST Standard Reference Database 20, Version 4.1, NIST X-ray Photoelectron Spectroscopy Database, http://srdate.nist.gov/xps/Default.aspx, (accessed 2018-01-19).
• 35)   M. E.  Machado,  E. B.  Caramao and  C. A.  Zini: J. Chromatogr. A, 1218 (2011), 3200.
• 36)   N.  Tsubouchi,  Y.  Mikawa,  Y.  Mochizuki,  T.  Kikuchi and  Y.  Ohtsuka: Energy Fuel., 31 (2017), 3885.
• 37)   H.  Mori,  K.  Asami and  Y.  Ohtsuka: Energ. Fuel., 10 (1996), 1022.
• 38)   Y.  Ohtsuka,  T.  Watanabe,  K.  Asami and  H.  Mori: Energy Fuel., 12 (1998), 1356.
• 39)   N.  Tsubouchi and  Y.  Ohtsuka: Fuel Process. Technol., 89 (2008), 379.
• 40)   J.  Leppdahti and  T.  Koljonen: Fuel Process. Technol., 43 (1995), 1.
• 41)   J.-J.  Xu,  M.  Makino and  M.  Kaiho: J. Jpn. Inst. Energy, 72 (1993), 1177.