The reactivity of the original hematite and lignite–modified hematite samples was thoroughly investigated to explore the effect of exogenous metals on the reactivity of the hematite OCs. Moreover, the reactivity of the doping metal salts or oxides (i.e., CaO, CuO, Mn2O3, and NiO) and the new crystal phases (i.e., CaFe2O4, CuFe2O4, MnFeO3, and NiFe2O4) are discussed. The reactivity of the Fe–K–O and Na–Fe–Si–O species are not discussed because it was difficult to determine the new crystal phase in the K–modified and Na–modified OCs.
3.2.1. Effect of K and Na Modifications
Figure 3 presents the TG and DTG curves of the different mixtures. As shown in
Figure 3a, compared to the inert SiO
2 sample, the active OC evidently promoted lignite conversion, and an apparent weightlessness process was observed. Furthermore, it was observed that the three mixtures (lignite–containing K–modified hematite) exhibited a slightly higher weight loss than hematite, indicating that the addition of K facilitates lignite conversion. Moreover, this trend became evident as the K content increased. These results preliminarily indicate that K modification improves the reactivity of hematite. As shown in
Figure 3b, there are four weight loss peaks in the DTG curves, which correspond to the different reaction stages. The weight loss peak at 96 °C is attributed to the evaporation of the water in lignite and the OCs (stage I). The second weight loss peak at 420 °C is assigned to the devolatilization of lignite (stage II). The small weight loss peak at about 660 °C is due to the secondary devolatilization of lignite (stage III). The fourth weight loss peak, i.e., the largest weight loss peak, is attributed to the reaction between lattice oxygen and lignite char (stage IV) [
41]. The former two weight loss peaks almost overlap in the DTG curves because the dehydration and devolatilization processes are not the rate–controlling steps of lignite conversion. The third weight loss peak for the inert SiO
2 sample is very small because of the extremely low degree of secondary lignite devolatilization. A fourth weight loss peak was not observed for this sample because SiO
2 could not provide lattice oxygen. An apparent weight loss peak (stage IV) was observed for the other four mixtures (the characteristics of this stage are shown in
Table 3) because a large amount of active lattice oxygen reacted with the lignite char. Compared to those of the hematite sample (700 °C and 910 °C), the initial and peak temperatures of the K5 (698 °C and 900 °C), K10 (563 °C and 768 °C), and K15 (540 °C and 735 °C) samples exhibited a noticeable forward shift in stage IV. This is because the addition of K promoted the generation of an active solid solution of K–Fe–O species, which has a strong catalytic activity to facilitate lignite char conversion [
42]. The lattice oxygen in K10 and K15 had a high reactivity and reacted with the lignite char at about 600 °C, causing the secondary devolatilization process to overlap with the char conversion process. Thus, the weight loss peak of the secondary volatilization of lignite was not observed. Moreover, the reactivity of the K–modified OCs exhibited an increasing trend as the K content increased because additional K–Fe–O species were generated. Thus, the K15 sample exhibited the highest reactivity among the four OCs in this work.
During the lignite conversion, the three mixtures (i.e., the lignite–containing Na5, Na10, and Na15 samples) exhibited weight losses of 43.6 wt.%, 46.7 wt.%, and 48.9 wt.%, respectively, which are slightly higher than that of hematite (43.5 wt.%), as shown in
Figure 4a. These results preliminarily indicate that Na modification also enhances the reactivity of hematite, promoting lignite conversion.
Similar to the K–modified samples, the Na–modified samples contained lattice oxygen with a high reactivity (no weight loss peak of secondary volatilization of lignite was observed), and their DTG curves displayed three apparent weight loss peaks. The first and second weight loss peaks are ascribed to the dehydration and devolatilization processes, respectively. Since the Na element easily absorbs water in the air, the first weight loss peak sharpened as the Na content in the OCs increased. In terms of the third weight loss peak (corresponding to the secondary devolatilization and char conversion), the initial and peak temperatures of the Na5, Na10, and Na15 samples were 552 °C and 832 °C, 527 °C and 748 °C, and 516 °C and 705 °C, respectively, which are lower than those of the hematite sample (700 °C and 910 °C), as shown in
Table 4. Furthermore, the initial and peak temperatures gradually decreased as the Na content in the OCs increased. These results suggest that Na modification increased the reactivity of hematite because of the formation of Fe–O–Na species in the modified OCs. Additionally, the OC reactivity increased as the Na content increased.
3.2.2. Effect of Ca Modification
The effect of Ca modification on the reactivity of hematite is discussed, as shown in
Figure 5 and
Table 5. The four mixtures (lignite–containing Ca5, Ca10, Ca15, and hematite) exhibited similar weight losses, indicating that Ca modification did not promote lignite conversion. As shown in
Figure 5b, two small weight loss peaks corresponding to the lignite dehydration (stage I) and devolatilization (stage II) processes, respectively, are almost overlapping, indicating that Ca modification did not promote the devolatilization of lignite. As the Ca content increased, the third weight loss peak (stage III) became more apparent (the weight loss peak at 673 °C for the Ca15 sample), suggesting that Ca modification can promote the secondary devolatilization of lignite, where the volatile products can react with lattice oxygen in this stage. The fourth weight loss peak is attributed to the lignite char conversion. The initial and peak temperatures of the Ca5 (709 °C and 851 °C), Ca10 (709 °C and 897 °C), and Ca15 (721 °C and 897 °C) samples were similar to those of hematite (700 °C and 910 °C). However, the reaction rates of the different lignite–containing OCs were substantially different. The Ca5, Ca10, and Ca15 samples had maximum reaction rates (i.e., the reaction rate of the OC and lignite char) of 0.86 wt.%/min, 1.09 wt.%/min, and 0.92 wt.%/min, respectively, which are lower than that of the hematite sample (1.82 wt.%/min). These results indicate that Ca modification decreases the reaction rate of OCs with lignite char.
Comparative experiments were conducted on the reactivity of the different lignite–containing OCs (i.e., the lignite–containing SiO
2, CaO, CaFe
2O
4, Ca15, and hematite) to determine why the reactivity of the Ca–modified hematite decreased, and the results are illustrated in
Figure 6 and
Table 6. CaO and CaFe
2O
4 were newly formed crystal phases in the Ca15 sample. Hematite was observed to exhibit the highest weight loss among those five mixtures, indicating that hematite of equal mass has the highest lattice oxygen content. As shown in
Figure 6b, there is an overlap with the first weight loss peak ascribed to the dehydration process for all the samples. The CaO sample can be observed to exhibit two large weight loss peaks at 400 °C and 660 °C. The weight loss peak at 400 °C is noticeably stronger than that of the other four samples, which is attributed to the catalytic effect of CaO on the lignite devolatilization process [
43]. The weight loss peak at about 660 °C is attributed to CaO, which can further promote the secondary devolatilization of lignite. Similar to the SiO
2 sample, the CaO sample did not exhibit a char conversion peak (stage IV) because it could not provide lattice oxygen. Although the DTG curves of CaFe
2O
4 and hematite basically coincided in the first three stages (the third peak is not noticeable), the reaction rate of CaFe
2O
4 (0.89 wt.%/min) was significantly lower than that of hematite (1.82 wt.%/min) in stage IV (the reaction between lattice oxygen and lignite char), indicating that CaFe
2O
4 has a lower reactivity than hematite. This is consistent with a previous finding [
44]. Consequently, the Ca–modified hematite had a lower reactivity than the original hematite because of the formation of large amounts of CaFe
2O
4.
Generally, it can be concluded that CaO plays a major role in promoting lignite devolatilization, while the formation of a solid CaFe2O4 solution inhibits lattice oxygen release, thereby decreasing the reactivity of OCs.
3.2.3. Effect of Cu, Mn, and Ni Modifications
The effects of the modification of transition metals (i.e., Cu, Mn, and Ni) on the reactivity of hematite are also discussed in this paper. As shown in
Figure 7a, the three mixtures (the lignite–containing Cu5, Cu10, and Cu15 samples) exhibited weight losses of 41.1 wt.%, 40.1 wt.%, and 43.3 wt.%, respectively, which are lower than that of hematite (43.5 wt.%). This is attributed to the CuO sample having a lower lattice oxygen content (20 wt.%) than the Fe
2O
3 sample (30 wt.%). As shown in
Figure 7b, the first weight loss peaks of all the samples overlap because of the dehydration process, but the second weight loss peak exhibits a slight increase as the Cu content increases, which is due to the small amount of free O
2 released by CuO/CuFe
2O
4 [
45,
46]. The third weight loss peaks of all the samples are extremely small, and they also almost overlap. For the fourth weight loss peak, the initial temperature of the four mixtures was about 700 °C, but the peak temperatures of the Cu–modified OCs were higher than that of hematite (
Table 7). Moreover, the Cu–modified OCs with lignite char had a lower reaction rate than hematite. These results reveal that Cu modification reduced the reactivity of hematite, inhibiting the reactions between lattice oxygen and lignite char; this trend became evident as the K content increased.
Furthermore, comparative experiments were conducted on the reactivity of the different OCs (i.e., the lignite–containing CuO, CuFe
2O
4, Cu15, and hematite), and the results are illustrated in
Figure 8 and
Table 8. The lignite–containing CuO and CuFe
2O
4 samples exhibited lower weight losses (35.5 wt.% and 39.5 wt.%) than hematite (43.5 wt.%) because the lattice oxygen content in CuO and CuFe
2O
4 was lower than that in Fe
2O
3. The CuO sample exhibited two noticeable weight loss peaks at 400 °C and 660 °C, which are attributed to the reactions between free O
2 released by CuO and the volatiles, as well as the conversion of residual lattice oxygen and lignite char, respectively. The second weight loss peaks of CuFe
2O
4 and Cu15 are also slightly higher than that of hematite, suggesting that Cu modification promoted lignite volatile conversion, and that the promotion effect of CuO was more pronounced than that of CuFe
2O
4 (the order is CuO > CuFe
2O
4 > Cu15 > hematite). In stage IV, the reaction rates of CuFe
2O
4 (0.64 wt.%/min) and Cu15 (0.79 wt.%/min) were significantly lower than that of hematite (1.82 wt.%/min), owing to the release of lattice oxygen (i.e., lattice oxygen was converted into gas phase oxygen) during the initial reaction stage of the Cu–modified OCs.
In summary, the crystalline phases of CuO and CuFe2O4 formed by Cu doping can promote the conversion of lignite volatiles but inhibit the conversion of lignite char.
Similar to the Ca–modified samples, the Mn–modified samples did not significantly promote lignite conversion because their weight loss values before and after Mn modification were almost consistent, as shown in
Figure 9a. Meanwhile, the first three small weight loss peaks of the four mixtures (i.e., the lignite–containing Mn5, Mn10, Mn15 and hematite samples) almost overlap (
Figure 9b), indicating that Mn modification had no noticeable promoting effect on the dehydration (stage I), devolatilization (stage II), and secondary devolatilization (stage III) of lignite. In stage IV, Mn15 (0.85 wt.%/min), Mn10 (0.73 wt.%/min), and Mn5 (0.78 wt.%/min) had substantially lower maximum reaction rates than hematite (1.82 wt.%/min), as shown in
Table 9, indicating that the addition of Mn decreases the reaction rate of OCs and lignite char.
Experiments were conducted on the reactivity of the different OCs (i.e., the lignite–containing Mn
2O
3, MnFeO
3, Mn15, and hematite samples), and the results are shown in
Figure 10 and
Table 10. As shown in
Figure 10a, the lignite–containing Mn
2O
3 sample exhibited a small weight loss (30.1 wt.%) because the lattice oxygen content in Mn
2O
3 (2.43 wt.%) was substantially lower than that in Fe
2O
3 (30 wt.%). Additionally, the lignite–containing MnFeO
3 sample exhibited a smaller weight loss (35.6 wt.%) than hematite (as shown in
Table 10). A noticeable weight loss peak at 620 °C (in stage III) can be observed in the DTG curves of Mn
2O
3 (
Figure 10b), which is attributed to the release of free O
2 from Mn
2O
3 to promote lignite conversion [
47]. Similar to that of the CuO sample, the char conversion process of the Mn
2O
3 sample advanced and coincided with the secondary devolatilization process because of the release of free O
2. During the char conversion process, the weight loss rates of MnFeO
3 (0.62 wt.%/min) and Mn15 (0.85 wt.%/min) were significantly lower than that of hematite (1.82 wt.%/min), indicating that the formation of the MnFeO
3 phase decreased the reaction rate of the OCs and lignite char, thereby inhibiting the lignite conversion.
In conclusion, the Mn2O3 and MnFeO3 phases in the Mn–modified hematite OCs both inhibit reactions between the OC and lignite char, but the Mn2O3 phase alone promotes reactions between the OC and lignite volatiles.
The effects of Ni modification on the reactivity of hematite were determined, and the results are shown in
Figure 11 and
Table 11. The TG curves (
Figure 11a) show that the four mixtures (i.e., the lignite–containing Ni5, Ni10, Ni15, and hematite samples) exhibited weight losses of 43.88 wt.%, 46.09 wt.%, 41.31 wt.%, and 43.5 wt.%, respectively, with Ni15 exhibiting a noticeably smaller weight loss than hematite. This indicates that excessive Ni loading inhibits lignite conversion because high Ni contents cause surface carbon deposition [
35,
48]. Additionally, high Ni contents result in the sintering of OC particles. As shown in
Figure 11b, the third weight loss peak increased slightly as the Ni content increased, demonstrating that the addition of Ni promotes the secondary devolatilization of lignite. Regarding the fourth weight loss peak, the initial and maximum reaction temperatures of the Ni5, Ni10, and Ni15 samples are consistent with those of the hematite sample (i.e., 700 °C and 910 °C, respectively), as shown in
Table 11. Additionally, the reaction rate of the Ni15 sample (1.39 wt.%/min) was substantially lower than those of the Ni5 (1.68 wt.%/min), Ni10 (1.75 wt.%/min), and hematite (1.82 wt.%/min) samples, indicating that excessive Ni loading is ineffective for lignite char conversion.
Figure 12 presents the experimental results for SiO
2, hematite, Ni15, NiO, and NiFe
2O
4, which were all mixed with lignite. Similar to the CuO and Mn
2O
3 samples, the lignite–containing NiO samples exhibited a smaller weight loss (38.7 wt.%) than hematite (43.5 wt.%) because the lattice oxygen content in NiO (21.33 wt.%) was also lower than that in Fe
2O
3 (30 wt.%), as illustrated in
Figure 12a. The DTG curves of the five mixtures almost overlap in the first two stages, indicating that the crystalline phases of NiO and NiFe
2O
4 did not affect the lignite dehydration and devolatilization processes, as illustrated in
Figure 12b. The initial and peak temperatures for the maximum weight loss peak of NiO (526 °C and 720 °C) and NiFe
2O
4 (589 °C and 880 °C) were significantly lower than those of Ni15 (760 °C and 910 °C) and hematite (700 °C and 910 °C), as shown in
Table 12, which is attributed to the high activity of NiO and NiFe
2O
4, which can react with lignite secondary devolatilization products and lignite char [
49,
50,
51]. The third weight loss peak of the Ni15 sample is ascribed to the formation of the NiFe
2O
4 phase.
In conclusion, the high reactivity of the Ni–modified OCs was due to the formation of the NiO and NiFe2O4 crystal phases.
Comparative experiments were also conducted using the M10 OCs to compare the reactivity of the different metal–modified OCs containing lignite. The results are shown in
Figure 13 and
Table 13.
As shown in
Figure 13, the largest weight loss peak (i.e., the reaction between lattice oxygen with lignite char) of K10 and Na10 are significantly forward–shifted in comparison to that of the other samples, with a significantly lower initial and peak temperature than that of the other samples (
Table 13), while the maximum weight loss rate of K10 and Na10 was less different from that of hematite. This indicates that the reactivity of the K–modified and Na–modified hematite OCs was higher than that of the other OCs. The peak value of K10 is larger than Na10, indicating that the modification of K is better than Na.
The peak temperature of Ca10, Cu10, Mn10, and Ni10 are all around 800 °C, so the performance is mainly compared by the peak value of the OC with lignite char. The addition of Ni promoted the conversion of lignite volatiles, and the reaction rates of lattice oxygen in the Ni10 sample and lignite char were almost the same as that of the hematite sample (the maximum reaction rates for hematite and Ni10 were 1.82 wt.%/min and 1.75 wt.%/min, respectively), demonstrating that using an appropriate amount of Ni for modification improves the performance of hematite. The Ca–modified, Cu–modified, and Mn–modified OCs exhibited relatively weak reactivity during the lignite char conversion. The order of reactivity is as follows: hematite (1.82 wt.%/min) > Ca10 (1.09 wt.%/min) > Cu10 (0.89 wt.%/min) > Mn10 (0.73 wt.%/min).