2.1 Physicochemical Characterization of the Materials Tested

The amount of Fe supported was not quantitative for the organic xerogel (OX) and carbonized material (CX) samples (6–7 wt%) (Table 1). However, for the activated materials (AX), the impregnation efficiency reached up to 85% (17 wt%). Generally, a higher surface area promotes a more quantitative immobilization of Fe on the support (Table 1). However, the surface chemistry of the samples may also influence their interaction with Fe during the immobilization process. Therefore, although the Fe content in the OX and CX samples is similar (6–7 wt%), their performance can be compared in terms of their differing textural properties.

The sol-gel methodology enables the design of mesopores (i.e., pores between 2 and 50 nm in width), often referred to as feeder pores, during the sol-gel process, while microporosity (i.e., pores <2 nm in width) can be developed independently during postsynthesis treatments. Consequently, the OX, CX, and AX samples exhibit similar mesoporosity, with pores in the 6-8 nm range (Figure S1a, Supporting Information), but they show significant differences in microporosity, as indicated by the volume of N₂ adsorbed at low relative pressures (Figure S1b, Supporting Information).

All the supports exhibit type I–IV isotherms according to the IUPAC classification,^[34] which are characteristic of micro- and mesoporous samples with a well-defined capillary condensation step at p/p0 above 0.5, which is indicative of a well-developed mesoporosity, consistent with uniform cylindrical mesopores and greater pore connectivity. This is confirmed by the pore size distribution (PSD), which shows that the porosity is made up of uniform mesopores ranging from 2 to 10 nm. The presence of a hysteresis loop also suggests a similar mesoporous structure across the samples. The carbonization process (i.e., heat treatment under an inert atmosphere) removes volatile matter in CX, resulting in the development of microporosity, which is reflected in higher adsorbed volumes at low p/p⁰ compared to the original polymer OX. The activation process used during the synthesis of AX further enhances this microporosity, as indicated by the increased adsorption at low relative pressures in Figure S2b, Supporting Information. This increase in microporosity contributes to a higher specific surface area (S_BET), as summarized in Table 1, following the trend OX < CX < AX (226, 619, and 1597 m² g⁻¹, respectively).

Incorporation of Fe into these samples partially blocks the pores, leading to a decrease in S_BET compared to the pristine samples: 158, 523, and 1325 m² g⁻¹ for OXFe(g), CXFe(g), and AXFe(g), respectively. A slight increase in the S_BET is observed when these samples undergo heat treatment to convert goethite to hematite, as seen in CXFe(h) and AXFe(h). However, this increase is not observed for OXFe(h), where a notable change in morphology occurs. The low thermal stability of the OX sample may induce secondary reactions with the Fe precursors, further blocking the porosity, as indicated by the reduction in both microporosity and S_BET. Additionally, the N₂ adsorption–desorption isotherm shifts to type II (Figure S2, Supporting Information), indicating the loss of mesopores. Accordingly, OXFe(h) is completely different from OXFe(g), not only in terms of its chemistry but also in porosity of the resultant material, with a denser structure clearly reflected by the increase of helium density and the decrease of the porous characteristics (Table 1).

The morphologies of Fe-impregnated samples were examined using scanning electron microscopy (SEM) (Figure 1). Iron species predominantly appeared as polyhedra and nanorods. However, the distribution of these iron species was more uniform in the CXFe and AXFe samples compared to OXFe (Figure 1a–c). For the OXFe material, a lower surface coverage was observed, with noticeable agglomeration of nanorods forming in certain areas (Figure 1a).

The analysis by X-ray diffraction (XRD) (Figure S3a, Supporting Information) revealed the presence of goethite in the samples synthesized by oxidative hydrolysis (OXFe(g), CXFe(g), and AXFe(g)). The subsequent heat treatment at 300 °C led to the transformation of goethite into hematite (OXFe(h), CXFe(h), and AXFe(h)) (Figure S3b, Supporting Information). Notably, the original polyhedral and nanorod morphologies of the iron species were preserved after this thermal transformation (Figure 1c,d).

As observed in previous studies conducted by the authors,^[35] the type of iron species present on the surface of carbon materials significantly impacts mercury adsorption performance. This performance is also influenced by the textural properties of the material and, distinctly, by the presence of SO₂.

2.2 Mercury Retention: The Role of Porous Texture

Figure 2 shows the mercury retention capacity alongside the surface area for all the carbon materials studied. As previously discussed, these materials exhibit variable textural characteristics but were designed with hierarchical porosity to enable a comprehensive evaluation of their effects on mercury removal and SO₂ resistance. It is worth noting that OX and CX samples have comparable total pore volumes and mesopore sizes but differ in micropore volume, leading to differences in S_BET (Table 1). In contrast, the AX samples have a more developed porosity with higher pore volumes across the entire porosity range. The mercury retention capacity varies according to the type of carbon material and the iron species present.

Similar mercury retention capacities (1–2 μg g⁻¹) were observed in the raw materials and the OX sample loaded with iron species (OXFe(g) and OXFe(h)), regardless of their textural properties (Figure 2, Table 1). This suggests that the carbonaceous matrix itself lacks significant activity for mercury capture and serves mainly to provide a support with appropriate porosity for the active phase. Additionally, the limited textural properties of the OXFe samples reduce the accessibility of Fe active sites, leading to low mercury removal activity irrespective of the type of iron species present.

Impregnation of the CX and AX with Fe had a significant impact on Hg capture efficiency, with the highest retention observed in CX and AX samples containing hematite (20 and 21 μg g⁻¹ for CXFe(h) and AXFe(h), respectively). Previous studies have shown that the outer layer of Fe³⁺ in the iron oxide (αFe₂O₃), which has an empty orbital structure, enhances the Fe/Hg⁰ interaction .^[36] In fact, earlier research by the authors on carbon foams demonstrated that αFe₂O₃ enhances mercury capture efficiency compared to other iron oxides and hydroxides under similar conditions.^[35] The results obtained in this work corroborate these findings, as both CXFe(h) and AXFe(h) exhibit higher mercury retention values than CXFe(g) and AXFe(g), respectively, reinforcing the role of αFe₂O₃ in Hg removal.

In terms of textural properties, it is notable that CXFe(h) and AXFe(h) exhibit similar mercury retention capacities, despite CXFe(h) having a considerably smaller surface area than AXFe(h) (Table 1, Figure 2). In contrast, for samples loaded with goethite (CXFe(g) and AXFe(g)), mercury retention is higher in AXFe(g) than in CXFe(g). Both samples have the same iron morphology, characterized by nanorods and polyhedral particles (Figure 1), but AXFe(g) benefits from a more developed microporosity, which likely enhances the availability of active sites and thus contributes to its superior Hg retention.

The results suggest that 1) the formation of αFe₂O₃ on the surface of carbon materials creates active sites for the chemisorption/oxidation of elemental mercury, but the critical factor is the overall porosity of the material, rather than solely the microporosity, as commonly suggested in most studies carried out so far; and 2) when FeOOH is the iron species, which results in fewer active sites available for Hg⁰ oxidation,^[36] a more developed pore structure with a higher surface area in AXFe(g) compared to CXFe(g) (Table 1) enhances mercury retention (Figure 2). The findings also indicate that a carbonization process alone may suffice, eliminating the need for activation and thus potentially lowering the cost of this technology.

2.3 Mercury Retention: The Role of SO₂

Inhibition of mercury capture could occur because carbon material is a catalyst for the oxidation of SO₂ to sulfuric acid or through adsorption of SO₃ (which can hydrolyze to sulfuric acid), with SOx competing with mercury for the same adsorption sites, specifically the Lewis base sites, on the carbon surface. Two carbon materials with differing pore structures were subjected to varying SO₂ concentrations to assess the effect on Hg⁰ adsorption, specifically focusing on those materials that exhibited the highest mercury retention capacity, that is, those loaded with αFe₂O₃ (CXFe(h) and AXFe(h)). Figure 3 displays the mercury retention capacity of CXFe(h) and AXFe(h) at different SO₂ concentrations, highlighting their contrasting textural properties. For CXFe(h), mercury retention remained relatively stable across the SO₂ concentration range. In contrast, AXFe(h), characterized by a more developed porosity, showed a progressive decrease in mercury removal as SO₂ concentration increased (Figure 3). This suggests that SO₂ competes with Hg⁰ for surface and lattice oxygen active sites, thereby reducing Hg⁰ adsorption in the material (M) via the well-known Mars–Maessen mechanism.^{[35, 37, 38]}

()

()

()

()

It is generally assumed that the first step, namely, the physisorption process (I), is favored by a higher surface area, as seen in AXFe(h) (Table 1). However, unlike CXFe(h), a more developed microporosity preferentially promotes the adsorption of SO₂ over Hg⁰ on the same active sites, supporting a mechanism of competitive adsorption between SO₂ and Hg^0[32] (Figure 3). Therefore, the findings of this study demonstrate that in flue gas atmospheres, an adsorbent with increased microporosity does not necessarily achieve higher Hg retention efficiency or better tolerance to acidic gases like SO₂, which is critical for designing a cost-effective technology optimized for mercury capture.

Although evaluating this effect is not the primary objective of this study, as it has already been investigated in various sorbents, it is important to note that hematite active sites may be consumed during retention experiments. However, this deactivation can be counteracted by the presence of O₂ in the gas stream, which facilitates the regeneration of the sorbents.^{[29, 35, 39]}