Abstract
Capturing gaseous mercury (Hg0) from sulfur dioxide (SO2)-containing flue gases remains a common yet persistently challenge. Here we introduce a low-temperature sulfur chemical vapor deposition (S-CVD) technique that effectively converts SO2, with intermittently introduced H2S, into deposited sulfur (Sd0) on metal sulfides (MS), facilitating self-sustained adsorption of Hg0. ZnS, as a representative MS model, undergoes a decrease in the coordination number of Zn–S from 3.9 to 3.5 after Sd0 deposition, accompanied by the generation of unsaturated-coordinated polysulfide species (Sn2–, named Sd*) with significantly enhanced Hg0 adsorption performance. Surprisingly, the adsorption product, HgS (ZnS@HgS), can serve as a fresh interface for the activation of Sd0 to Sd* through the S-CVD method, thereby achieving a self-sustained Hg0 adsorption capacity exceeding 300 mg g−1 without saturation limitations. Theoretical calculations substantiate the self-sustained adsorption mechanism that S8 ring on both ZnS and ZnS@HgS can be activated to chemical bond S4 chain, exhibiting a stronger Hg0 adsorption energy than pristine ones. Importantly, this S-CVD strategy is applicable to the in-situ activation of synthetic or natural MS containing chalcophile metal elements for Hg0 removal and also holds potential applications for various purposes requiring MS adsorbents.
Similar content being viewed by others
Introduction
Mercury pollution has aroused global concern mainly ascribed to the volatility, insolubility, and long-range transport of gaseous elemental mercury (Hg0), which is the focus of mercury abatement1,2,3,4. Effective control of Hg0 pollutant necessitates the use of functional materials that exhibit high activity, stability, and tolerance to flue gas conditions5,6,7. Recent developments in materials for Hg0 capture include carbon-based8,9, oxide-based10, and noble metal-based materials11,12. However, the presence of SO2 in any flue gas has been known to negatively impact Hg0 removal, leading to surface sulfation or the occupation of active sites13. More unfortunately, the non-ferrous metal smelting industry is considered as the largest single source of Hg0 emissions, where high concentrations of SO2 and Hg0 co-exist14. Therefore, achieving large-capacity adsorption of Hg0 at high concentration of SO2 remains a significant challenge.
While many sulfur-based materials have shown a degree of resistance to SO2 in Hg0 adsorption15,16, their capacities often experience significant suppression due to active site depletion or deactivation, particularly at high SO2 concentrations17. This limitation frequently requires off-line regeneration under harsh conditions (e.g., heating or acidification treatments with irreversible destruction) or even necessitates the replacement of adsorbents18,19. To address the challenge of active depletion or deactivation, a more cost-effective and convenient method involves the continuous replenishment of active sites in situ at the interface of sulfur-based materials. Recognizing that the performance of metal sulfides (MS) relies heavily on the quantity of active sulfur sites20, a viable strategy is to directly convert SO2 to sustainably replenish surface active sites, thereby turning the negative effect of SO2 into a positive one. However, the high average S−O bond energy of SO2 (548 kJ mol−1) necessitates high-temperature conditions (>2000 °C) for its decomposition21,22, whereas the preferential reaction of SO2 with flue gas O2 impedes the feasibility of this pathway23. Notably, the assistance of H2S can lower the S−O bond breaking energy barrier (139 kJ mol−1) and further reorganize the S−S bond to generate elemental sulfur through the Claus reaction24,25. Fortunately, H2S or its raw materials (Na2S or NaHS) are easily accessible and commonly used for heavy metals removal from various wastewaters in non-ferrous smelters26,27.
However, two key challenges persist in achieving our objectives. Firstly, it involves effectively generating fresh sulfur on MS surface. More importantly, it pertains to activating the deposited sulfur (Sd0) for site replenishment instead of allowing it to aggregate into the inert S80 state (octatomic ring structure with poor Hg0 adsorption activity28,29). To address these challenges, we have developed a sulfur chemical vapor deposition (S-CVD) method using the Claus reaction between excessive SO2 in flue gases and intermittently added H2S. This method facilitates the deposition of gas-phase sulfur species with high controllability and scalability. Furthermore, an active interface is crucial to bonding with Sd0 to create unsaturated coordination sites rather than coordination-saturated S8. Notably, the incorporation of anchoring sites to bond with sulfur can maintain its unsaturated state30. Chalcophile elements exhibit a natural tendency to lose outer electrons to form an 18-electron outermost structure (s2p6d10), which in turn combines with sulfur (3s23p4) to form an ionic compound under ambient conditons31,32. Thus, MS containing chalcophile metals emerges as promising candidates to bond with Sd0, preventing it from falling into a saturated-coordinated ring structure.
Hence, this work employs the proposed in-situ S-CVD method on chalcophile MS to counteract the negative effects of SO2 and achieve self-sustained Hg0 adsorption, enabling in-situ reactivation without the need to replace spent adsorbents. Variety of experimental conditions and characterization methods, such as scanning electron microscopy (SEM), X-ray absorption fine structure (XAFS), and density functional theory (DFT) calculations, were devoted to evaluating the self-sustained adsorption performance, revealing the deposition process of Sd0, identifying the formation of unsaturated coordination environments, and calculating the energy changes to interpret the self-sustained adsorption mechanism. The results indicate that Sd0 can be efficiently activated to polysulfide (Sn2−, named Sd*) species by chalcophile MS, including the formed HgS itself, ensuring self-sustained Hg0 adsorption. This in-situ S-CVD approach provides a promising solution to active site depletion and poisoning issues and offers an avenue for efficient and continuous heavy metal removal using MS materials.
Results
Establishment of in-situ S-CVD method for SO2 deposition
The in-situ S-CVD method was established for flue gas SO2 deposition. To initiate S-CVD, a small amount of H2S (100 ppm) was injected into the SO2-containing flue gas upstream of Al2O3@MS adsorbents (Supplementary Fig. 1). In actual non-ferrous smelting processes, flue gas particle-bond mercury (Hgp) and oxidized mercury (Hg2+) can be respectively removed by an electrical precipitator and scrubber, resulting in a subsequent flue gas with high concentrations of SO2 and Hg0 (Supplementary Fig. 2). Extraction of approximately 0.1‰ of total SO2 for on-site conversion to H2S27,33 can satisfy the needs of S-CVD, and Hg0 will be removed by the proposed self-sustained adsorption method on adsorbents (Fig. 1a). In this method, MS functions as the fresh surface for the initial S-CVD and Hg0 adsorption, and then the spent MS (i.e., MS-HgS) acts as new surface for further S-CVD and Hg0 adsorption, ultimately achieving the sustained adsorption by HgS itself (Fig. 1b).
During the S-CVD process, the reaction ratio of H2S and SO2 was monitored as 2.1: 1 and chemical composition analysis exhibited that the sulfur content in different Al2O3@MS increased by 2.1%−3.0% after 180 min of S-CVD (Supplementary Fig. 3). These confirm the occurrence of the Claus reaction on the Al2O3@MS surface, which is a critical step in the S-CVD process. Furthermore, the optimal adding sequence of H2S and SO2 was investigated to understand the generation mechanism of Sd0. The results demonstrated that the activity of Al2O3@MS pretreated with H2S followed by SO2 was significantly lower than that pretreated with SO2 followed by H2S (Supplementary Fig. 4). Meanwhile, pretreatment only by SO2 cannot enhance the activity of Al2O3@MS. This finding indicates that the formation of Sd0 on Al2O3@MS surface followed the Eley-Rideal mechanism, in which SO2 is first adsorbed on adsorbent surface and then reacts with gaseous H2S to produce Sd0 (Eq. 1, 2):
Importantly, the negative Gibbs free energy (ΔG0 = −91 kJ mol−1 at 25 °C) of the Claus reaction and the much lower concentration of H2S (100 ppm) compared to SO2 (≥5000 ppm) used in S-CVD guarantee the sufficient reaction of added H2S.
Deposited sulfur activation on MS for Hg0 removal
The surface properties of metal sulfides play a vital role in S-CVD process. Both synthetic and natural MS served as the deposition surface (Fig. 2a). In light of Goldschmidt geochemical classification of the elements, metals with chalcophile nature have higher affinity towards sulfur, which could accelerate the stimulation of Sd0 (refs. 31,32). Thus, a range of typical chalcophile metal elements, including Cu, Zn, In, Cd, Pb, and Sn, were chosen as candidates for the synthesis of MS, while some siderophile metals, including Mn, Fe, Co, and Ni, were offered as contrasts (Fig. 2b). As depicted in Fig. 2c and Supplementary Table 1, after 15 min of S-CVD, various Al2O3@MS-Sd showed substantial differences in enhancing their Hg0 adsorption performances, of which all the chalcophile Al2O3@MS-Sd exhibited significantly increase in their adsorption capacities. Notably, Al2O3@CuS-Sd demonstrated a remarkable increase the Hg0 adsorption capacity, rising from 178.9 to 640.4 mg mol−1(within 180 min and normalized to MS molar mass). Similarly, Al2O3@ZnS-Sd exhibited a significant enhancement, with adsorption capacity increasing from 11.0 to 457.3 mg mol−1. Al2O3@CdS-Sd, Al2O3@In2S3-Sd, Al2O3@PbS-Sd, and Al2O3@SnS-Sd also showed considerable improvements in their Hg0 adsorption capacities, reaching 454.4, 431.7, 401.6, and 275.9 mg mol−1, respectively. However, the adsorption capacities of Al2O3@Fe2S3-Sd, Al2O3@CoS-Sd, Al2O3@MnS-Sd, and Al2O3@NiS-Sd were comparatively lower, reaching 207.4, 165.3, 139.7, and 125.4 mg mol−1, respectively. In addition, the Al2O3-Sd, Al2O3@ZnSO4-Sd, and Al2O3@Na2S-Sd did not show improved performance compared to their raw materials (Supplementary Fig. 5), implying that the isolated presence of metal or sulfur sites cannot directly activate the deposited Sd0.
Further, to quantitatively explore the role of metal sites in MS on the activation of Sd0, we further construct the relationship between the Hg0 adsorption capacity increment (Qi = QAl2O3@MS-Sd − QAl2O3@MS) and metal-sulfur (M−S) bond energy (measured by bond length34, Supplementary Table 2) of the investigated metal sulfides. As depicted in Supplementary Fig. 6, the Qi of chalcophile MS showed a negative relationship with the increase of M−S bond length. However, for siderophile MS, higher M−S bond length instead led to relatively higher Qi. It can be deduced that the activity of Sd0 on the sulfide interface highly depends on the geochemical characteristics of metal element and the M−S affinity. Notably, Hg also belongs to chalcophile elements (Fig. 2b), and it is supposed that HgS itself could potentially play a role in activating Sd0 (see later in Fig. 5e).
We then choose ZnS as a representative model to optimize the reaction conditions owing to its significant performance enhancement (~42 times) and simple metal and sulfur speciation. The sulfur content in Al2O3@ZnS-Sd exhibited a three-stage growth pattern with increasing S-CVD time (Fig. 3a). The growth rate of Sd0 decreased from 2.5 mg g−1 min−1 at 0−5 min to 0.2 mg g−1 min−1 at 5−30 min then to 0.06 mg g−1 min−1 at >30 min (Supplementary Table 3). The decreasing formation rate indicates that the fresh surface of the adsorbent was gradually covered by Sd0, and once the surface was completely covered, subsequent Sd0 would generate on the existing Sd0 layer, resulting in a final slow but steady growth rate. Figure 3b presents the relationship between Sd0 increment and Hg0 removal efficiency of Al2O3@ZnS-Sd. The results showed that the removal efficiency gradually increased to 71.7% when the ratio of Sd0/Zn increased to 2.7 (corresponding to 60 min of S-CVD); while, as the ratio further increased to 4.1 (240 min), 6.0 (480 min), and 13.5 (1440 min), the removal efficiency remained at a stable level of around 83%. This indicates that excess deposited Sd0 did not work, presumably attributed to the inevitable aggregation of excess Sd0 into inert S80 (ref. 29). When the adsorption capacity is normalized to the mole of Sd0 (red curve in Fig. 3b), it reached a maximum of 467.5 mg mol−1 at the Sd0/Zn ratio of 2.1 (15 min) and then gradually decreased to 89.3 mg mol−1 as the ratio increased to 13.5. Therefore, 30 min of S-CVD was chosen as the optimal condition by integrating the Hg0 removal efficiency and Sd0 utilization. Besides S-CVD time, reaction temperatures and other flue gas components also played significant roles. Thermogravimetric analysis (TGA) results indicated the high thermal stability of Al2O3@ZnS-Sd below 200 °C (Supplementary Fig. 7). The increase in adsorption temperature from 60 to 120 °C improved the Hg0 removal efficiency of Al2O3@ZnS-Sd from 23.8% to 89.9% within 180 min (Supplementary Fig. 8). However, further elevation of temperature to 140 °C and 160 °C resulted in decreasing activity to 86.6% and 75.9%, respectively, presumably due to the re-decomposition of partially captured Hg0 (Supplementary Fig. 9). Moreover, Al2O3@ZnS-Sd demonstrated high tolerance to different gas components (Supplementary Fig. 10). The addition of 5000 ppm SO2 in adsorption process resulted in enhanced removal efficiency by 3.2% at 120 °C, and further addition of 5% O2 or 100 ppm NO had slight influence with a reduction of 2.6% and 6.8%, respectively. The introduction of 5000 ppm SO2 + 4% H2O showed negative effect on Hg0 adsorption performance of Al2O3@ZnS-Sd, presumably due to the competing adsorption between H2O and Hg0 on the active sites35 and the hydrophilicity of Al2O3 (ref. 36); however, its removal efficiency maintained at a stable level of ~70% without reduction for 180 min adsorption.
Long-term experiments were conducted to clarify the SO2 effect on Hg0 removal during the S-CVD and adsorption process under the optimal condition (H2S = 100 ppm, temperature = 120 °C, S-CVD time = 30 min) (Fig. 3c). In stage I (S-CVD process), without the presence of SO2, Al2O3@ZnS initially exhibited a temporary increase in Hg0 removal but rapidly decreased once the S-CVD was stopped. This improvement might be attributed to the sulfuration of residual metal salt precursors on the surface. However, when SO2 was added during the S-CVD process, the Hg0 removal sharply increased to around 90%. Furthermore, SO2 continued to exhibit an extraordinary effect on the performance of Al2O3@ZnS-Sd after the S-CVD process. As shown in stage II (without H2S), the slope of curve without SO2 addition in the adsorption process (black line) was lower than that with SO2 addition (red one), indicating a positive effect of SO2, which is contrary to our traditional perceptions. Moreover, as the SO2 concentration increased from 0 ppm to 5000 ppm and to 33,000 ppm, the Hg0 removal after 1440 min reaction increased from 46.1% to 64.0% and to 75.0%, respectively. To verify this unexpected positive effect, SO2 was added intermittently to the SO2-free system in stage II (Fig. 3d). The results showed that once 5000 ppm of SO2 was introduced, the Hg0 removal efficiency was instantly improved by about 35%; however, once the SO2 was turned off, it dropped back down to a lower level immediately. Thus, SO2 was identified as the positive component for Hg0 adsorption over the Al2O3@ZnS-Sd. Through Hg0 temperature-programmed desorption (Hg0-TPD) experiments, we analyzed the changes in de-mercury products (Supplementary Fig. 9). The desorption peaks of spent Al2O3@ZnS-Sd in the absence of SO2 located at 210−265 °C, which are attributed to the decomposition temperature of β-HgS37. While, after adsorption in the presence of SO2, there appeared a new peak centered at higher temperature of 365 °C, which is assigned to the decomposition temperature of α-HgS38, thereby improving its Hg0 adsorption stability. Additionally, the performance of Al2O3@MS without S-CVD assistance was significantly reduced at high concentrations of SO2, excluding the promotional effect of SO2 on Al2O3@MS itself (Supplementary Fig. 11).
Self-sustained adsorption performance of Al2O3@MS-Sd
Although Al2O3@ZnS-Sd exhibited high performance and SO2 can further promote the Hg0 adsorption, it still has a limited number of active sites and requires replacement of the adsorbent once depleted. If we can periodically replenish the Sd0 on the spent Al2O3@ZnS-Sd surface, self-sustained adsorption of Hg0 can be achieved without the need for adsorbent replacement. Hence, simulated flue gas ((2.5 ± 0.05) mg m−3 Hg0, 5000 ppm SO2, 4% H2O, total flow rate = 300 mL min−1, and reaction temperature = 120 °C) was applied to investigate the self-sustained adsorption performance of Al2O3@ZnS-Sd. 100 ppm of H2S was intermittently injected into the flue gas for 30 min per 3, 6, and 24 h to replenish Sd0 (Fig. 4a–c). The Hg0 removal of Al2O3@ZnS-Sd can reach around 95% after each round of S-CVD except the first round (83.2%). After the supplemented Sd0 was consumed by Hg0, the removal efficiency gradually decreased. Meanwhile, as the number of S-CVD increased, the 3, 6, and 24 h breakthrough ratios gradually converged to ~20%, ~35%, and ~78%, respectively. Additionally, owing to the higher performance of Al2O3@CuS-Sd at low temperatures (Supplementary Fig. 12), it displayed the self-sustained adsorption performance at 60 °C, with an initial Hg0 removal efficiency of ~90% and a 24 h breakthrough ratio of ~50% in each round of reaction (Supplementary Figs. 13). Moreover, Al2O3@ZnS-Sd can restrore its original activity after Hg0 desorption and secondary S-CVD (Supplementary Fig. 14). Additionally, we evaluated the Hg0 re-emission of Al2O3@ZnS-Sd after ten rounds of reaction. At the reaction temperature, the Hg0 re-emission concentration can be reduced from 1.7 mg m−3 to 0.05 mg m−3 (emission standard for non-ferrous smelting flue gas in China) in 30 min with the assistance of S-CVD, and once the temperature dropped to room temperature, the Hg0 concentration rapidly decreased to 0 mg m−3 (Supplementary Fig. 15). Thus, the S-CVD strategy not only can fulfill the self-sustained adsorption of Hg0, but also inhibit the re-emission of adsorbed mercury.
Besides synthesized Al2O3@MS, natural sulfide ores that contain chalcophile metal elements, such as chalcopyrite and sphalerite, also enable their potential for self-sustained Hg0 adsorption. As depicted in Fig. 4d, at conditions of near-actual flue gas SO2 concentration (6%), natural sphalerite ore achieved ~95% Hg0 removal after each round of S-CVD, and the breakthrough ratio converged to ~80%. Chalcopyrite exhibited enhanced Hg0 adsorption performance with intermittent S-CVD at a lower temperature (40 °C), which reached a 100% initial Hg0 removal efficiency and had a 24 h breakthrough ratio of ~50% in each round of reaction (Fig. 4e). Thus, directly utilization of natural sulfide ores as adsorbents can effectively lower the cost for Hg0 pollution control as well as improve the adsorption capacity taking advantage of their self-sustained adsorption properties.
Figure 4f shows the macrophotographs of Al2O3@ZnS in different reaction stages. After the first round of S-CVD, the color of Al2O3@ZnS changed from white to pale yellow, verifying the formation of Sd0. Then, the color of spent Al2O3@ZnS-Sd gradually turned black after 140 h reaction (six rounds of S-CVD and Hg0 adsorption). The cross-sectional view of spent Al2O3@ZnS-Sd pellets (after ten rounds) in Fig. 4f depicted an obvious black shell. The energy dispersion X-ray spectroscopy (EDS) mapping images and the selected line scanning curves of the cross-section of Al2O3@ZnS revealed that the contents of S and Zn elements on Al2O3 pellet gradually decreased from outside in (Supplementary Fig. 16). After reaction, S and Hg elements were more scattered in the outer layer and exhibited synchronous linear increase at the boundary layer (~50 μm); while, there presented a decline in Zn element at this boundary layer (Fig. 4g). In addition, the S and Hg contents on the surface of the spent Al2O3@ZnS-Sd was much higher than the Zn content (Supplementary Fig. 17). This indicates a layer-by-layer outward deposition of Sd0 and adsorption of Hg0 on Al2O3@ZnS surface. Moreover, assume that the Zn: S ratio in raw Al2O3@ZnS is 1:1, the ratio of Hg to Sd0 in spent Al2O3@ZnS-Sd was calculated as 1.04 in light of the ESD result of the cross-section (Supplementary Fig. 18), giving an indication that Sd0 atoms were fully utilized for Hg0 adsorption.
Mechanism for self-sustained Hg0 adsorption on Al2O3@MS-Sd
Identifying the Sd0 activation and Hg0 adsorption behaviors on Al2O3@ZnS contributes to a profound insight into the self-sustained adsorption mechanism of Hg0. The Brunauer−Emmett−Teller (BET) surface area, total pore volume, and average pore size of Al2O3, Al2O3@ZnS, and Al2O3@ZnS-Sd were not significantly different (Supplementary Table 4), suggesting that the coating of ZnS and the deposition of Sd0 did not affect the pore structure of Al2O3. The X-ray diffraction (XRD) pattern of Al2O3@ZnS-Sd showed that, besides the diffraction peaks assigned to γ-Al2O3 (JCPDS no. 79-1558) and sphalerite ZnS (JCPDS no. 77-2100), no peaks related to elemental sulfur emerged (Fig. 5a), suggesting the amorphous structure of formed Sd0. The Raman spectrum of Al2O3@ZnS-Sd charactered the peaks at 155.6, 223.9, 443.8, and 475.3 cm–1, which are related higher than Raman shift of S80 (Fig. 5b). This indicates that the surface ZnS can change its S–S vibration of Sd0 and prevent its aggregation39. In the X-ray photoelectron spectroscopy (XPS) S 2p spectrum of Al2O3@ZnS-Sd, aside from the peaks ascribed to S2− (161.7 eV and 162.9 eV)40 and SO42− (168.4 eV and 169.6 eV)41, new peaks related to Sn2− (163.4 eV and 164.6 eV)42 occurred (Fig. 5c). Additionally, the proportion of Sn2− in Al2O3@CuS-Sd increased from 21.6% to 29.7% (Supplementary Fig. 19, Supplementary Table 5), demonstrating the consistency of the sulfur chemical state on chalcophile metal sulfides. However, in comparison, Al2O3-Sd featured its characteristic peaks of S 2p3/2 at binding energies of 164.0 eV and 167.8 eV (Supplementary Fig. 20), which were ascribed to S80 and adsorbed unreacted SO2 species, respectively43. Moreover, the binding energy of Zn2+ 2p3/2 in Al2O3@ZnS-Sd shifted from 1022.0 eV to 1021.8 eV after Sd0 deposition (Fig. 5d), indicating the formation of unsaturated coordination environments44. To better observe the microscopic changes and the dynamic evolution of deposited Sd0, pure ZnS was further synthesized in the same way without adding Al2O3 pellets to directly serve as support for S-CVD process. The transmission electron microscope (TEM) images of ZnS-Sd showed a decrease in the contrasts of zinc atoms (Supplementary Fig. 21), indicating the formation of Zn defects19. The XAFS S L-edge spectra confirmed the formation of Sn2− species in ZnS-Sd (Supplementary Fig. 22a). The extended XAFS (EXAFS) Zn K-edge spectra further revealed a decrease from 3.9 to 3.5 in the coordination number of Zn to S atoms in ZnS-Sd compared to that in pristine ZnS (Supplementary Fig. 22b and c, and Supplementary Table 6), further certifying the formation of unsaturated coordination sites. The XPS depth profiling results depicted that with the Ar+ etching depth increased to 6 mm, the average valence state of Sn2− decreased, suggesting a shortening of the Sn2− chain length close to the ZnS surface (Supplementary Fig. 23). Furthermore, in-situ Raman spectra revealed conversion of S8 to Sn2− in ZnS-Sd at elevated temperatures (Supplementary Fig. 24), which is also demonstrated by Fourier transform infrared spectroscopy (FTIR) and 13C nuclear magnetic resonance (NMR) using propylene as an indicator (Supplementary Fig. 25). Therefore, these imply that Sd0 does not simply physically accumulate on the Al2O3@ZnS surface in form of S80 (Eq. 3), but can be activated by Zn atoms and generated chemically bonded Sn2− (Sd*) with unsaturated coordination environments (Eq. 4):
The kinetic simulation revealed that the Hg0 adsorption over Al2O3@ZnS-Sd closely followed the pseudo-first-order kinetic model (Supplementary Fig. 26), emphasizing the important role of the external surface area for Hg0 adsorption progress. Notably, after Hg0 adsorption, there brought out the crystal of β-HgS (JCPDS no. 75-1538), which characterized its main diffraction peaks at 26.4°, 30.6°, 43.8°, and 51.9°, in the XRD pattern of spent Al2O3@ZnS-Sd (Fig. 5a). The Raman spectrum of spent Al2O3@ZnS-Sd also verified the formation of Hg−S bonds that located at 249.6 cm−1 and 343.6 cm−1 (Fig. 5b)45. Besides, the proportion of Sn2– species in XPS spectra of spent Al2O3@ZnS-Sd decreased from 36.7% to 3.0%, while the proportion of S2– increased from 35.4% to 82.9% (Fig. 5c, Supplementary Table 5). Moreover, the Hg 4f spectrum of spent Al2O3@ZnS-Sd exhibited the characteristic peaks of Hg2+ 4f7/2 and 4f5/2 centering at 100.3 eV and 104.3 eV, respectively (Supplementary Fig. 27). This signifies that adsorbed Hg0 can combine with the active Sd* on Al2O3@ZnS-Sd to form HgS:
As the 24 h breakthrough rate increased and stabilized with increasing S-CVD rounds, and HgS crystal was observed in the spent Al2O3@ZnS-Sd, we suspect that the adsorbent surface was gradually covered by produced HgS, which may also possess the ability to activate Sd0 owing to the chalcophile nature of Hg element46. Upon this, Al2O3@HgS was synthesized and used to testify the performance of Al2O3@HgS-Sd for Hg0 adsorption. As shown in Fig. 5e, Al2O3@HgS-Sd exhibited an initial Hg0 removal efficiency of ~96% and an 24 h breakthrough ratio of ~80% in each round of experiment, which is close to that of Al2O3@ZnS-Sd at the tenth round. As shown in Fig. 5f, compared with Al2O3@HgS, the XPS S 2p spectrum of Al2O3@HgS-Sd brought out new sulfur species located at 163.7 eV and 164.9 eV, which were between the binding energies of Sn2– in Al2O3@ZnS-Sd and S80 in Al2O3-Sd. This indicates that the average chemical valence of Sd* on Al2O3@HgS-Sd was slightly higher than that on Al2O3@ZnS-Sd, explaining the decrease in Hg0 adsorption on Al2O3@ZnS with increasing S-CVD rounds. The Hg 4 f spectra showed that the locations of Hg2+ in Al2O3@HgS-Sd shifted to lower binding energies compared to those of Al2O3@HgS (Supplementary Fig. 28). Moreover, the relationship between the Qi of Al2O3@HgS-Sd and the Hg–S bond energy likewise fell within the negative correlation trend of chalcophile MS (Fig. 2c, Supplementary Fig. 29, Supplementary Table 2).
Notably, as the segmented linear fitting results in Fig. 5g present, with the reaction proceeded, the adsorption rate of Al2O3@ZnS-Sd at 120 °C decreased from 1.6 mg g–1 h–1 to 1.1 mg g–1 h–1, which converged to that of Al2O3@HgS-Sd (1.1 mg g–1 h–1). The adsorption rate of Al2O3@CuS-Sd at 60 °C decreased from 1.5 mg g–1 h–1 to 1.4 mg g–1 h–1, which was also converged to that of Al2O3@HgS-Sd (1.3 mg g–1 h–1) (Supplementary Fig. 30). This validates the gradual transition of Sd0 activation and Hg0 adsorption from the raw MS surface to the self-sustained HgS surface, in accordance with the concept presented in Fig. 1b. This verifies the promoting effect of MS on Sd0 is gradually replaced by HgS. Importantly, disregarding the effect of increasing adsorbent size caused by the formation of HgS layer, theoretically, the Hg0 adsorption capacity of Al2O3@ZnS-Sd can be continuously increased due to the self-sustained adsorption performance of HgS itself (Fig. 5h). Thus, according to the adsorption rate, the adsorption capacity of Al2O3@ZnS-Sd can achieve (54.2 + 25.1d) mg g−1 (d > 6 days), thereby breaking the saturation limitations and realizing multilayer adsorption.
DFT calculations were applied to elaborate the Hg0 self-sustained adsorption mechanism on ZnS. ZnS (111) model was established and optimized to investigate the Gibbs free energy (ΔG) of Sd0 activation and the adsorption energy (Eads) of Hg0. Given that Sn2– dominated in the Hg0 adsorption process, the ΔG from S8 ring to Sn (n = 6, 4, 2, and 1) chains were first calculated. As presents in Supplementary Fig. 31, the ZnS-S8 can spontaneously convert to ZnS-S6 and then to ZnS-S4 with a negative ΔG of −77.9 kJ mol−1, while the conversion of ZnS-S4 to ZnS-S2 has a positive ΔG of 158.6 kJ mol−1. This indicates the most stable structure of ZnS-S4. Moreover, compared with ZnS, the Zn–S bond length in ZnS-S4 surface increased from 2.31 Å to 2.38–2.48 Å (Supplementary Fig. 32), in line with the decrease in Zn–S coordination number in ZnS-Sd. The Eads of Hg0 adsorption on ZnS, ZnS-S8, and ZnS-S4 were calculated as –46.2, –67.6, and –125.0 kJ mol−1, respectively (Fig. 5i). The highest Eads of ZnS-S4 verifies the important role of S4 chain on Hg0 adsorption. Further, considering the formation of HgS on Al2O3@ZnS-Sd after self-sustaied adsorption of Hg0, we constructed the ZnS@HgS structure for subsequent Sd0 activation and Hg0 adsorption. The negative ΔG (–68.4 kJ mol−1) from ZnS@HgS-S8 to ZnS@HgS-S4 demostrated its spontaneous convertion process. The Eads of Hg0 on ZnS@HgS-S4 (–73.1 kJ mol−1) was higher than those of ZnS@HgS (–26.0 kJ mol−1) and ZnS@HgS-S8 (–27.9 kJ mol−1). These confirm the role of HgS in the activation of Sd0 and further adsorption of Hg0, supporting the proposed Hg0 self-sustained adsorption mechanism.
Discussion
Figure 6a and Supplementary Table 7 compare the Hg0 adsorption capacities of Al2O3@ZnS-Sd with various reported adsorbents. The adsorption capacity of carbon- and oxide-based adsorbents were generally lower than 2 mg g–1 and 10 mg g–1, respectively, and their performance was severely inhibited by SO2. Among the reported sulfide-based adsorbents, nano-CuS exhibited the highest capacity of 122.4 mg g–1; however, it decreased to 89.4 mg g–1 in the presence of SO2 and H2O (ref. 16). Additionally, our previous work found that under scaled-up conditions, ~1 mm Al2O3@CuS only had the normalized saturated adsorption capacity of 21.0 mg g–1 (ref. 41). Another recently reported in-situ acid etching method can boost the adsorption capacity of ZnS to 53.8 mg g–1 (ref. 19), but it decreased to 1.4 mg g–1 under the scale-up conditions (Supplementary Fig. 33). Impressively, the self-sustained Al2O3@ZnS-Sd not only reversed the poisoning effect of SO2 but also reached a Hg0 adsorption capacity of 303.9 mg g–1 (normalized to ZnS coating amount, and with 24-h breakthrough ratio of ~80%) after 10 rounds of reaction, which is over 250, 60, and 8 times higher than the average of reported carbon-, oxide-, and sulfur-based adsorbents, respectively. Moreover, as the self-sustained adsorption process continues, the Hg0 adsorption capacity of Al2O3@ZnS-Sd will be increasing, thereby breaking the capacity limitations.
In summary, the Al2O3@MS-Sd adsorbents assisted with intermittent S-CVD can reverse SO2 poisoning effects and have great potential for efficient and cost-effective Hg0 removal from SO2-containing flue gases. This study demonstrates the crucial role of HgS, whether synthesized or formed by Hg0 adsorption on Al2O3@MS-Sd, in activating of Sd0 into Sd* (Eqs. 6 and 7). Therefore, this enables the self-sustained adsorption of Hg0 on Al2O3@MS-Sd surface-like chain reactions (Eqs. 8–10), realizing the multilayer adsorption (Fig. 6b).
Methods
Materials
The raw materials used in this work were purchased in chemical purity (>99.5%) from Sinopharm Chemical Reagent Co., Ltd and used without purification. Natural sulfide ores (chalcopyrite and sphalerite) were provided by provided by a non-ferrous smelter in Henan, China.
Preparation of adsorbents
Different metal sulfides (MS, M = Cu, Zn, Cd, In, Hg, Pb, Sn, Mn, Fe, Co, and Ni) were synthesized via a one-step hydrothermal method as the interface for S-CVD and Hg0 adsorption. According to our previous study, ~1 mm commercial γ-Al2O3 pellets were chosen as the support for MS (named Al2O3@MS) to improve gas mass transfer41. Taking Al2O3@ZnS as an example, firstly, 3.50 mL of H2O containing 2.50 mmol of ZnSO4·7H2O was added into 4.75 g of γ-Al2O3 pellets followed by 30 min ultrasonic treatment. Then, the mixture was dried at 60 °C for 9 h to obtain Al2O3@ZnSO4. After that, the pellets were poured into 50.0 mL of H2O containing 2.50 mmol of Na2S·9H2O and transferred to a 100 mL Teflon-lined autoclave. After reacting at 120 °C for 12 h, the resultant Al2O3@ZnS was collected by vacuum filtration, washed with deionized water and ethanol, and dried at 60 °C in an oven for 12 h. Other Al2O3@MS adsorbents were synthesized in the same way, except that different metal precursors, including CuSO4·5H2O (2.50 mmol), CdCl2 (2.50 mmol), InCl3·4H2O (1.67 mmol), HgCl2 (2.50 mmol), (CH3COO)2Pb (2.50 mmol), SnCl2 (2.50 mmol), MnSO4·H2O (2.50 mmol), FeCl3 (1.67 mmol), CoCl2·6H2O (2.50 mmol), and NiCl2·6H2O (2.50 mmol), were used instead of ZnSO4·7H2O.
In-situ low-temperature sulfur chemical vapor deposition
The in-situ S-CVD was conducted in a self-made fixed-bed reactor system as shown in Supplementary Fig. 1. To prevent the pre-reaction between H2S and SO2 in the pipeline (heat protection at 120 °C), a three-way quartz reaction tube was used to introduce different gas components, of which the main tube (10.0 mm inner diameter) was used to feed 5000 ppm SO2, and the side tube (4.0 mm inner diameter) embedded in the main reaction tube was used to feed 100 ppm H2S. The H2S concentration was detected by a gas chromatograph (Agilent GC 8860) equipped with a flame photometric detector. To verify the feasibility of the S-CVD strategy, we used SiO2 wafer as substrate to conduct the reaction. The SEM images of SiO2-Sd exhibited a homogeneous surface with a few grooves and the ESD mapping images showed the uniform distribution of S element (Supplementary Fig. 34a–f). Moreover, the SEM image of the cross-section of SiO2-Sd showed the formation of a deposition layer (Supplementary Fig. 34g, h).
Specifically, in this work, single S-CVD (15 min) was used to evaluate the enhanced effect of S-CVD on the activities of different Al2O3@MS adsorbents, while Al2O3@MS or natural sulfide ores treated with intermittent S-CVD (30 min each round) were applied to assess their self-sustained adsorption performance for Hg0.
Characterizations
The chemical composition of the as-prepared adsorbents was analyzed by X-ray fluorescence spectroscopy on an Epsilon 3X instrument (Netherlands). The thermal stability of the adsorbents was analyzed by a thermogravimetric analyzer (NETZSCH, STA 2500 Regulus) at a heating rate of 5 °C min−1 from 20 °C to 500 °C in N2 atmosphere. The BET-specific surface area and pore structure were detected by an automatic porosity analyzer apparatus (Quantachrome, Autosorb-iQ, USA). The XRD patterns were obtained on Shimadzu XRD-6100 (Japan) with Cu Kα radiation (scan speed: 8° min−1, and 2θ range of 10−80°). The SEM images of adsorbents were performed on ZEISS Sigma 300 (Germany), and EDS/mapping images using Oxford Xplore 30 (UK). XPS spectra were determined on an X-ray photoelectron spectrometer (Thermo Scientific ESCALAB Xi+, USA) equipped with a mono Al Kα X-ray source. TEM images were carried out on FEI Tecnai F20. The XPS depth profiling is measured by Ar+ ion etching material surface layers at different depths. XAFS spectra of S L-edge and Zn K-edge were detected by synchrotron radiation light sources (see detail in Supplementary Methods 1). Raman spectra were measured on a Raman spectrometer (Horiba LabRAM HR Evolution, Japan) with 532 nm line of Ar+ laser for excitation. FTIR spectra were recorded on a Fourier-Transform infrared spectrometer (Nicolet 6700, USA) using the potassium bromide pellet technique. 13C NMR spectra were performed on a nuclear magnetic resonance spectrometer (Bruker Avance Neo 400WB, Germany).
Gaseous mercury adsorption assessment
Hg0 adsorption performances of as-prepared adsorbents were assessed in the same reactor system. The Hg0 vapor was obtained by an Hg0 permeation device (VICI Metronics) loaded in a U-shaped glass tube. The H2O vapor was produced by a steam generator. The concentrations of both Hg0 and H2O were controlled by adjusting the water bath temperature and carrier gas (N2) flow rate. Other gas component including O2, SO2, and NO, were obtained directly from compressed gas in cylinders. The gas components of the simulated flue gas, including Hg0 (1.5−2.5 mg m−3), SO2 (5000−60,000 ppm, when used), and H2O (4%, when used), O2 (5%, when used), and NO (100 ppm, when used) were mixed evenly before entering the fixed-bed reactor. The total flow rate was controlled at 300−360 mL min−1, the adsorbent mass used in each experiment was 0.3−1.0 g. The detailed experiment conditions were summarized in Supplementary Table 8. Lumex RA 915+ was used to record the inlet and outlet Hg0 concentrations and before each experiment, the inlet concentration should remain stable (±0.05 μg m−3) for more than 10 min. The exhaust gas was absorbed by 0.1 mol L−1 potassium permanganate solution and activated carbon before discharged.
The Hg0 removal efficiency (η, %) and Hg0 adsorption capacity (Q, mg·g−1) were calculated according to following equations:
where Hg0in and Hg0out (mg m−3) are the inlet and outlet Hg0 concentrations, respectively, m (g) is the adsorbent mass, f (mL min−1) represents the total flow rate, and t (min) donates the adsorption time.
Hg0-TPD experiments were conducted to identify the Hg0 species adsorbed on the adsorbent surface and its stability. A certain amount of spent adsorbent was heated from 50 to 450 °C with a heating rate of 5 °C min−1 in pure N2 to desorb the adsorbed mercury. The signal of desorbed Hg0 was also detected by Lumex RA 915+.
Kinetics and theory calculation
Hg0 adsorption kinetic models and theoretical calculations of Sd0 activation and Hg0 adsorption behavior are detailed in Supplementary Methods 2 and 3.
Data availability
All data supporting the findings of this study are available within the paper and the Supplementary information files or available from the corresponding author upon request.
References
Obrist, D. et al. Tundra uptake of atmospheric elemental mercury drives Arctic mercury pollution. Nature 547, 201–204 (2017).
Krabbenhoft, D. P. & Sunderland, E. M. Global change and mercury. Science 341, 1457–1458 (2013).
Schaefer, K. et al. Potential impacts of mercury released from thawing permafrost. Nat. Commun. 11, 4560 (2020).
Lin, Y. et al. Minamata Convention on Mercury: Chinese progress and perspectives. Natl Sci. Rev. 4, 677–679 (2017).
Wang, D. et al. Multipollutant Control (MPC) of flue gas from stationary sources using SCR technology: a critical review. Environ. Sci. Technol. 55, 2743–2766 (2021).
Islamoglu, T. et al. Metal-organic frameworks against toxic chemicals. Chem. Rev. 120, 8130–8160 (2020).
Xu, H. et al. Heterogeneous reaction mechanisms and functional materials for elemental mercury removal from industrial flue gas. ACS EST Engg. 1, 1383–1400 (2021).
Li, J. et al. Atomically dispersed manganese on a carbon-based material for the capture of gaseous mercury: mechanisms and environmental applications. Environ. Sci. Technol. 54, 5249–5257 (2020).
Yang, W. et al. A sulfur-tolerant MOF-based single-atom Fe catalyst for efficient oxidation of NO and Hg0. Adv. Mater. 34, 2110123 (2022).
Xu, H. et al. MnOx/graphene for the catalytic oxidation and adsorption of elemental mercury. Environ. Sci. Technol. 49, 6823–6830 (2015).
Zhao, S. et al. Combined effects of Ag and UiO-66 for removal of elemental mercury from flue gas. Chemosphere 197, 65–72 (2018).
Zhang, H. et al. Nanoconfinement of Ag nanoparticles inside mesoporous channels of MCM-41 molecule sieve as a regenerable and H2O resistance sorbent for Hg0 removal in natural gas. Chem. Eng. J. 361, 139–147 (2019).
Zhang, H. et al. A review on adsorbent/catalyst application for mercury removal in flue gas: effect of sulphur oxides (SO2, SO3). J. Clean. Prod. 276, 124220 (2020).
Wu, Q. et al. Flow analysis of the mercury associated with nonferrous ore concentrates: implications on mercury emissions and recovery in China. Environ. Sci. Technol. 50, 1796–1803 (2016).
Li, G. et al. Promoting SO2 resistance of a CeO2(5)-WO3(9)/TiO2 Catalyst for Hg0 oxidation via adjusting the basicity and acidity sites using a CuO doping method. Environ. Sci. Technol. 54, 1889–1897 (2020).
Yang, Z. et al. Multiform sulfur adsorption centers and copper-terminated active sites of nano-CuS for efficient elemental mercury capture from coal combustion flue gas. Langmuir 34, 8739–8749 (2018).
Xu, H. et al. Enhancing the catalytic oxidation of elemental mercury and suppressing sulfur-toxic adsorption sites from SO2-containing gas in Mn-SnS2. J. Hazard. Mater. 392, 122230 (2020).
Kong, L. N. et al. Outstanding resistance of H2S-modified Cu/TiO2 to SO2 for capturing gaseous Hg0 from nonferrous metal smelting flue gas: performance and reaction mechanism. Environ. Sci. Technol. 52, 10003–10010 (2018).
Li, H. et al. In situ acid etching boosts mercury accommodation capacities of transition metal sulfides. Nat. Commun. 14, 1395 (2023).
Li, H. et al. Mechanisms of gas-phase mercury immobilized by metal sulfides from combustion flue gas: a mini review. Energ. Fuel 36, 6027–6037 (2022).
Kreysa, G. & Schütze, M. Corrosion Handbook. (Wiley-VCH Weinheim, Germany, 2004).
Greenwood, N. N. & Earnshaw, A. Chemistry of the Elements 2nd. (Elsevier, 2012).
Fleig, D., Andersson, K. & Johnsson, F. Influence of operating conditions on SO3 formation during air and oxy-fuel combustion. Ind. Eng. Chem. Res. 51, 9483–9491 (2012).
Kumar, M. & Francisco, J. S. Elemental sulfur aerosol-forming mechanism. Proc. Natl Acad. Sci. 114, 864–869 (2017).
Kamyshny, A., Gun, J., Rizkov, D., Voitsekovski, T. & Lev, O. Equilibrium distribution of polysulfide ions in aqueous solutions at different temperatures by rapid single phase derivatization. Environ. Sci. Technol. 41, 2395–2400 (2007).
Liang, Q., Jiang, L., Duan, N., Xu, F. & He, H. Exploration of the arsenic removal performance fluctuation using H2S from highly acidic wastewater in copper smelting. J. Clean. Prod. 376, 134311 (2022).
Sun, X. et al. Looping upcycling SO2 into value-added H2S by fast-induced reduction process for heavy metals treatment in nonferrous smelting industry. Fuel 331, 125867 (2023).
Chung, W. J. et al. The use of elemental sulfur as an alternative feedstock for polymeric materials. Nat. Chem. 5, 518–524 (2013).
Hong, Q. et al. Insight into the interfacial stability and reaction mechanism between gaseous mercury and chalcogen-based sorbents in SO2-containing flue gas. J. Colloid Interface Sci. 577, 503–511 (2020).
Zhang, T., Wang, J., Wu, P., Lu, A.-Y. & Kong, J. Vapour-phase deposition of two-dimensional layered chalcogenides. Nat. Rev. Mater. 8, 799–821 (2023).
Sarah-Jane, B. Encyclopedia of Geochemistry, Encyclopedia of Earth Sciences Series (ed. White, W. M.) 1–5 (Springer International Publishing, 2016).
Wieser, P. E. & Jenner, F. E. Encyclopedia of Geology (Second Edition) (eds Alderton, D. & Elias, S. A.) 67–80 (Academic Press, 2021).
Sun, X. et al. Roles of the comproportionation reaction in SO2 reduction using methane for the flexible recovery of elemental sulfur or sulfides. Environ. Sci. Technol. 58, 960–969 (2023).
Jain, A. et al. Commentary: The Materials Project: a materials genome approach to accelerating materials innovation. APL Mater. 1, 011002 (2013).
Liu, D., Li, C., Jia, T., Wu, J. & Li, B. Novel metal sulfide sorbents for elemental mercury capture in flue gas: a review. Fuel 357, 129829 (2024).
Li, Z. et al. Preparation of Al2O3-coated expanded graphite with enhanced hydrophilicity and oxidation resistance. Ceram. Int. 44, 16256–16264 (2018).
Hong, Q. et al. Stepwise ions incorporation method for continuously activating PbS to recover mercury from Hg0-rich flue gas. Environ. Sci. Technol. 54, 11594–11601 (2020).
Rumayor, M., Diaz-Somoano, M., Lopez-Anton, M. A. & Martinez-Tarazona, M. R. Mercury compounds characterization by thermal desorption. Talanta 114, 318–322 (2013).
Steudel, R. & Eckert, B. In Elemental Sulfur and Sulfur-Rich Compounds I 230 Topics in Current Chemistry-Series (ed. Steudel, R.) 1–79 (2003).
Hong, Q. et al. Regulation of the sulfur environment in clusters to construct a Mn–Sn2S6 framework for mercury bonding. Environ. Sci. Technol. 56, 2689–2698 (2022).
Hong, Q. et al. Adsorption of gaseous mercury for engineering optimization: from macrodynamics to adsorption kinetics and thermodynamics. ACS EST Engg. 1, 865–873 (2021).
Hao, J. et al. An in-depth study of Zn metal surface chemistry for advanced aqueous Zn-ion batteries. Adv. Mater. 32, 2003021 (2020).
Smart, R. S. C., Skinner, W. M. & Gerson, A. R. XPS of sulphide mineral surfaces: metal‐deficient, polysulphides, defects and elemental sulphur. Surf. Interface Anal. 28, 101–105 (2015).
Hao, X. et al. Zinc vacancy-promoted photocatalytic activity and photostability of ZnS for efficient visible-light-driven hydrogen evolution. Appl. Catal. B-Environ. 221, 302–311 (2018).
Minceva-Sukarova, B., Najdoski, M., Grozdanov, I. & Chunnilall, C. J. Raman spectra of thin solid films of some metal sulfides. J. Mol. Struct. 410–411, 267–270 (1997).
Frost, B. R., Mavrogenes, J. A. & Tomkins, A. G. Partial melting of sulfide ore deposits during medium- and high-grade metamorphism. Can. Mineral. 40, 1–18 (2002).
Acknowledgements
This work was partly supported by the National Key Research and Development Program of China (2022YFC3901100, N.Y.; 2022YFC3703800, H.X.) and the National Natural Science Foundation of China (No. 21976118, N.Y.; No. 52070129, Z.Q.; No. 22176121, H.X.).
Author information
Authors and Affiliations
Contributions
Q.H., H.X., Z.Q., and N.Y. conceived the idea and designed the research; Q.H. and X.S. prepared materials and performed characterizations; J.L. and W.H. assembled the test system; Q.H. and H.X. analyzed and interpreted the results; and Q.H., H.X., Z.Q., L.Z., and N.Y. contributed to the writing and revising of the paper.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Dongjing Liu, Zhilou Liu and the other, anonymous, reviewer for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Hong, Q., Xu, H., Sun, X. et al. In-situ low-temperature sulfur CVD on metal sulfides with SO2 to realize self-sustained adsorption of mercury. Nat Commun 15, 3362 (2024). https://doi.org/10.1038/s41467-024-47725-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-024-47725-3
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.