Metallic 1T/1T′ phase TMD nanosheets with enhanced chemisorption sites for ultrahigh-efficiency lead removal

Abstract

Two-dimensional (2D) materials, as adsorbents, have garnered great attention in removing heavy metal ions (HMIs) from drinking water due to their extensive exposed adsorption sites. Nevertheless, there remains a paucity of experimental research to remarkably unlock their adsorption capabilities and fully elucidate their adsorption mechanisms. In this work, exceptional lead ion (Pb²⁺) (a common HMI) removal capacity (up to 758 mg g⁻¹) is achieved using our synthesized metallic 1T/1T′ phase 2D transition metal dichalcogenide (TMD, including MoS₂, WS₂, TaS₂, and TiS₂) nanosheets, which hold tremendous activated S chemisorption sites. The residual Pb²⁺ concentration can be reduced from 2 mg L⁻¹ to 2 μg L⁻¹ within 0.5 min, meeting the drinking water standards following World Health Organization guideline (Pb²⁺ concentrations <10 μg L⁻¹). Atomic-scale characterizations and calculations based on density functional theory unveil that Pb²⁺ bond to the top positions of transition metal atoms in a single-atom form through the formation of S-Pb bonds. Point-of-use (POU) devices fabricated by our reported metallic phase MoS₂ nanosheets exhibit treatment capacity of 55 L-water g⁻¹-adsorbent for feed Pb²⁺ concentration of 1 mg L⁻¹, which is 1-3 orders of magnitude higher than other 2D materials and commercial activated carbon.

Introduction

Heavy metal ions (HMIs) in drinking water poses a grave threat to human health owing to their inherent characteristics of high toxicity, rapid mobility, and non-biodegradability^1,2,3. Among the various HMIs, lead ions (Pb²⁺), cadmium ions (Cd²⁺), and mercury ions (Hg²⁺) warrant particular concern due to their well-documented detrimental effects^4,5. Although, numerous technologies have been applied for removing these ions, including chemical precipitation⁶, solvent extraction⁷, ion exchange⁸, and membrane separation^9,10,11, their removal efficiency is not sufficient to meet the stringent requirements of extremely low HMI concentrations in drinking water (World Health Organization (WHO) guideline values are 10 μg L⁻¹, 3 μg L⁻¹ and 6 μg L⁻¹ for Pb²⁺, Cd²⁺ and Hg²⁺, respectively)¹².

Adsorption stands out as an effective strategy for the HMIs decontamination from drinking water due to its exceptional uptake capacity, remarkable selectivity, rapid kinetics and simplicity of design and operation^13,14,15. According to the hard/soft acid/base (HSAB) theory⁹, Lewis soft acids (Pb²⁺, Cd²⁺, Hg²⁺) prefer binding to Lewis soft base groups (compounds containing S, N, or O). Based on this theory, the ideal materials for the adsorption of HMIs should be equipped with rich Lewis soft base groups. Consequently, various materials, particularly two-dimensional (2D) materials like hexagonal boron nitride (h-BN)¹⁶, graphene oxide (GO)¹⁷, transition metal carbides/nitrides (MXenes)¹⁵, and transition metal dichalcogenides (TMDs)¹⁸, have been developed for HMIs removal.

TMDs, with their unique layered structure and abundant sulfur sites, are promising candidates for HMIs removal. Current research focuses on improving the sulfur atom accessibility of MoS₂ by increasing interlayer spacing^19,20 or exfoliating it into monolayer nanosheets⁹. Phase tuning can also enhance HMIs adsorption, demonstrating higher Pb²⁺ capacity and faster kinetics than the original counterpart^21,22. Additionally, MoS₂ composites with hydrogels can improve dynamic HMIs processing under continuous flow operations²³. However, the underlying mechanisms, such as adsorption sites, the morphology of adsorbed Pb²⁺, and Pb-S bond formation, remain unexplored.

In this work, we developed nearly perfect 2D adsorbents–1T/1T′ TMD nanosheets (MoS₂, WS₂, TaS₂, and TiS₂)–synthesized through electrochemical lithium intercalation based exfoliation method^24,25. Beneficial from the effective phase tuning (2H to 1T & 1T') and high yield exfoliation (single & bilayer yield >90%), the abundant exposed S chemisorption sites render these nanosheets high adsorption capacity up to 758 mg g⁻¹, rapid adsorption kinetics within 0.5 min, and minimal residual Pb²⁺ concentrations down to 2 μg L⁻¹. The adsorption mechanism and point-of-use (POU) water purification device based on the TMD nanosheets are also elucidated and demonstrated. We elucidated the mechanism of single-atom Pb adsorption through S-Pb bond formation, with Pb adsorbed exclusively on the top transition metal site.

Results and discussion

Ultrahigh-efficiency Pb²⁺ removal

The metallic 1T/1T′ phase TMD nanosheets, including MoS₂, WS₂, TaS₂ and TiS₂, were exfoliated from their bulk counterparts (Supplementary Figs. 1 and 2) via an electrochemical lithium intercalation and exfoliation method (see “Method”)²⁴. These nanosheets are ultrathin (mono- or bi-layer, high yield than 92%), metallic (1T/1T′ phase), and negatively charged (Supplementary Figs. 3 and 4)²⁶.

A batch of experiments were conducted to assess the adsorption performances of these exfoliated nanosheets (MoS₂, WS₂, TaS₂, TiS₂) towards various cations, including several toxic transition metal cations (Pb²⁺, Cu²⁺, Ni²⁺, and Zn²⁺) and some background cations (K⁺, Ca²⁺, Na⁺, and Mg²⁺) (see “Methods” and Supplementary Fig. 5). All of these nanosheets demonstrated nearly 100% removal of Pb²⁺ with initial concentration of 2 mg L⁻¹ (Fig. 1a, c, e, g). A high removal efficiency was also demonstrated toward Cu²⁺ (87–90%), but the removal efficiencies for K⁺, Ni²⁺, Na⁺, Zn²⁺, Ca²⁺ and Mg²⁺ were very low (10–50%) (Fig. 1a, c, e, g, Supplementary Tables 1–4 and Supplementary Figs. 6, 7).

Fig. 1: Removal of various cations by metallic 1T/1T′ phase TMD nanosheets.

a, c, e, g Removal percentage of various cations by exfoliated TMD (MoS₂, WS₂, TaS₂, TiS₂) nanosheets in individual ion solution (~2 mg L⁻¹). b, d, f, h The corresponding distribution coefficients K_d for various cation species. Pb²⁺ removal by exfoliated TMD nanosheets in the presence of Na⁺, Ca²⁺ or tap water impurities (i: MoS₂, j: WS₂, k: TaS₂, l: TiS₂). DI, Deionized water. m Isotherm of Pb²⁺ adsorption by TMD (MoS₂, WS₂, TaS₂, TiS₂) nanosheets. C_e represents the equilibrium concentration. n Kinetics curve of Pb²⁺ adsorption by TMD (MoS₂, WS₂, TaS₂, TiS₂) nanosheets, inset shows the enlarged detail for the range of 0–1 min. o Overall performance comparison between MoS₂ and representative adsorbents for Pb²⁺ removal. p Comparison of the maximum adsorption capacity (q_max) and distribution coefficient (K_d) of TMD (MoS₂, WS₂, TaS₂, TiS₂) nanosheets developed in this work (see inset shadings) with those of other benchmark materials reported in the literatures. The details can be found in Supplementary Table 5.

To compare the affinity of TMD nanosheets, the distribution coefficients (K_d) for each cation were calculated (see “Methods” and Fig. 1b, d, f, h). The MoS₂, WS₂, TaS₂ and TiS₂ nanosheets exhibited superior K_d values of 5.57 × 10⁷, 3.59 × 10⁷, 3.90 × 10⁷ and 1.58 × 10⁷ mL g⁻¹ for Pb²⁺, respectively, which are 2 to 4 orders of magnitude higher than those for other metal cations (Cu²⁺, Zn²⁺, K⁺, Ni²⁺, Na⁺, Mg²⁺, Ca²⁺), revealing their superior selectivity/affinity toward Pb²⁺. This is attributed to the tremendous S sites (Lewis soft base groups) in exfoliated TMD nanosheets that can offer strong soft-soft interactions with the Lewis soft acids (Pb²⁺)⁹.

Considering the cations of Na⁺ and Ca²⁺ are ubiquitous in drinking water, the interference of these cations with high concentration on the adsorption capacities of TMD nanosheets toward Pb²⁺ were further investigated. Results showed that 20 mM NaNO₃, 20 mM Ca(NO₃)₂, and laboratory tap water have no detectable impact on TMDs′ removal efficiency towards Pb²⁺ (Fig. 1i–l).

Further experiments were conducted to reveal the underlying adsorption isotherms and kinetics. As shown in Fig. 1m, the adsorption isotherms of Pb²⁺ on MoS₂, WS₂, TaS₂ and TiS₂ nanosheets can be well fitted with Langmuir model (Supplementary Note 1), indicating monolayer Pb²⁺ adsorption on TMD nanosheets²⁷. According to our density functional theory (DFT) calculations, each Pb atom is stabilized by three S atoms through Pb-S bonds (Supplementary Fig. 8), resulting in a maximum stoichiometric Pb/MoS₂ ratio of 3:2 for saturated adsorption. Consequently, the theoretical adsorption capacity of MoS₂ for Pb²⁺ is estimated at 863 mg g⁻¹. Similarly, the calculated theoretical adsorption capacities for WS₂, TaS₂, and TiS₂ are 557 mg g⁻¹, 564 mg g⁻¹, and 1233 mg g⁻¹, respectively. The pH of 6 was chosen to explore the saturated adsorption capacity of MoS₂ for Pb²⁺ while preventing potential interference from hydroxide precipitates under alkaline conditions (Supplementary Note 2). This adjustment significantly boosted the saturated adsorption capacity of Pb²⁺ by MoS₂, yielding an enhanced capacity of 758 mg g⁻¹ (Supplementary Fig. 9a). Similarly, Pb²⁺ capacities for WS₂, TaS₂, and TiS₂ increased to 401 mg g⁻¹, 311 mg g⁻¹, and 574 mg g⁻¹, respectively (Supplementary Fig. 9b–d). Consequently, MoS₂, WS₂, TaS₂, and TiS₂ achieved 88%, 72%, 55%, and 47% of their theoretical capacities, respectively. The lower ratios observed for TaS₂ and TiS₂ can be attributed to surface oxidation^28,29, which hinders Pb²⁺ adsorption. MoS₂, although not reaching 100%, may be affected by sulfur vacancies or the slightly acidic pH of 6, since protons have competitive adsorption with Pb²⁺ and thereby reduce the adsorption capacity in comparison with the theoretical value. The kinetic processes for MoS₂, WS₂, TaS₂, and TiS₂ nanosheets in Pb²⁺ adsorption are shown in Fig. 1n. The rapid kinetics of Pb²⁺ adsorption by four TMD nanosheets are evident, with MoS₂ and WS₂ reducing the concentration from 2 mg L⁻¹ to 2 μg L⁻¹ in just 0.5 and 0.75 min, respectively. While TaS₂ and TiS₂ achieve this reduction to 3 μg L⁻¹ in 4 min. This performance aligns perfectly with the stringent WHO guideline value for Pb²⁺ concentrations (10 μg L⁻¹) in drinking water.

Additionally, the stability of Pb²⁺ adsorption on TMD nanosheets was also demonstrated, ensuring that it does not lead to secondary pollution (refer to Supplementary Fig. 10). Upon a comprehensive assessment of MoS₂ in comparison to other representative adsorbents reported in the literatures, we observe that the MoS₂ material developed in this study excels in nearly all aspects (as indicated in Fig. 1o, p and Supplementary Table 5). This superiority includes parameters such as adsorption capacity (q_max), distribution coefficient (K_d), removal efficiency (R), adsorption time (t), and the final concentration of Pb²⁺ after adsorption (C_f). The mass production of the MoS₂ nanosheets is easily achieved by scaling up the electrode in our method (see Supplementary Fig. 11). This scalability natures significantly demonstrates their potential in practical application for the removal of Pb²⁺.

Clarification of Pb²⁺ adsorption sites

The low-magnification transmission electron microscope (TEM) images of TMD-Pb revealed the absence of Pb particles on the nanosheets (Supplementary Fig. 12). The selected area electron diffraction patterns obtained from a flat area of the nanosheets exhibited the (100) and (110) planes of the TMDs (insets in Supplementary Fig. 12), confirming the excellent crystallinity of the TMDs, even after Pb²⁺ adsorption. The scanning electron microscopy (SEM) images coupled with energy-dispersive X-ray (EDX) mapping (see Supplementary Fig. 13) demonstrated the uniform distribution of S, Pb, and transition metal atoms (Mo, W, Ta, and Ti) across the entire nanosheets after Pb²⁺ adsorption.

The high-resolution high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images of TMD-Pb nanosheets displayed the regularly arranged bright dots (Fig. 2a, f, k, p), belonging to the periodic arrangement of the transition metal atoms. Clearly, some brighter dots (red circles) also emerged in these images (Fig. 2a, f, k, p and Supplementary Figs. 14–16), belonging to individual Pb atoms^30,31,32, suggesting the single-atom forms of Pb²⁺ adsorption. Such sing-atom Pb exactly overlap the transition metal atom sites in TMD crystal (Fig. 2b, g, l, q), revealing that the sing-atom Pb are situated on the top sites of transition metal atoms. The HAADF-STEM signal intensity analysis revealed that d_Pb-M = d_M-M (d stands for interatomic distance, M refers to the transition metal atom) (Fig. 2c, h, m, r), further indicating that Pb single-atom are precisely overlap transition metal atom sites.

Fig. 2: HAADF-STEM images of single-atom Pb adsorbed onto single/-bilayer TMD nanosheets.

HAADF-STEM images of a MoS₂-Pb, f WS₂-Pb, k TaS₂-Pb, p TiS₂-Pb. Insets: the corresponding simulated atomic arrangement model of 1T′-MoS₂, 1T′-WS₂, 1T-TaS₂ and 1T-TiS₂. The brighter dots in red circles represent the single-atom Pb. b, g, l, q The corresponding magnified view of STEM images. c, h, m, r The intensity profiles along the yellow dots shown in the magnified STEM images. d, e, i, j, n, o, s, t The initial Pb adsorption sites in each TMD (MoS₂, WS₂, TaS₂, TiS₂) structure and the finally determined adsorption sites through HAADF-STEM results. Encircled numbers mean the different Pb adsorption sites in each TMD structure.

Additionally, the S atoms within the lattice structures of MoS₂-Pb, WS₂-Pb, TaS₂-Pb, and TiS₂-Pb are almost invisible, indicating the formation of octahedral (1T) or distorted octahedral (1T′) phases in the synthesized TMDs³³. One-dimensional zigzag chains composed of Mo and W atoms were conspicuously observed and consistently appeared throughout the entire HAADF-STEM images (Fig. 2a, f). This distinctive asymmetric characteristic arises from the presence of the distorted 1T′ structure, thus confirming the exfoliated MoS₂ and WS₂ are 1T′ phases^34,35. However, this asymmetric structural feature is notably absent in the HAADF-STEM images of both TaS₂-Pb and TiS₂-Pb (Fig. 2k, p), indicating that TaS₂ and TiS₂ adopt a 1T structure³⁶. This distinction between 1T′-MoS₂ or WS₂ and 1T-TaS₂ or TiS₂ is further supported by the characteristic peaks observed in the Raman spectra (Supplementary Fig. 17). In principle, the adsorption of Pb²⁺ onto 1T′-MoS₂ or WS₂ offers six distinct types of adsorption sites due to their inherent asymmetric structures (Fig. 2d, i). While 1T-TaS₂ or TiS₂ exhibit symmetrical structures, resulting in a more limited set of three discernible adsorption sites available for Pb²⁺ binding (Fig. 2n, s). Nevertheless, experimental data have conclusively determined that only the top metal (Mo/W/Ta/Ti) sites act as the confirmed adsorption sites for Pb²⁺ binding (Fig. 2e, j, o, t).

To determine whether the adsorption of Pb²⁺ by MoS₂, WS₂, TaS₂, and TiS₂ is a physical or chemical process, we conducted synchrotron-based X-ray absorption fine structure (XAFS) measurements. The normalized X-ray absorption near-edge structure (XANES) spectra of the Pb L₃-edge for TMD-Pb samples closely resemble those of PbO or PbS (Fig. 3a), indicating that the average oxidation state of Pb species is approximately 2, and the primary form of existence Pb atoms is single-atom form (because the oxidation state of Pb species in the form of Pb particles or clusters is expected to be zero)^37,38,39. Further insights into the chemical bonding were obtained through R-space curves (Fig. 3b) derived from the extended X-ray absorption fine structure (EXAFS) spectra of the Pb L₃-edge, revealing the presence of Pb-S bonds in TMD-Pb samples, which is consistent with the standard PbS. No discernible peak corresponding to Pb-Pb bonds, like the peak observed in Pb foil, further confirms the formation of single-atom Pb. Additionally, no peaks related to Pb compounds were detected in the XRD spectra of TMD-Pb samples (Fig. 3c), further illustrating an atomic dispersion of Pb in TMD crystal lattice with the crystalline domains below the detection limit of XRD. A similar phenomenon, where no metallic particles could be detected, was also observed for other metal cations adsorbed on TMD nanosheets, even after saturation of adsorption (see Supplementary Fig. 18).

Fig. 3: Chemisorption via forming Pb-S bond.

a Normalized X-ray absorption near-edge structure (XANES) spectra at the Pb L₃-edge of Pb foil, PbO, PbS, MoS₂-Pb, WS₂-Pb, TaS₂-Pb and TiS₂-Pb. Inset shows the enlarged detail for the range of 13,010–13,080 eV. b Corresponding extended X-ray absorption fine structure (EXAFS) spectra. FT Fourier-transformed. c XRD patterns of MoS₂-Pb, WS₂-Pb, TaS₂-Pb and TiS₂-Pb that collected onto the PVDF substrate. d–g Wavelet-transform images of MoS₂-Pb, WS₂-Pb, TaS₂-Pb and TiS₂-Pb at Pb L₃ edge. R denotes radical distance. Inset in each figure show the gradient scale. h XPS Pb 4f spectra of MoS₂-Pb, WS₂-Pb, TaS₂-Pb, TiS₂-Pb and Pb(NO₃)₂. i–l XPS Mo3d, W4f, Ta4f, Ti2p for pristine exfoliated TMD nanosheets and after Pb²⁺ adsorption. Ex represents exfoliated, “arb. u.” denotes arbitrary units.

Due to its powerful resolution in both k and R spaces, wavelet transform (WT)-EXAFS analysis was applied to investigate the atomic configuration of TMD-Pb⁴⁰. Pb foil exhibited the maximum intensity of Pb-Pb signals in R-space at R = 3.5 Å and Pb-Pb signals in k-space at k = 7.5 Å⁻¹ (Supplementary Fig. 19a). PbO displayed resolved spectra of Pb-O signals in R-space at R = 1.7 Å and Pb-O signals in k-space at k = 8.0 Å⁻¹ (Supplementary Fig. 19b). In contrast, the WT contour plots of MoS₂-Pb, WS₂-Pb, TaS₂-Pb, and TiS₂-Pb (Fig. 3d–g) reveal a single intensity maximum at around R = 2.3 Å and k-space of 6.0–8.0 Å⁻¹, exclusively assigned to the Pb-S contribution. This analysis unequivocally confirms the absence of Pb-Pb bonds, indicating that no Pb clusters or particles are formed in TMD-Pb. This aligns perfectly with the findings from HAADF-STEM and XRD results, providing strong evidence for the atomically dispersed nature of Pb after adsorption on TMD structures. Moreover, these atomically dispersed Pb ions are bound to the S atoms of TMDs, forming stable S-Pb bonds that help stabilize the adsorption configuration. Consequently, these experimental findings strongly support the conclusion that Pb²⁺ adsorption on TMDs represents a form of chemical adsorption, with the S atoms of TMDs serving as the primary adsorption sites. Further evidence for chemical adsorption is drawn from the adsorption kinetics of Pb²⁺ by single/bi-layer TMD nanosheets (Supplementary Fig. 20). The excellent fit of the data to the pseudo-second-order kinetic model provides compelling evidence that the chemisorption process governs the rate-determining step in the adsorption process (Supplementary Note 3)⁵. To deepen our understanding of the interaction between Pb²⁺ and TMDs, we conducted X-ray photoelectron spectroscopy (XPS) on both pristine Pb(NO₃)₂ and Pb²⁺ adsorbed onto the TMDs. The emergence of a new peak at a binding energy of 141.0 eV was attributed to Pb 4f (Supplementary Fig. 21), confirming the successful adsorption of Pb²⁺ onto MoS₂, WS₂, TaS₂, TiS₂. In the high-resolution spectra of Pb 4f depicted in Fig. 3h, it is observed that the peaks corresponding to Pb 4f_7/2 and 4f_5/2 in TMD-Pb have shifted to lower binding energies (138.6 and 143.5 eV, respectively) compared to those of Pb(NO₃)₂ (139.5 and 144.3 eV, respectively). This shift suggests a strong interaction between TMDs and Pb²⁺, with TMDs transferring electrons to Pb²⁺²⁷. Therefore, the peaks attributed to Mo3d_5/2-3/2, W4f_7/2-5/2, Ta4f_7/2-5/2 and Ti2p_3/2-1/2, as well as the corresponding S2p_3/2-1/2 peaks, all shifted to higher binding energies after Pb²⁺ adsorption (Fig. 3i–l and Supplementary Fig. 22).

DFT calculations

We further apply the DFT calculations to investigate the corresponding adsorption behaviors for Pb²⁺ on different TMD surfaces regarding both thermodynamic trends and electronic structures. For four types of TMD materials, we have compared the adsorption energies of different metal ions on TMD surfaces based on the most stable configurations (Fig. 4a–d). For 1T′-MoS₂ and 1T′-WS₂, there are 6 possible adsorption sites while 1T-TaS₂ and 1T-TiS₂ have 3 different adsorption sites due to the higher symmetrical level of lattice structures (Supplementary Fig. 23). For most of the surfaces, we have demonstrated that Pb²⁺ are much more preferred to adsorb on the TMD surface than other metal ions due to the much lower adsorption energies. There is only one exception, where Cu²⁺ display a slightly stronger adsorption trend than Pb²⁺ on 1T-TiS₂. These different adsorption energies highly agree well with their removal efficiency (Fig. 1a, c, e, g). The stronger adsorption strengths of the Pb²⁺ benefit the further removal of the Pb²⁺ from the solution. Among different TMD surfaces, the adsorption trends of Pb²⁺ show an order of 1T′-MoS₂ > 1T-TiS₂ > 1T′-WS₂ > 1T-TaS₂, which is supportive of the removal efficiency of different TMDs (Fig. 1m). Then, we further compare the adsorption energies of Pb²⁺ on different possible sites on the surface based on the structures. For both 1T′-MoS₂ and 1T′-WS₂, it is noted that the Mo-top and W-top sites supply the most stable configurations (Fig. 4e, f). From adsorption site 1 to 6, the adsorption energies and the average Pb-S bond lengths exhibit a converse trend, in which the S-top sites with the longest Pb-S bond lengths are the most energetically unfavored adsorption sites. For 1T′-MoS₂, the S-top sites are meta-stable sites, where the Pb²⁺ is easy to migrate to the nearby stable Mo-top sites driven by the much stronger adsorption trends with largely reduced energy (~1.5 eV). A similar phenomenon is also revealed on 1T′-WS₂, where the W-top sites become an energy well, resulting in the stabilizations of Pb²⁺ from nearby meta-stable sites. It is worth mentioning that the most stable adsorption configurations of Pb²⁺ slightly deviate from the Mo-top sites, which is attributed to the reduced symmetrical level of the nearby S sites (Supplementary Figs. 24 and 25). Meanwhile, for 1T-TaS₂ and 1T-TiS₂, we discover that the three possible adsorption sites also exhibit the opposite trend between the adsorption energies and the Pb-S bond lengths (Fig. 4g, h). The metal-top sites still display the dominant advantages as the most stable adsorption configurations with the lowest adsorption energies. For Pb²⁺ adsorbed on metal-top sites, the migrations to nearby S-top sites are very difficult, which need to overcome an energy barrier of 1.51 eV and 1.01 eV for 1T-TaS₂ and 1T-TiS₂, respectively. In addition, we notice that the metal-top adsorptions are highly stable without deviations due to the higher symmetrical level of lattice, resulting in slightly decreased Pb-S bond lengths than that of 1T′-MoS₂ and 1T′-WS₂ (Supplementary Fig. 26). These energy favorable M-top sites (M = Mo, W, Ta, Ti) in 4 TMDs for Pb²⁺ adsorption are highly consistent with the HAADF-STEM results (Fig. 2).

Fig. 4: Adsorption mechanisms of Pb²⁺ on 1T/1T′ TMDs revealed by DFT calculations.

The adsorption energies of different metal ions on a 1T′-MoS₂, b 1T′-WS₂, c 1T-TaS₂, and d 1T-TiS₂. The comparisons of adsorption energies and Pb-S average bond lengths for different adsorption sites on e 1T′-MoS₂, f 1T′-WS₂, g 1T-TaS₂, and h 1T-TiS₂. The PDOS of comparisons for Pb adsorptions on i 1T′-MoS₂ (Mo-4d, S-3p represent the 4d and 3p orbitals of 1T′-MoS₂, Site1 corresponds to Pb adsorption site shown in Fig. 2d), j 1T′-WS₂ (W-5d, S-3p represent the 5d and 3p orbitals of 1T′-WS₂, Site1 corresponds to Pb adsorption site shown in Fig. 2i), k 1T-TaS₂ (Ta-5d, S-3p represent the 5d and 3p orbitals of 1T-TaS₂, Site1 corresponds to Pb adsorption site shown in Fig. 2n), and l 1T-TiS₂ (Ti-3d, S-3p represent the 3d and 3p orbitals of 1T-TiS₂, Site1 corresponds to Pb adsorption site shown in Fig. 2s). Pb(H₂O)₄²⁺ represents hydrated lead ion, E_F denotes Fermi level, Pb-6s and Pb-6p represent the 6s and 6p orbitals of Pb(H₂O)₄²⁺. “arb. u.” denotes arbitrary units.

To further reveal the electronic interactions between Pb²⁺ and TMD surfaces, we have compared the projected partial density of states (PDOS) before and after adsorptions of Pb²⁺ on the most stable metal-top sites (Site 1) (Fig. 4i–l). For Pb(H₂O)₄²⁺, it is noted that the dominant peak of Pb orbitals locate far above the Fermi level (E_F). After the adsorption on TMD surfaces, the corresponding s, p orbitals have evidently downshifted towards the E_F, which is induced by the strong electron transfer from TMDs. For all the TMD surfaces, we notice the good matching between S-3p orbitals and Pb-6s orbitals, which proves the formation of Pb-S bonds due to the chemisorption (Supplementary Fig. 8), align well with XAS results (Fig. 3d–g). Moreover, the p-p orbital overlapping between Pb and S sites is also observed, supporting the efficient electron transfer. The efficient s-p and p-p coupling guarantees stable Pb-S bond formation. In contrast, the main peak positions of d orbitals of Mo, W, Ta, and Ti show obvious mismatch with s, p orbitals of Pb sites, which suppress the electron transfer as well as bonding formation, leading to the preferences for Pb-S bond formation.

Portable point-of-use (POU) device

The regeneration of TMDs after Pb²⁺ adsorption can be achieved by employing potent chelating agents to bind with Pb²⁺. In this study, we chose ethylenediaminetetraacetic acid (EDTA) to regenerate MoS₂. Following treatment of the MoS₂-Pb sample with EDTA, XPS results displayed a significant reduction in the Pb signal (Fig. 5a). Even after five regeneration cycles, MoS₂ maintained a Pb²⁺ removal efficiency of approximately 94% (Fig. 5b). The slight decrease in removal efficiency after regeneration could be attributed to factors such as aggregation or partial oxidation of the MoS₂ nanosheets⁹.

Fig. 5: Regeneration and membrane filtration performance.

a Overall XPS full scanning spectra of MoS₂-Pb regenerated by EDTA. “arb. u.” denotes arbitrary units. b Regeneration performance of MoS₂ for Pb²⁺ removal was evaluated at an initial Pb²⁺ concentration of 2 mg L⁻¹, pH of 5, and room temperature. c Cross-sectional SEM image of MoS₂ membrane. PVDF polyvinylidene difluoride. d Schematic of the pressure assisted Pb²⁺ filtration experiment through MoS₂ membrane. e Performance of MoS₂ membranes for filtering a feedwater containing 0.25 and 1 mg L⁻¹ Pb²⁺, respectively. The treatment capacity is defined as the total effluent volume at the breakthrough point, which was reached when the effluent Pb²⁺ concentration reached 10 μg L⁻¹. Inset shows the enlarged detail for the effluent concentration range of 0–80 μg L⁻¹, with the green region indicating effluent Pb²⁺ concentrations below 10 μg L⁻¹. f Cross-sectional HAADF-STEM image of MoS₂ membrane after Pb²⁺ filtration experiment. g Schematic of the single-atom lead adsorbed into the nanochannel of laminate MoS₂ membrane. h XPS depth profile of the MoS₂ membrane after filtering Pb²⁺ solution. i Comparison of the treatment capacities (L-water g⁻¹-material) of MoS₂ membrane and other adsorptive membranes reported in the literatures. The details can be found in Supplementary Table 6.

The capacity, selectivity, and regenerability are crucial criteria when considering single-layer TMD nanosheets as potential components for developing a portable POU filter for Pb²⁺ removal from contaminated water. Such a filter should also possess attributes like flexibility, user-friendliness, and cost-effectiveness⁴¹. To demonstrate their capability, we selected MoS₂ due to its superior adsorption capacity and strong affinity for removing Pb²⁺ compared to other TMD nanosheets. Subsequently, we fabricated a MoS₂ lamellar membrane through vacuum filtration of the TMD nanosheets solution on a porous substrate (“Methods”). As depicted in Fig. 5c, the layer-by-layer structure of the MoS₂ membrane is clearly visible in the SEM cross-sectional image, with a membrane thickness of approximately 500 nm. Following this, we conducted pressure-assisted filtration experiments on the MoS₂ membrane (“Methods” and Fig. 5d) using feed water with varying Pb²⁺ concentrations from 0.25 to 1 mg L⁻¹. Due to the inherent limitations of MoS₂ in its adsorption capacity for Pb²⁺, a breakthrough point is reached once adsorption saturation is achieved. As demonstrated in Fig. 5e, when subjected to a feed concentration of 1 mg L⁻¹ Pb²⁺, the effluent volume can reach up to 200 ml before the Pb²⁺ concentration exceeds 10 μg L⁻¹, which is the WHO guideline value for Pb²⁺ concentration in drinking water¹². Additionally, the water flux was measured at approximately 5 L m⁻² h⁻¹ bar⁻¹ (Supplementary Fig. 27). By reducing the feed concentration of Pb²⁺ to 0.25 mg L⁻¹, the effluent volume can be further increased to 670 ml. Moreover, the MoS₂ membrane was regenerated after reaching saturation with Pb²⁺ by flushing it with EDTA solution. Following EDTA cleaning, a second filtration cycle was conducted, and the regenerated MoS₂ membrane was still able to reduce the Pb²⁺ concentration from 1 mg L⁻¹ to less than 10 μg L⁻¹. It exhibited a treatment volume of 184 mL, which is equivalent to 92% of the original value of a fresh MoS₂ membrane (200 ml). Furthermore, the recovery rate was 97% when decreasing the Pb²⁺ concentration from 0.25 mg L⁻¹ to below 10 μg L⁻¹ (650 ml vs 670 ml). These high recovery rates confirmed the excellent reusability of the MoS₂ membrane for Pb²⁺ filtration. After adsorption, the MoS₂ membrane was characterized by cross-sectional HAADF-STEM image as shown in Fig. 5f, which reveals the presence of single-atom Pb that identifies as brighter spots within the 2D nanochannel of the MoS₂ membrane (Fig. 5g). These Pb atoms are adsorbed on the top Mo sites, consistent with the HAADF-STEM image of depicting Pb²⁺ adsorption on 2D MoS₂ nanosheets (Fig. 2a). Furthermore, the predominant mechanism for Pb²⁺ removal, which involves Pb²⁺ adsorption into the 2D nanochannel rather than membrane rejection⁹, was reaffirmed through XPS etch measurements along the membrane stacking direction. As shown in Fig. 5h and Supplementary Fig. 28, noticeable Pb 4f intensity was observed at all depths (100–400 nm) of the MoS₂ membrane rather than only on the surface layers. The treatment capacity of the adsorptive membrane was determined by the total effluent volume at the breakthrough point. When compared with other reported adsorptive membranes (MXene¹⁸, graphdiyne²⁹, activated carbon²⁸, ion exchanger²⁸), our developed MoS₂ membrane exhibited a treatment capacity that was 1-3 orders of magnitude higher. This remarkable performance surpassed most other adsorptive membranes, including commercial ones (Fig. 5i and Supplementary Table 6). This exceptional performance can be attributed to the high adsorption capacity and selectivity of MoS₂ for Pb²⁺, as well as the full accessibility of S atoms within the 2D nanochannel for binding with Pb²⁺.

Additionally, the stability of the MoS₂ membrane in aqueous environment is noteworthy. After immersing it in water for a month, there is no apparent loosening of the membrane or detachment from the substrate (Supplementary Fig. 29a). This stability aligns with previous findings that the robust Van der Waals forces between MoS₂ nanosheets can effectively prevent the layered MoS₂ nanosheets from delaminating in water⁴². Similar membrane stability in deionized water was also observed for WS₂, TaS₂, and TiS₂ (Supplementary Fig. 29b–d). Finally, four TMD membranes were soaked in aqueous solutions with pH values ranging from 0 to 14 for 3 days, no significant differences on XRD spectra were observed between the pristine MoS₂/WS₂/TaS₂/TiS₂ membranes and after 72 h of soaking (Supplementary Fig. 30), suggesting a well-maintained lamellar configuration, demonstrating their stable lamellar structure for membrane-based filtration applications under varying pH conditions.

Based on our comprehensive characterizations in combination with theoretical calculation, Fig. 6 illustrates the successful high-yield preparation of single/bi-layer TMD nanosheets featuring metallic 1T/1T′ phases, and highlights their outstanding performance for the Pb²⁺ removal application. Here is a breakdown of the key factors contributing to this success, as summarized in Fig. 6a: 1) Precise lithium ion (Li⁺) intercalation: The electrochemical lithium intercalation and exfoliation method was employed with precise control over the amount of Li⁺ intercalated. This approach ensured the optimal exfoliation conditions, resulting in the highest yield of single/bi-layer TMD (≥92%). This maximized the exposure of the in-plane S atoms within the TMD bulk, creating an abundance of Pb²⁺ adsorption sites and ultimately leading to a high adsorption capacity. 2) Electron transfer induced phase change: The controlled electron transfer from intercalated lithium atoms to the layered TMD materials induced significant phase changes, transitioning from the 2H phase to the 1T′ phase. This phase change facilitated the uniform distribution of additional electrons or negative charges on the S sites of the metallic phase TMD nanosheets. These charged sites attracted positively charged Pb²⁺, forming strong S-Pb bonds at the top metal sites, thereby achieving rapid Pb²⁺ adsorption kinetics. 3) Efficient adsorption sites: The combination of abundant and activated (due to extra electron doping) S adsorption sites resulted in minimal residual concentrations of Pb²⁺ after the adsorption process. For the adsorption sites identification, theoretical calculations elucidated that metallic 1T′ phase MoS₂ or WS₂ possessed six stable adsorption sites, labeled as S1 (M-top), S2 (M-top), S3 (S-top), S4 (S-top), S5 (S-hollow) and S6 (S-hollow), while 1T-TaS₂ and 1T-TiS₂ possessed three stable adsorption sites, labeled as S1 (M-top), S2 (S-top), S3 (S-hollow). Among these, only the adsorption sites-S1 (M-top) and S2 (M-top) for 1T′ phase MoS₂ or WS₂, as well as S1 (M-top) for 1T phase TaS₂ or TiS₂ with the lowest adsorption energies, were capable of accommodating Pb²⁺ for adsorption. This finding aligns with our observations through HAADF-STEM, as depicted in Fig. 6b. Therefore, compared to other TMDs studied for Pb²⁺ adsorption (Supplementary Table 7), our work not only achieves remarkable adsorption performance-such as a capacity of up to 758 mg g⁻¹, rapid kinetics as fast as 0.5 min, and minimal residual content as low as 2 μg L⁻¹, but also elucidates the mechanism of single-atom Pb adsorption exclusively on the top transition metal sites through the formation of S-Pb bonds with TMDs.

Fig. 6: Schematic illustration of metallic 1T/1T′ phase TMD nanosheets for ultrahigh-efficiency lead removal.

a Through optimizing electrochemical lithium intercalation and exfoliation method, we synthesized a serial of 1T/1T′ TMD nanosheets with high single/bi-layer yields, that tremendous S sites were exposed and activated through electron doping and phase transformation, which contributed to TMD′ superior Pb²⁺ removal performances in the form of high adsorption capacity, fast adsorption kinetics and minimal Pb²⁺ residual. E denotes energy, and E_f represents fermi level. b 1T′ phase MoS₂ or WS₂ possess two stable adsorption sites: S1 (M-top) and S2 (M-top), while 1T phase TaS₂ or TiS₂ only show one stable adsorption site: S1 (M-top).

In summary, our study presents a significant advancement in the efficient removal of Pb²⁺ through the development of a series of metallic 1T/1T′ phase TMD nanosheets, including MoS₂, WS₂, TaS₂, and TiS₂, all characterized by a high yield of single/bi-layer structures with tremendous activated S chemisorption sites. Our achievements encompass rapid adsorption kinetics, minimal residual content, high adsorption capacity, exceptional selectivity, and unparalleled removal efficiency, surpassing previously reported Pb²⁺ adsorbents. Through a combination of comprehensive multimodal characterizations and rigorous theoretical calculations, these exceptional Pb²⁺ removal performances are attributed to the activation of surface S chemisorptive sites on TMD nanosheets, facilitated by electron doping during Li⁺ intercalation. The single-atom form with the formation of S-Pb bonds by Pb²⁺ adsorption mechanism is proposed and proven. Furthermore, our work extends its implications by demonstrating the potential application of MoS₂ membranes as highly effective POU filters for Pb²⁺ removal. These findings underscore the remarkable potential of metallic 1T/1T′ phase TMD nanosheets for mitigating heavy metal contamination in drinking water to align with WHO guideline values. The determined adsorption mechanism might be used to improve the HMI adsorption performance via tuning the microenvironment of the adsorption sites.

Methods

Materials

Molybdenum disulfide (MoS₂, Innochem), Tungsten disulfide (WS₂, Macklin), Tantalum disulfide (TaS₂, Sixcarbon Technology), Titanium disulfide (TiS₂, Sigma-Aldrich), N-methylpyrrolidone (NMP, Aladdin), Polyvinylidene difluoride (PVDF, Sigma-Aldrich), Natural graphite (Bay Carbon), Copper foil (Aritech Chemazone Private), Lithium foil (DodoChem), Polypropylene (pp) film (Celgard 2300), Lithium hexafluorophosphate dissolved in ethylene carbonate, dimethyl carbonate and methylene carbonate mixture (1 M LiPF₆ in EC:DMC:EMC, 1:1:1 vol.%, DodoChem), Lead sulfide (PbS, 99.995%, Aladdin), Lead (II) nitrate (Pb(NO₃)₂, ≥99.0%, Fisher chemical), Copper nitrate trihydrate (Cu(NO₃)₂.3H₂O, AR, Aladdin), Cadmium nitrate tetrahydrate (Cd(NO₃)₂.4H₂O, Aladdin), Potassium nitrate (KNO₃, ≥99.0%, Aladdin), Sodium nitrate (NaNO₃, ≥99.0%, VWR Chemicals), Nickel(II) nitrate hexahydrate (Ni(NO₃)₂.6H₂O, Duksan Pure Chemicals), Zinc nitrate hexahydrate (Zn(NO₃)₂.6H₂O, 98%, Alfa Aesar), Calcium nitrate tetrahydrate (Ca(NO₃)₂.4H₂O, ≥97%, Alfa Aesar), Magnesium nitrate hexahydrate (Mg(NO₃)₂.6H₂O, Alfa Aesar). Ethylenediaminetetraacetic acid (EDTA, Aladdin), Nitric acid (HNO₃, 69%, VWR Chemicals), Hydrogen peroxide (H₂O₂, 30%, Anaqua global international Inc. Limited), Acetone (CH₃COCH₃, 99.5%, Anaqua global international Inc. Limited). The Milli-Q water (Millipore, Billerica, MA) was used in all experiments.

Synthesis of single/bi-layer TMD nanosheets (MoS₂, WS₂, TaS₂, TiS₂)

MoS₂, WS₂, TaS₂, and TiS₂ nanosheets were prepared using an electrochemical lithium intercalation-based exfoliation method (Supplementary Fig. 3a). The process commenced with the intercalation of Li⁺ into the interlayer spaces of bulk TMDs (4 mg), achieved through the application of galvanostatic discharge conditions on a coin cell. The current and cutoff voltage are set as follows: 0.05 mA and 0.9 V for MoS₂; 0.025 mA and 0.9 V for TaS₂ and TiS₂; 0.08 mA and 0.7 V for WS₂. Following the intercalation step, exfoliation was performed by subjecting the lithiated TMD samples to sonication in DI water (10 ml). The introduction of Li⁺ into the interlayer spacing resulted in an expansion of the interlayer gaps, thereby weakening the van der Waals forces that bound the layers together. Concurrently, the generation of hydrogen gas during sonication further enhanced the efficiency of exfoliation, typically completing within a few minutes (5-10 min). The exfoliated TMD nanosheets were subsequently purified through multiple rinses with DI water. Please handle the unreacted lithium foil with care after disassembling the coin cell.

Synthesis of the TMD membranes

The purified TMD nanosheets were redispersed in DI water, and subsequently vacuum-filtrated onto a porous polymer substrate (PVDF, 220 nm pore size, 25 mm in diameter) to fabricate the TMD membrane⁴³. The thickness of the membrane was precisely tuned by controlling the volume of the filtrated solution.

Adsorption of metal cations by TMD nanosheets

In this study, batch experiments were conducted to investigate the adsorption of various metal cations onto TMD nanosheets (MoS₂, WS₂, TaS₂, and TiS₂). Nine different metal cations, including Pb²⁺, Cu²⁺, Cd²⁺, K⁺, Ni²⁺, Na⁺, Zn²⁺, Ca²⁺, and Mg²⁺, were evaluated in the form of nitrate salts to eliminate any potential interference from anions. The procedure involved mixing the TMD nanosheets solution with each individual metal cation at a concentration of 2 mg L⁻¹, followed by 12 h of stirring. After this incubation period, the TMD nanosheets with adsorbed cations were separated using 0.22 μm PVDF syringe filters, and the remaining cation concentration in the filtrate solution was determined using Inductively coupled plasma optical emission spectroscopy (ICP-OES) or Inductively Coupled Plasma Mass Spectrometer (ICP-MS) (Supplementary Fig. 5). The removal rate (R) was calculated using the following equation:

$$R=[({C}_{0}{{{\rm{\hbox{-}}}}}{C}_{f})/{C}_{f}]\times 100\%,$$

(1)

where C₀ and C_f represent the initial and final concentrations of the cations (mg L⁻¹), respectively. The distribution coefficient (K_d) was calculated using the following formula:

$${K}_{d}=[({C}_{0}{{{\rm{\hbox{-}}}}}{C}_{f})/{C}_{f}]V/m,$$

(2)

where V represents the volume of the solution (ml) and m denotes the mass of adsorbent (g).

The concentration of TMD nanosheets was determined by digesting the sample in a mixture of HNO₃ and H₂O₂ solution, followed by testing the concentration of soluble Mo/W/Ta/Ti species via ICP-OES⁷. These batch experiments were conducted at room temperature to assess adsorption capacity, selectivity, and kinetic processes. To evaluate adsorption capacity, various concentrations of Pb²⁺ ranging from 2 to 80 mg L⁻¹ were individually introduced into the TMD nanosheets solution. After 12 h of continuous stirring, the residual Pb²⁺ concentrations were quantified using ICP-MS. In order to investigate the kinetics of the Pb²⁺ adsorption process, a mixture of TMD with Pb²⁺ at a concentration of 2 mg L⁻¹ was stirred for varying durations: 1, 2, 4, 8, 16, 30, and 60 min, respectively. Subsequently, the remaining Pb²⁺ concentration was promptly determined using ICP-MS after separating the Pb²⁺ adsorbed TMD nanosheets.

Filtration of Pb²⁺ through MoS₂ membrane

The filtration experiment involving Pb²⁺ through the MoS₂ membrane was conducted at room temperature using a pressure cell (Sterlitech) with an internal volume of 300 ml. To initiate the experiment, a MoS₂ membrane (1 cm in diameter) was loaded into the cell’s chamber, which was then filled with Pb²⁺ solutions of varying concentrations (1 mg L⁻¹ and 0.25 mg L⁻¹). The filtration experiments were initiated by applying a pressure of 1 bar using compressed N₂. Subsequently, 10 mL samples of the filtrate were periodically collected and subjected to analysis using ICP-MS. To regenerate the TMD membrane after the Pb²⁺ filtration, it was thoroughly flushed with an EDTA solution to remove the adsorbed Pb²⁺ from the membrane. This regenerated membrane was then ready for use in subsequent filtration experiments.

Material characterization

SEM (Thermo Scientific Quattro S, operated at 20 kV), TEM (Tecnai G2 spirit Twin, accelerating voltage of 120 kV), STEM (collected on a double Cs-corrected JEOL-ARM300 at 80 kV), XPS (Thermo Scientific K-Alpha Nexsa, Al Kα source gun, the binding energy was calibrated via the C1s peak at 284.8 eV), XAS (Beijing Synchrotron Radiation Facility, beamline 4B9A, transmission mode), Raman (WITEC 300R Raman system, 532 nm), XRD (D2 PHASER XE-T, Cu Kα radiation source, 30 kV voltage and 10 mA current), AFM (Dimension 3100, Veeco, CA), Zeta potential (Malvern Zetasizer Nano series), ICP-OES (Optima 8000 spectrometer). ICP-MS (PE Nexion 2000) was used for detecting the concentration of cation with μg L⁻¹ level).

Calculation setup

In this work, the removal efficiency of Pb²⁺ on TMD materials is explored using DFT calculations with CASTEP version 7.0⁴⁴, embedded in Material Studio 7.0 (released in 2014). In particular, we have used the generalized gradient approximation (GGA) with Perdew-Burke-Ernzerhof (PBE) functionals to indicate the exchange-correlation interactions^45,46,47. Meanwhile, the cutoff energy has been set to 440 eV with the ultrafine quality settings and the ultrasoft pseudopotentials are considered for all the elements. During the energy minimization processes, we have applied the k-point setting to coarse quality based on the Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm⁴⁸. For the TMD surfaces, we have cleaved the (001) surfaces from single layer 1T′-MoS₂, 1T′-WS₂, 1T-TaS₂, and 1T-TiS₂ in a 6 × 6 × 1 supercell. To achieve accurate geometry optimizations, the following stringent convergency criteria have been applied for geometry optimizations, where the Hellmann-Feynman forces should not exceed 0.001 eV/Å, and the total energy difference should not be over 5 × 10⁻⁵ eV/atom. The calculation of the adsorption energy (E_ad) follows the equation:

$${E}_{{ad}}={E}_{{TMD}+{Metal}}-{E}_{{TMD}}-{E}_{{Metal}}$$

(3)

In the above equation, ${E}_{{TMD}+{Metal}}$ represents the total energy of metal ions adsorbed on TMD surfaces, ${E}_{{TMD}}$ and ${E}_{{Metal}}$ are the total energy of the TMD and metals.