Layered Nano‐Mosaic of Niobium Disulfide Heterostructures by Direct Sulfidation of Niobium Carbide MXenes for Hydrogen Evolution

MXene‐transition metal dichalcogenide (TMD) heterostructures are synthesized through a one‐step heat treatment of Nb2C and Nb4C3. These MXenes are used without delamination or any pre‐treatment. Heat treatments accomplish the sacrificial transformation of these MXenes into TMD (NbS2) at 700 and 900 °C under H2S. This work investigates, for the first time, the role of starting MXene phase in the derivative morphology. It is shown that while treatment of Nb2C at 700 °C leads to the formation of pillar‐like structures on the parent MXene, Nb4C3 produces nano‐mosaic layered NbS2. At 900 °C, both MXene phases, of the same transition metal, fully convert into nano‐mosaic layered NbS2 preserving the parent MXene's layered morphology. When tested as electrodes for hydrogen evolution reaction, Nb4C3‐derived hybrids show better performance than Nb2C derivatives. The Nb4C3‐derived heterostructure exhibits a low overpotential of 198 mV at 10 mA cm−2 and a Tafel slope of 122 mV dec−1, with good cycling stability in an acidic electrolyte.


Introduction
MAX phases, such as Ti 3 AlC 2 , are nano lamellar structures generally comprising an early transition metal element (M), a group 13-15 element (A), and carbon and/or nitrogen (X) with a composition of M n+1 AX n where n can be 1-4. [1] As the metallic bond between MA is weaker than the mixed ionicmetalliccovalent bond between MX, elements at the Asite (e.g., Al) can be selectively removed to result in layered MX structures (e.g., Ti 3 C 2 ), [2] named MXenes. [3] The typical HF acid etching in an aqueous environment converts MAX into MXene while rendering MXene sur faces terminated with O, OH, and F functionalities. [4] Since their discovery in 2011, [2] MXene has so far found a plethora of wideranging applications such as in energy storage devices, [5] electrocatalysis, [6,7] sensing, [8] and environmental remediation. [9] Each of these applications capitalizes on a specific combination of unique properties offered by MXenes: a layered 2D structure that can accommodate ions, [10] an excellent electrical conductivity (a conductivity up to 24 10 3 S cm −1 was measured for Ti 3 C 2 ), [11] a hydrophilic nature (depending on the surface terminations), [3] and rich electrochemically active sites, [7] among others.
Shortly after their discovery, research efforts have been made to modify MXenes further to improve their performances. [12] For instance, alternative etching methods, [13] such as HCl/LiF mixtures, [14] NH 4 HF 2 , [15] molten salt etching, [16] and alkali etching [17] have been proposed to favorably alter the surface ter minations, [14] interlayer spacings, [18] electronic, [19] and electro chemical [20] properties of MXene. In contrast to techniques such as layerbylayer selfassembly [21] or pillaring [20,22] of the MXene sheets where organic molecules need to be present during syn thesis, onestep modification methods often offer simplicity and more significant potential for industrialscale applications. The first report on the latter approach was the onestep conversion of Ti 3 C 2 into TiO 2 C hybrids by oxidation of the MXene precur sors in air, CO 2 , or hydrothermally. [23] The oxidation conditions in CO 2 can be adjusted so that MXene is only partially converted to achieve ternary heterostructures such as Nb 2 CCNb 2 O 5 , resulting in significantly enhanced electrochemical [24] and pho tocatalytic [25] performances compared to the parent Nb 2 C. These hierarchical heterostructures benefit from synergistic effects MXene-transition metal dichalcogenide (TMD) heterostructures are synthesized through a one-step heat treatment of Nb 2 C and Nb 4 C 3 . These MXenes are used without delamination or any pre-treatment. Heat treatments accomplish the sacrificial transformation of these MXenes into TMD (NbS 2 ) at 700 and 900 °C under H 2 S. This work investigates, for the first time, the role of starting MXene phase in the derivative morphology. It is shown that while treatment of Nb 2 C at 700 °C leads to the formation of pillar-like structures on the parent MXene, Nb 4 C 3 produces nano-mosaic layered NbS 2 . At 900 °C, both MXene phases, of the same transition metal, fully convert into nanomosaic layered NbS 2 preserving the parent MXene's layered morphology. When tested as electrodes for hydrogen evolution reaction, Nb 4 C 3 -derived hybrids show better performance than Nb 2 C derivatives. The Nb 4 C 3 -derived heterostructure exhibits a low overpotential of 198 mV at 10 mA cm −2 and a Tafel slope of 122 mV dec −1 , with good cycling stability in an acidic electrolyte.
of a conductive carbon/MXene network, a layered 2D MXene structure, and a transition metal oxide semiconductor, that is, making them electrochemically and/or catalytically active.
Recently, hybrids of MXenetransition metal sulfides have gained growing attention. [26] Examples include Ti 3 C 2 MoS 2 hybrids fabricated either by chemical synthesis of MoS 2 in the presence of Ti 3 C 2 MXene [27,28] or by physical admixing of the two. [29] MoS 2 is one of the transition metal dichalcogenides (TMDs) class of 2D materials. [30] TMDs find extensive applica tion in energy storage, [31] electrochemical desalination, [32] and electrocatalysis. [33] However, most TMDs suffer from low elec trical conductivity and unfavorable strains during charge/dis charge. [34] This translates to limited electron mobility to active sites and low stability upon cycling in electrocatalysis. Further more, a general problem of 2D structures is the reduced cata lytic activity of the basal plane. Heterostructure engineering can significantly boost the catalytic properties of TMDs by improving conductivity, increasing edge terminations, acti vating the basal plane, or creating defects. [35] Therefore, hetero structures obtained by integrating TMDs and MXenes benefit from the latter's metallic conductivity, resulting in a more effi cient charge transfer. This is shown by a sixfold enhancement of the hydrogen evolution reaction (HER) activity of Ti 3 C 2 MoS 2 hybrid compared with sole MoS 2 and a much larger enhance ment than that of sole Ti 3 C 2 . [28] Nevertheless, the enhance ment of the TMDheterostructure properties is directly related to the interface quality of the heterojunction and minimization of Fermi level pinning. [26] A high hybridization degree can be achieved by converting or templating one of the heterostructure components into the other through atomicscale transforma tion. MXene templating is still mostly limited to metal oxides. [36] The direct conversion of MXene into TMD via in situ thermal sulfidation was so far explored to produce MXene TMD heterostructures of Mo 2 TiC 2 MoS 2 , [37] Mo 2 CMoS 2 , [38] and Ti 3 C 2 TiS 2 , [39] out of their parent delaminated MXenes. These MXenederived TMDs are distributed in a unique sandwich like fashion in the matrix of parent MXene. This morphology is accessible to electrolytes allowing for fast charge transfer. The intimate contact between the two phases provides strong adhesion and a nanohybrid interface [38] that is otherwise not achieved through TMD precursors' growth on MXene via other routes exemplified earlier. A variety of (mixed metal) MXenes remain unexplored as the precursor for TMD fabrication. Fur thermore, no study has been done comparing the effect of dif ferent MXene phases (e.g., with different n, such as comparing MXene with n = 1 to that of n = 3), of the same transition metal, on the resulting derivative of any kind.
The present study thus investigates the onestep thermal sulfidation technique for fabricating two niobium MXeneTMD heterostructures, namely Nb 2 CNbS 2 and Nb 4 C 3 NbS 2 . Unlike other previous reports on MXene sulfidation, no delamination, pretreatment with sulfur, [37,38] or modification with carbon [39] were employed. The degree of the MXenetoTMD conversion is shown to depend on the sulfidation temperature and dura tion, whereby hybrids of MXeneNbS 2 or only NbS 2 structures could be conveniently obtained. The sidebyside comparison of structures derived from Nb 2 C and Nb 4 C 3 MXenes reveals that each MXene undergoes unique conversion mechanisms, resulting in NbS 2 with different nanomorphologies, such as pillars or mosaics. These heterostructures are further examined as electrocatalysts for HER.

Results and Discussion
The most common method of performing thermal sulfidation of different materials is by i) mixing the precursor with sulfur powder/sulfur salt, [40] which will likely react at higher temper atures or ii) placing sulfur powder upstream of the furnace, which in the form of vapor, upon heating, is carried to the main part of the furnace, where the precursor is located. H 2 gas is often combined to form H 2 S and facilitate the transfer. [41] These methods are limited by the amount of sulfur initially placed in the system and its evaporation rate. Above about 450 °C, sulfur quickly reacts with H 2 , leading to a large amount of released H 2 S in the first few minutes of the treatment, which quickly decreases due to sulfur depletion, leading to a nonuniform feed rate of H 2 S during treatment. [42] Direct use of H 2 S gas, on the other hand, allows a constant flow and sulfursource feed rate, providing more control over the sulfidation process, especially for longer treatment times (Figures S1 and S2, Supporting Information). Nb 4 C 3 was treated by two different methods to verify this effect: directly using H 2 S gas or reacting sulfur powder with H 2 upstream of the tube furnace. Figure 1A and Figure S3, Supporting Information, show the Xray diffracto grams of Nb 4 C 3 thermally treated with H 2 S and H 2 /S methods, respectively.
At 600 °C, no significant sulfidation is observed in either system, with only a broadening of MXene peaks. At 700 °C, a clear difference appears between the two methods. The sample treated with H 2 S still shows the main features of MXene even at prolonged times (Nb 4 C 3 H 2 S70060′) with a pronounced broadening of the peaks. The (002) peak from MXene also shifts toward higher 2θ degree (Table 1), from 5.93° 2θ (Nb 4 C 3 ) to 7.10° 2θ (Nb 4 C 3 H 2 S60010′), and 7.18° 2θ (Nb 4 C 3 H 2 S70010′), indicating a decrease in the interlayer spacing between the MXene sheets. This change was already detected even when sulfidation did not occur ( Figure S3, Supporting Information) at 500 and 600 °C, which can be explained by the removal of interlayer water and other etching products. [43] In the H 2 /S setup, the sample clearly shows conversion to NbS 2 matching the trigonal structure (PDF 89-3041), with no remaining peaks of the parent MXene when treated at 700 °C. The sharp differ ences between the two systems can be attributed to the amount of H 2 S (and ultimately S) provided by each setup in the same time range (as qualitatively represented in Figure S2B, Sup porting Information). This becomes obvious when the tem perature increases to 900 °C, where the sample in H 2 /S shows mainly NbS 2 features within 10 min, while in H 2 S, evidence of remaining Nb 4 C 3 is seen even after 60 min treatment.
The faster conversion rate in the H 2 /S system can also be verified by the sulfur content obtained by energydispersive Xray (EDX) analysis (Table S1, Supporting Information). The chemical composition obtained by EDX analysis corroborates the Xray diffraction (XRD) data. The sulfur contents obtained by EDX analysis are crossexamined with CHNS elemental analysis, and the values are in close agreement with each other. No sulfur is detected below 600 °C while it reaches 31 at% when treated at 900 °C for 60 min. These changes are accompanied by decreased oxygen content and complete fluorine removal (probed via EDX) with increasing treatment temperature and time, confirming the (partial) removal of terminal functional groups from the MXene surface. The similar composition of Nb 4 C 3 S90010′ and Nb 4 C 3 S90060′ indicates that in this setup, complete conversion is already achieved in 10 min.
The H 2 /S setup leads to faster conversion of the material. Hence, the partial conversion of the MXene becomes chal lenging. Partial conversion is highly desirable since a combi nation of the properties of the derived material and the parent MXene can be achieved, for example, the redox activity of metal oxides with the conductivity of MXene. [44] Besides, the fast con version with the H 2 /S method leads to heterogeneous reactions and the formation of coarsened particles spread within the orig inal layered structure, as observed in scanning electron micro graphs ( Figure S4, Supporting Information). That considered, the Nb 2 C phase was only treated with the H 2 S system.
For Nb 2 C, a similar change in composition occurs, ultimately leading to tetragonal NbS 2 at 900 °C in 60 min ( Figure 1B). In this case, the transition seems to occur faster than in Nb 4 C 3 , as observed by comparing the percentage of sulfide incorporated (Table S2, Supporting Information) and the A sulfide /A MXene ratio (Table 1) in the different treatment conditions between the two MXene phases. This could be explained by the slightly larger surface area of the Nb 2 CMXene than Nb 4 C 3 MXene ( Figure S5, Supporting Information), which enhances the exposure to the H 2 S, facilitating the conversion process. Similar to Nb 4 C 3 , Nb 2 C patterns also presented a shift of (002) peaks to higher angles after the heat treatments (Table 1).
Furthermore, the samples treated at 700 °C show reflec tions corresponding to Nb 2 O 5 in addition to Nb 2 C and NbS 2 . The Nb 2 O 5 formation in the case of Nb 2 C precursor can be explained by the higher atomic O/Nb ratio in Nb 2 C compared to Nb 4 C 3 (1.47 compared to 0.60, respectively; Tables S1 and S2, Supporting Information). During heat treatment, the large amount of oxygen promotes Nb oxidation in Nb 2 C, as not all functionalities are lost at 700 °C. [45] The Nb 2 O 5 reflections remain even when the treatment is extended to 60 min at 700 °C ( Figure 1B), indicating that the oxide neither decomposes nor is further converted during the treatment with H 2 S. Once Nb 2 O 5 is formed, its sulfidation is thermodynamically unfavored at 700 °C, requiring higher temperatures than Nb 2 C to form NbS 2 . [46] This can explain the absent Nb 2 O 5 signal in the Nb 2 CH 2 S90060′ sample: the higher temperature efficiently removes surface functionalities. Besides, H 2 S decomposition is favored above 900 °C, leading to a more significant amount of H 2 that helps to reduce oxygen ated groups. [42] In the case of the sample Nb 2 CH 2 S70010′, in addition to the formation of NbS 2 and Nb 2 O 5 , a pronounced reflec tion also appears at 2θ = 41.5° ( Figure 1B). Although the latter reflection could not be assigned to a certain phase, it may originate from cubic NbC and other niobium sulfide --species such as Nb 0.9 S. The rather short exposure time to H 2 S (10 min) can explain the formation of the latter sulfur deficient niobium sulfide stoichiometry. The scanning electron microscopy (SEM) images presented in Figure 2 and Figure S6, Supporting Information, show the morphology of the materials before and after different sulfida tion conditions. The original layered MXene structure is still observed for all samples, and the flake particles remain essen tially of the same size. For treatments at 900 °C, such layers are mainly composed of hexagonal and trigonal nanomosaics, characteristic of NbS 2 morphology. The gradual development of morphologies seen at 700 °C further leads to the morphologies seen at 900 °C ( Figure 2B-D). At certain particles ( Figure 2B), hexagons appear at the edge plane, with a higher surface and exposure to H 2 S gas. The same features are also seen at places where the layeropening is more pronounced. Such morpholo gies are very different from the samples treated in the H 2 /S setup ( Figure S4, Supporting Information). Large rhombohedral Figure 2. Scanning electron micrographs of A) Nb 4 C 3 , B) Nb 4 C 3 -H 2 S700-10′, C) Nb 4 C 3 -H 2 S700-60′, D) Nb 4 C 3 -H 2 S900-60′, E) Nb 2 C, F) Nb 2 C-H 2 S700-10′, G) Nb 2 C-H 2 S700-60′, and H,K) Nb 2 C-H 2 S900-60′. Transmission electron micrographs of I) Nb 2 C-H 2 S700-10′ and L) Nb 2 C-H 2 S900-60′. J) Elemental composition via energy-dispersive X-ray spectroscopy of the assigned points in image (I).
particles appear on top and grow from the layers, resulting in a much more heterogeneous material.
A strikingly different feature occurs in the Nb 2 C treated at 700 °C ( Figure 2F; Figure S6D-F, Supporting Informa tion). Several pillars growing mainly from the basal plane are observed throughout the whole material. Such pillars occur between smaller layers and grow considerably at the external surface, up to 1 µm. When the reaction time is increased from 10 min (Nb 2 CH 2 S70010′) to 60 min (Nb 2 CH 2 S70060′), the pil lars seem to increase in quantity. At 900 °C (Nb 2 CH 2 S90060′), however, there are no signs of such morphology, indicating that such structures are decomposed at higher temperatures.
To further investigate these morphologies, transmission electron microscopy (TEM) images were acquired (Figure 2I,L; Figure S7, Supporting Information). In Nb 2 CH 2 S90060′, the layers composed of prismatic NbS 2 nanomosaics are nicely seen. The lattice spacing of 0.285 nm matches the (101) plane of NbS 2 also observed in XRD. The same structures occur for the Nb 4 C 3 sample treated under the same conditions. For Nb 2 CH 2 S70010′, the pillars previously seen in SEM appear together with the layered material. The pillars seem to grow parallel to the basal plane in many regions, protruding from in between the layers. The lattice spacing of the pillars is ≈0.388, 0.314, and 0.250 nm, which matches the (001), (100), and (101) reflections of Nb 2 O 5 , respectively. The base from where the pillars grow also shows organized spacing with ≈0.62 and 0.285 nm, matching the (003) and (101) reflections of NbS 2 , respectively. While other characterizations indicate the pres ence of residual MXene in this sample, it was not possible to identify characteristic spacings of Nb 2 C in this sample. Possibly the remaining MXene occurs at the inner layers of the material, which cannot be identified in TEM as only the edges and par tially delaminated layers can be analyzed. Nevertheless, in the Nb 4 C 3 H 2 S70010′ sample, spacing corresponding to the (002) plane of Nb 4 C 3 can be more easily found ( Figure S7A, Sup porting Information), confirming the partial conversion already observed in XRD.
EDX spectra also confirmed the composition ( Figure 2J): While the layered particles are mainly composed of Nb and S, the pillars show the presence of Nb and O. Al appears at the far end/top of most of these structures, which could be part of the pillar growing mechanism acting as a catalyst. The evidence that the pillars observed in the Nb 2 C samples treated at 700 °C are composed of Nb 2 O 5 also aligns with XRD data (Figure 1B), since neither the pillars nor Nb 2 O 5 peaks are seen in Nb 2 CH 2 S90060′. To understand the decomposition mecha nism of such pillars, Nb 2 C was heattreated only in an inert atmosphere (Argon) up to 900 °C. In this condition, NbO 2 is observed ( Figure S8, Supporting Information). This indicates that the oxide pillars were reduced upon temperature increase and only further reduced/converted during the sulfidation pro cess with the introduction of H 2 S, leading to the absence of pil lars or oxide in the Nb 2 CH 2 S90060′ sample.
Raman spectra of Nb 4 C 3 and Nb 2 C treated with H 2 S are presented in Figure S9, Supporting Information. Before the treatment, NbC modes appear below 300 cm −1 , corresponding to A 1g outofplane vibration (ω 4 ≈260 cm −1 ) and E g inplane oscillations (ω 2 ≈180 cm −1 ). [47] The functional terminations (NbF, NbO, NbOH) appear at 450-800 cm −1 . [48] Compared to Nb 4 C 3 , Nb 2 C bands present a shift toward higher frequen cies as M n+1 X n O tend to be stronger with the decrease of n. [49] The difference in the NbO bond strength could be one of the factors that lead to the growth of Nb 2 O 5 pillars in Nb 2 C and their absence in Nb 4 C 3 . These modes gradually disappear with increasing temperature and time as this bonding is broken to form NbS 2 . This transition is more abrupt in Nb 2 C as previ ously discussed that this phase has a faster conversion. With the treatment, typical modes of 2HNbS 2 emerge at 386 cm −1 (A 1 ), 322 cm −1 (E 2 ), and 153 cm −1 (defects). [50] Carbon modes are also present between 1200 cm −1 and 1700 cm −1 . The band at lower wavenumber (≈1350 cm −1 ) is defined as the Dband. It is associ ated with disordered amorphous sp 3 carbon, while the Gband at ≈1600 cm −1 is characteristic of graphitic sp 2 carbon mode. [51] For Nb 4 C 3 , C modes initially increase with treatment, as free C is formed while niobium is sulfidized. [52] The broadness of the Dband is further decreased when the temperature rises to 900 °C, with a Dmode and Gmode typical of defective graphitic carbon formed at higher temperatures. Even at high temperatures, the presence of carbon modes suggests a hybrid structure between NbS 2 , carbon, and possibly residual MXene. Conversely, carbon modes disappear upon treatment of Nb 2 C. The less shielding of niobium layers and the single carbon layer possessed by the 211 MAXphase leads to less protection of the carbon atoms and fur ther decomposition.
To understand the surface chemistry of Nb 4 C 3 before and after sulfidation and the nature of Nb at the surface, XRay photoelectron spectroscopy (XPS) was studied for both Nb 4 C 3 (Figure 3A,C; Figure S10A,C, Supporting information) and Nb 4 C 3 S90060′ ( Figure 3B,D; Figure S10B,D, Supporting information). The data of both samples were calibrated using adventitious carbon (CC/CH at 284.8 eV). For the Nb 4 C 3 sample, Figure 3A shows the C 1s spectrum where the peak at 282.7 eV can be assigned to CNb in MXene. [53] Figure 3C shows the Nb 3d spectrum of Nb 4 C 3 where the doublet peaks at 203.8 and 206.6 eV are assigned to NbC in Nb 4 C 3 MXene. The doublets found at higher binding energies are assigned to NbO of different oxidation states. [53] For Nb 4 C 3 S90060′ sample, Figure 3B shows the C 1s spectrum where the CNb was not observed, which could be attributed to the NbS 2 nanomosaic layer formation upon sulfidation. For the Nb 3d spectrum of Nb 4 C 3 S90060′ in Figure 3D, it is impossible to differentiate between Nb 4 C 3 and NbS 2 as the NbS and NbC doublet peaks binding energies are too close to be distin guished from each other. Previous literature of NbS doublets were found at (203.4 ± 0.2 eV and 206.1 ± 0.2 eV) for 3RNbS 2 and (204.0 ± 0.2 eV and 206.7 ± 0.2 eV) for 2H (or 1H) NbS 2 . [54] Since the CNb peak was not observed in the C 1s spectrum of the Nb 4 C 3 S90060′ surface, it is reasonable to assume that the Nb 3d spectrum does not include the NbC doublet. There fore, we were able to assign (203.4 and 206.1 eV) to Nb +4 S in 3RNbS 2 and (203.8 and 206.5 eV) to Nb +4 S in 2HNbS 2 . [54] We also studied the C 1s and Nb 3d spectra ( Figure S10A,B, Supporting information) for both samples after sputtering to study the subsurface chemistry. The CNb peak in C 1s spectrum relatively increased in Nb 4 C 3 ( Figure S10A, Supporting Infor mation), but more importantly, it showed up in Nb 4 C 3 S90060′ ( Figure S10B, Supporting Information). This reveals the existence of Nb 4 C 3 under the nanomosaic NbS 2 layer. The presence of NbC signal in Nb 4 C 3 S90060′ sample after sput tering demonstrates that the conversion of the MXene into the sulfide takes place from outer to inner layers. Furthermore, this indicates that the samples treated at milder conversion condi tions (700 °C) are indeed partially converted and composed of MXeneNbS 2 hybrid. The Nb 3d spectra of both samples after sputtering ( Figure S10C,D, Supporting Information) reveal no Nb 2 O 5 peak at high binding energy.
To investigate the electrocatalytic behaviors of the asprepared samples, linear sweep voltammetry (LSV) was carried out in an H 2 bubbled 0.5 m H 2 SO 4 aqueous electrolyte. As shown in Figure 4A, Nb 4 C 3 pristine, Nb 4 C 3 H 2 S60010′, Nb 4 C 3 H 2 S70010′, and Nb 4 C 3 H 2 S90060′ exhibit similar electrocatalytic activities with almost the same onsetoverpotentials. It can be seen that all samples show similar overpotentials of around 200 mV at 10 mA cm −2 . A similar trend can be found for Nb 2 CH 2 S70010′ and Nb 2 CH 2 S90060′, in which the respective overpotentials of 250 and 252 mV are needed to deliver a current density of 10 mA cm −2 . Given the lower overpotential for Nb 4 C 3 and Nb 4 C 3 derived sam ples compared to Nb 2 C and its derived samples, we focus herein on the electrocatalytic behavior of the formers. Generally, the following three possible pathways can be used to describe the HER mechanism in the acidic electrolyte (Equations (1)-(3)): [55] + → + −

Volmer step: H e H ads
(1) where H ads stands for the adsorped H on the catalyst surface. There fore, two possible mechanisms can be predicted, viz. Volmer-Heyrovsky or Volmer-Tafel. The Tafel slopes of Nb 4 C 3 pristine, Nb 4 C 3 H 2 S60010′, Nb 4 C 3 H 2 S70010′, and Nb 4 C 3 H 2 S90060′ are calculated to be 166, 143, 129, and 122 mV dec −1 , respectively ( Figure 4B). The lower Tafel slope of Nb 4 C 3 H 2 S90060′ indicates faster electron transfer than the others. The large Tafel slopes suggest that a Volmer-Heyrovsky step generated the hydrogen, and H (Volmer step) adsorption is the limiting step. [55] To further study the HER activity, each sample's electrochem ically active surface area (EASA) was calculated by performing cyclic voltammetry (CV) at different scan rates ( Figure S11, Supporting Information). In this work, a nonFaradaic poten tial region with the center of open circuit potential is used. The opencircuit potential varies from one sample to the other, so a different potential range is employed in this work to avoid any Faradic contribution from any of the samples for accurate estimation of the EASA. The EASA differs from the Brunauer-Emmett-Teller (BET)specific surface area as the first deter mines the electrochemically available surface while the latter probes the area accessible for nitrogen physisorption.
The EASA values for Nb 4 C 3 pristine, Nb 4 C 3 H 2 S60010′, Nb 4 C 3 H 2 S70010′, and Nb 4 C 3 H 2 S90060′ were obtained as 35.1, 5.1, 2.6, and 0.7 cm 2 , respectively. The EASA decreases significantly after treatments, and the higher temperatures result in smaller EASA values. The very high EASA of Nb 4 C 3 is attributed to the large dspacing that renders the interlayer spacing accessible for ions and available for electrochemical processes. The reduction in EASA upon treatment can be explained by collapsing MXene's interlayer spacing during the heating cycle. Unlike normalizing the current by geometric area of the electrode, normalizing by the specific surface area or EASA reflects the intrinsic catalytic activity of the material rather than size dependence. [56] When LSV curves are nor malized to EASA ( Figure 4C), the Nb 4 C 3 H 2 S90060′ sample shows the highest HER activity. This suggests that even though this sample has lower EASA, it is intrinsically much more catalytically active than untreated MXene. Therefore, further enhancement in the kinetics of the HER is expected for Nb 4 C 3 H 2 S90060′ in the future by approaches to increase its EASA, such as milling. Table S3, Supporting Information, lists the electrocatalytic performance of various NbS 2 morphologies employed as HER catalysts in recent literature. The Nb 4 C 3 H 2 S90060′ in the present work shows a relatively low overpotential (198 mV at 10 mA cm −2 ) and Tafel slope (122 mV dec −1 ), revealing higher electrocatalytic activity than most of the reported NbS 2 and MXenes as HER catalysts. The improved performance of the latter sample can be attributed to the nano mosaic mor phology achieved where more edges are exposed. The long term electrocatalytic stability was conducted in a 0.5 m H 2 SO 4 electrolyte with a current density of 10 mA cm −2 for 24 h. As shown in Figure 4D, Nb 4 C 3 H 2 S90060′ exhibits promising electrochemical stability.

Conclusions
In summary, heterostructures of MXeneTMD were obtained by a onestep thermal sulfidation of niobium MXenes. Our work for the first time demonstrated that by a facile manipulation of the treat ment temperature and duration, different degrees of MXeneto TMD conversion are obtained. Our results showed that the parent MXene slab thickness plays a crucial role in the derived TMD nanostructure. Nb 2 C MXene is more susceptible to conversion into NbS 2 than Nb 4 C 3 MXene when treated at 700 °C for 10 min due to higher H 2 S exposure of NbCNb layers in Nb 2 C rather than NbCNbCNbCNb layers in Nb 4 C 3 . Tuning treatment con ditions yielded various nanostructures depending on the MXene type, such as pillars in Nb 2 C and mosaics in the case of Nb 4 C 3 MXene. The treatment at 700 °C resulted in MXeneNbS 2 hybrids, whereas treatment at 900 °C for 60 min resulted in the conversion of MXene into NbS 2 nanomosaic heterostructures with C/MXene core. Overall, the H 2 S gas method has better control over the con version degree than the H 2 gas + S powder method.
When used as electrodes for HER, the Nb 4 C 3 MXene treated in H 2 S gas at 900 °C for 60 min exhibited the highest HER activity among all samples tested, and promising cycling sta bility. Our study encourages the possibility of constructing a plethora of MXenebased TMD heterostructures for use as elec trode materials for energy storage and conversion, among other applications. Further finetuning the derivatization parameters to optimize the electrochemical performance of the produced heterostructures in future work may provide the path toward applications in more complex electrocatalytic applications and improved performances.

Experimental Section
Synthesis of Nb 2 AlC and Nb 4 AlC 3 MAX Phases: Nb 2 C and Nb 4 C 3 were synthesized by selective etching of the Al from their parent MAX phases, Nb 2 AlC and Nb 4 AlC 3 , respectively. Nb 2 AlC and Nb 4 AlC 3 were prepared via solid-state synthesis processes by mixing niobium (Nb, Alfa Aesar, < 45 µm, 99.8%), aluminum (Al, Alfa Aesar, 7-15 µm, 99.5%), and carbon (C, Alfa Aesar, 7-11 µm, 99%) with atomic ratios of 2:1.3:0.95 and 4:1.5:2.7, respectively. Each mixture was added to yttria-stabilized zirconia balls in a high-density polyethylene jar and was mixed using a Turbula mixer for 3 h at 56 rpm. The samples were then pressed at room temperature into pellets. Nb 2 AlC pellets were heated from room temperature to 1600 °C. The samples were held at the peak temperature for 4 h, then left to cool naturally to room temperature. Nb 4 AlC 3 pellets were heated through a three-temperature-step program: first at 750 °C for 30 min, second at 1450 °C for 30 min, and third at 1700 °C for 1 h. The heating rate used for both samples was 300 °C h −1 . All the heating and cooling steps were in an alumina tube furnace under a continuous flow of argon (Ar) at a flow rate of 100 sccm.
Synthesis of Nb 2 C and Nb 4 C 3 MXenes: Both Nb 2 AlC and Nb 4 AlC 3 were etched using hydrofluoric acid (48-51% by mass HF, Acros Organics, 10 mL for each 1 g of MAX phase powder). The mixtures were heated to 40 °C in an oil bath, mechanically agitated using Teflon coated magnetic stirrer, and held for 90 h. Afterward, the products were washed with de-ionized (DI) water to remove the acid and etching side-products. The washing procedure starts with dividing the mixture into several centrifuging tubes (one tube for each 0.5 g of starting MAX phase), centrifuging at 3500 rpm for 5 min, discarding the supernatant, refilling the tube with DI water, and finally redispersing the sediment using vortex machine. This washing step was repeated several times until the pH level exceeded a value of 6. Finally, the as-prepared product was dried using vacuum-assisted filtration overnight.
MXene Sulfidation: Nb 4 C 3 was placed in a quartz crucible inside a quartz tube positioned in a one-zone furnace. Two different methods were used to perform the sulfidation, named H 2 /S and H 2 S ( Figure S1, Supporting Information).
Method H 2 /S: To produce H 2 S gas, a second crucible containing sulfur (S) powder was placed upstream in the tube outside the furnace, with a heating jacket fixed around it. The system was purged with Ar at 100 sccm for 2 h, heated at a rate of 300 °C h −1 to the target temperature (500, 600, 700, or 900 °C) with holding times of 10 or 60 min. The Ar flow was kept at 50 sccm during heating, holding, and cooling steps. 20 min before reaching the holding temperature, H 2 gas flow was started at 10 sccm and turned off after holding time. Throughout the H 2 flow period, the heating jacket was kept at 550 °C ( Figure S2A, Supporting Information).
Method H 2 S: Like method H 2 /S, initial purging, Ar flow, heating rates, and temperatures were kept the same. H 2 S gas was directly used instead of the combination of H 2 + S powder. H 2 S gas was initiated when the target temperature was reached and kept flowing during holding time at 50 sccm. The samples were named according to the target temperature and holding time, "S" for method H 2 /S, and "H 2 S" for method H 2 S. For example, Nb 4 C 3 -S700-10′ stands for Nb 4 C 3 sulfidized at 700 °C for 10 min in the H 2 /S system. Nb 4 C 3 -H 2 S900-60′ stands for Nb 4 C 3 sulfidized at 900 °C for 60 min with H 2 S gas. For the sulfidation of Nb 2 C, only method H 2 S was applied, and the nomenclature follows the same pattern.
Material Characterization: SEM was carried out using a ZEISS-Gemini SEM500 system coupled to an EDX detector (Oxford Instruments for EDX analysis). Acceleration voltages of 1-3 kV were used for imaging and 15 kV for spectroscopy. The samples were analyzed without any conductive sputtering.
TEM (JEOL 2100F) was performed at an acceleration voltage of 200 kV. The samples were dispersed in ethanol by tip sonication for 30 s, drop-cast onto a copper grid coated with a lacey carbon film, and dried at room temperature overnight.
XRD was conducted with a D8 Advance diffractometer (Bruker AXS) with a copper source (Cu-Kα, 40 kV, 40 mA). Signal was collected using a 1D Lynxeye detector at 0.02 2θ step with a rate of 1 s/step. XPS measurements were performed using the Thermo-Fisher K-Alpha Plus XPS. An air-free holder is used during the whole process to protect the samples from oxidation. The photon source was a monochromatized Al K α line (hν = 1486.6 eV). The spectra were acquired using a spot size of 400 µm. A flood gun with combined electrons and low-energy Ar ions was used during the measurements. A dual monoatomic and gas cluster Argon ion source was used for depth profiling and sample cleaning.
Raman spectra (Renishaw inVia Raman microscope) were acquired with an Nd-YAG laser (532 nm) at 0.5 mW power with 10 s acquisition time for 10 accumulations using an objective lens with a numeric aperture of 0.75. At least 12 points were collected for each sample. The data show a representative spectrum of each sample.
Nitrogen gas sorption measurements were performed with an Autosorb iQ system (Quantachrome, Anton Paar) at −196 °C. The Nb 2 C and Nb 4 C 3 samples were first degassed at 200 °C at 10 2 Pa for one day. The specific surface areas were calculated by applying the BET model [57] within the linear pressure range (up to 0.3 relative pressure). [58] Elemental analysis (CHNS) was performed with Vario Micro Cube System using sulfanilamide for calibration and a reduction temperature of 850 °C. An OXY cube oxygen analyzer at 1450 °C was employed for the quantitative analysis of elemental oxygen.
Electrochemical Measurements: The electrochemical performance was conducted by a Bio-Logic SP200 portable electrochemical workstation with an RRDE-3A rotating ring disk electrode system. A standard SVC-2 three-electrode system with a working electrode of as-prepared samples, a Hg/Hg 2 SO 4 electrode with saturated KCl supporting solution as a reference electrode, and a graphite rod as a counter electrode were used to investigate the hydrogen evolution activity in 0.5 m H 2 SO 4 aqueous electrolyte. The working electrode was prepared by dispersing the as-prepared materials (10 mg) in a mixture of 750 µL water, 250 µL ethanol, and 10 µL Nafion 117 solution (≈5% in a mixture of lower aliphatic alcohols and water, Sigma Aldrich), followed by 1 h sonication. After that, 7 µL dispersion was drop-cast on the glassy carbon electrode and dried at room temperature. The mass loading of active materials was ≈1 mg cm −2 . LSV was measured at 5 mV s −1 by rotating the working electrode at 1600 rpm with flowing H 2 in the electrolyte. All LSV curves were iR-corrected, and the potentials were converted in reference to a reversible hydrogen electrode.
The CV polarization curves were used to calculate EASA. EASA values were obtained from the hydrogen adsorption/desorption region, so only the double-layer charging current is used. In this work, a non-Faradaic potential region with the center of open circuit potential was used. EASA was estimated from the electrical double-layer capacitance (C dl ) of the as-prepared materials. The C dl was analyzed via CV at scan rates of 20-100 mV s −1 .
The EASA was calculated using Equation (4): [59] where A is the geometric area of the electrode, and C s is the capacitance from a smooth planar surface per unit area. An average value of C s = 40 µF cm −2 is used in this work. [59] Supporting Information Supporting Information is available from the Wiley Online Library or from the author.