High Crystal Quality 2D Manganese Phosphorus Trichalcogenide Nanosheets and their Photocatalytic Activity

Transition metal phosphorus trichalcogenides (MPX3, X = S, Se) are layered materials possessing high chemical diversity and wide range of applications in a broad wave length spectrum. Theoretical studies reveal that auspicious activity of photocatalytic water splitting can be realized from them. However, experimental efforts have so far been challenged with the synthesis bottleneck. Described herein is the general chemical vapor deposition (CVD) growth method and photocatalytic activity of these materials. A novel route to systematically grow MnPX3 nanosheets on flexible carbon fiber substrate is reported. The temperature profile of the CVD process is carefully optimized that confer a facile and successful conversion of oxide precursor to phospho‐trichalcogenide with high crystallinity. Moreover, the obtained manganese‐based phosphorus trichalcogenide nanosheets demonstrate promising activity in sacrificial agent‐free photocatalytic water splitting under simulated solar light (AM 1.5G). This study provides a significant stepping stone in exploring the fascinating world of functional 2D materials and pursuing performance enhancement.

The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/adfm.201800548. effective charge/electron transport, appreciable stability, and immensely exposed active surface sites, 2D layered transition metal dichalcogenides (MX 2 ) have generated a tremendous amount of interest in catalyzing the dissociation of water into its component parts. [2] Fascinating strategies have been communicated so far aiming at their controllable synthesis and performance enhancement. [3] In catalyzing hydrogen evolution reaction (HER), the crucial role of foreign P site in MX 2 is greatly appreciated as evidenced in MoS 2 , [4] WS 2 [5] and CoS 2 . [6] The extraordinarily high activities exhibited by CoPS, [7] CoS/P, [8] MoP/S [9] shine more light on this postulate. Moreover, it has been pragmatic that increasing the atomic content of P in transition metal phosphide [10] such as, Ni x P y , [11] Mo x P y , [12] and Co x P y , [13] causes the increased HER activity. Therefore, the presence of a transition metal, chalcogen, phosphorus, and the 2D feature as integral part of the catalyst can bring about the realization of excellent catalysis. In this aspect, transition metal phosphorus trichalcogenides (MPX 3 ) are the best candidates. MPX 3 is one of the families of layered 2D materials whose electronic and structural features provide appealing etiquettes for catalysis. [14] This family demonstrates higher chemical diversity and structural complexity as compared to the conventional layered MX 2 thereby opening a wide range of applications in a broad wave length spectrum. [15] Despite its long history (beginning in the 19th century) [16] of main scientific direction, the development of MPX 3 for catalysis is still at its infant stage. This is associated with the difficulties in controllably synthesizing nanoscaled MPX 3 crystals. The methods employed so far such as chemical vapor transport (CVT) require quite long time of reaction up to couple of days, weeks or even months. [17] It involves a mere reaction of stoichiometric amounts of the components (metal, chalcogen, and phosphorus) in evacuated tube. The resultant MPX 3 is generally big in size and low in specific surface area to volume ratio. Therefore, the CVT method is not a practicable way to control the morphology of MPX 3 and hence thorough investigation of this family in various applications has been thwarted due to the synthesis bottleneck. Serendipitously, chemical vapor deposition (CVD) is well acknowledged route for swiftly growing various MX 2 nanostructures with enormously exposed active sites for catalysis. [18] Unlike MX 2 , the growth of MPX 3 via this method requires a careful optimization of temperature as the group-V

Introduction
The production of hydrogen molecule through water splitting has remained a significant issue of investigation in connection to the effort of sustainably powering the globe. [1] Considering the atom, P, has to be vaporized along with the chalcogen atom, X, and completely react with the metal oxide precursor. Our group has recently reported the photocatalytic hydrogen evolving activity of ultrathin NiPS 3 nanosheet which uncovers the promising feature of this family toward water splitting. [19] Herein, we further introduce a "preheating" process into the CVD growth to confirm the formation of fused mixture of P and X atoms before they are carried to the back zone for reacting with the metal oxide precursor. Being a theoretically proved promising photocatalyst with direct band gap, [20] manganese-based transition metal phosphorus trichalcogenides (MnPX 3 ) are chosen and synthesized with our developed method. This novel route resulted in a successful synthesis of monoclinic MnPS 3 and hexagonal MnPSe 3 nanosheets with high crystalline quality and phase purity. Their band gaps and the phase structure were studied in detail. Notably, their valance band (VB) and conduction band (CB) energy levels are constituted from the P 2 X 6 4− cluster orbitals. Going from P 2 S 6 4− to P 2 Se 6 4− , the electronic structure and the nature of energy levels are altered following the difference between S and Se atoms. The light absorption property shows an obvious red shift for MnPSe 3 . Interestingly, the synthesized nanosheets demonstrated unassisted photocatalysis for hydrogen generation in sacrificial agent-free water. The hydrogen evolution rate is about 6.5 µmol g −1 for MnPSe 3 nanosheets, which shows a better performance than that of MnPS 3 nanosheets (3.1 µmol g −1 ). Being the first experimental report regarding the catalytic property of facilely synthesized MnPX 3 , our finding can be an essential stepping stone in pursuit of performance enhancement and further exploration of other 2D materials in this emerging family.

Results and Discussion
We followed a systematic two step method (see the Experimental Section) which is a novel route for obtaining MnPX 3 at the level of nanometer thickness. First, MnO 2 nanosheet (NS) was grown hydrothermally on preacidified carbon fiber (CF) substrate. The morphology, as obtained from scanning electron microscope and transmission electron microscope (TEM), the X-ray diffraction (XRD) pattern and high resolution TEM image are illustrated in Figures S1 and S2 (Supporting Information). It is obvious that the nanosheet grew vertically covering the entire surface of the substrate with crystallographic pattern matching to tetragonal MnO 2 (PDF#72-1982). In the second step, the obtained MnO 2 NS/CF was placed at the back zone of CVD furnace wherein the front zone was loaded by a mixture of P and S/Se (accordingly) powders. Two growth routes were conducted (Figure 1). In route-I, we tried direct heating of the front zone along with the back zone simultaneously, which is commonly used in the CVD synthesis of MX 2 nanomaterials. [18] Nonetheless, prior selenization occurred inhibiting the pure phase synthesis of MnPSe 3 . According to the XRD pattern (Figure 1b), both cubic MnSe and hexagonal MnPSe 3 coexisted in the grown material. Thus, initial creation of thiophosphate/selenophosphate atmosphere (route-II, Figure 1) in the CVD system was found vividly crucial to successfully get the desired MnPX 3 phase. For this, the front zone was preheated at 110 °C for 10 min to make a fused mixture of phosphorus and chalcogen atoms that lead to a phase pure product. This phase purity and the crystalline structure of the as-grown materials were ascertained from XRD analysis. It is depicted in the XRD pattern of route-II that the peaks for MnPSe 3 matches well to standard diffractogram of hexagonal crystalline structure (PDF# 89-1414) and that of MnPS 3 ( Figure S3, Supporting Information) appears in monoclinic arrangement (PDF#78-0495). A typical metal phospho-trichalcogenide crystal can be regarded as a composition of two cations (M 2+ ) and one [P 2 X 6 ] 4− anion cluster in which 2/3 of the octahedral centers is occupied by M 2+ . The P-P dimers that are covalently bonded to chalcogen with a tetrahedral like arrangement (3X atoms with Adv. Funct. Mater. 2018, 28, 1800548 each P atom) occupy the remaining 1/3 part. They are layered structures having a weak van der Waals interaction between the layers and a P 2 X 6 anion unit within each lamella. In MnPSe 3 , the atomic layers stack in R3 space group such that each Se atom is bonded to single P atom and coordinated by two Mn sites in hexagonal arrangement ( Figure S3b, Supporting Information). Coming to MnPS 3 , structural difference arises owing to the decreased bond distance (PS) and bond angles (SPS) as compared to its selenide equivalent. The atomic layers stack in the C2/m space group of monoclinic structure ( Figure S3b, Supporting Information). As can be seen from the images of scanning electron microscopy ( Figure S4, Supporting Information), their morphology resembles hexagonal nanosheets grown on CF. The nanosheets cover the surface of the CFs uniformly. For a clearer visualization using atomic force microscopy (AFM), we transferred the sample from CF to SiO 2 /Si substrate and the images are shown in Figure 2a,b. It is evident that the transferred nanosheets are as thin as ≈6 nm and ≈28 nm with lateral size of 1.5 and 0.45 µm for MnPS 3 and MnPSe 3 , respectively. Moreover, energy dispersive X-ray spectroscopic (EDX)-elemental analysis confirms homogenous distribution ( Figure 2c, Figure S5, Supporting Information) of the components with atomic ratio as expected in MnPX 3 ( Figure S6, Supporting Information). Further crystallographic information can be collected from high resolution transmission electron microscope (HRTEM) and selected area electron diffraction (SAED) patterns. Figure 2d shows a clear crystal fringes with lattice spacing of 0.3 and 0.51 nm which are indexed to the crystal planes of (−201) and (110), respectively. Complementary information can be elucidated from SAED pattern depicted in Figure 2e where well-defined diffraction spots are clearly visualized corroborating the presence of characteristic planes for monoclinic MnPS 3 (PDF#78-0495). Figure 2f,g depicts the corresponding results for MnPSe 3 . From the HRTEM image, it is apparent that the lattice spacing of 0.371, 0.208, and 0.288 nm can be related to the crystal planes of (104), (211), and (113), respectively. Moreover, the SAED result unequivocally reveals a pure hexagonal MnPSe 3 with crystal planes consistent with XRD pattern (PDF# 89-1414). Beside structural identification, we carried out characterization with Raman spectroscopy to correlate the finger print peaks due to polarization driven modes of vibrations (Figure 3a,b). Being archetypal to most members of layered MPX 3 , [21] a band bellow 150 cm −1 is attributed to cation vibrations and a small peak representing E u mode of vibration is apparent in both MnPS 3 and MnPSe 3 at around 155 cm −1 . [22] As can be seen in Figure 3a, the Raman modes originating from P 2 S 6 units with D 3d symmetry (center of inversion along PS 3 groups) are identified as A 1g (245.9, 381.8, and 491 cm −1 ) and E g (225.5 and 581 cm −1 ). [23] The out-of plane vibrations of P 2 S 6 entities (with P-P dimers) that are characterized by the opposite movement of S 3 P-PS 3 units can be ascertained from the strongly polarized A 1g mode appearing at 381.8 cm −1 . [24] Prominent modes due to the PS bond symmetric stretching vibration are also corroborated from the peaks of Raman spectrum at 245.9 and 491 cm −1 . Moreover, those peaks appearing at 274 and 581 cm −1 are in response to the in-plane vibrations of P 2 S 6 unit and represented by E g mode. In the case of MnPSe 3 , belonging to hexagonal system, it possesses noncentrosymmetric PSe 3 units and hence do not exhibit center of inversion. Another compelling remark can be made via comparison of the  two Raman spectra of MnPS 3 versus MnPSe 3 . The strongest peak at about 381 cm −1 in MnPS 3 shifts to about 216 cm −1 in MnPSe 3 . As mentioned above, this peak is meant for the stretching vibration mode of PX 3 units. The observed shift, therefore, tangibly corroborates the fact that chalcogen atoms are mainly involved in this motion no matter how the contribution from vibrational coupling be excluded. Analogous observation was made for FePS 3 versus FePSe 3 . [22] Insight into the surface chemical states of the synthesized MnPX 3 was gained from X-ray photoelectron spectroscopy (XPS). The Mn 2+ state can be ascertained from the deconvoluted peaks of Mn P2p 1/2 and Mn P2p 3/2 at binding energies of 652 and 640 eV along with shake-up satellite peak at 645 eV ( Figure S7, Supporting Information). Figure 3c-f shows the XPS spectra of P, S, and Se core levels from which the expected peaks for each elements as in the formation of MnPX 3 are verified. The S 2p XPS spectrum shows well resolved peaks for 2 p 1/2 and 2 p 3/2 at binding energies of 161.9 and 163.1 eV, respectively, suggestive of Mn-S and P-S entities. [23] The Se 3d spectrum (Figure 3f) depicts deconvoluted peaks at 54.22 and 55.14 eV corresponding to Se 3d 3/2 and Se 3d 5/2 , respectively, which are in good agreement with the reported values for Se 3d core level characteristics in MnPSe 3 . [23] In particular, it is conceivable to identify if chemically nonequivalent atoms of the same element are present in the investigated samples. It is tempting to describe that that there exists an obvious shift in  the binding energy of P as seen in MnPS 3 and MnPSe 3 . This might be associated to the slightly higher electronegativity of S as compared to Se. Given the fact that the core level binding energy of the central atom increases as the electronegativity of the atoms attached to it increases, P 2p in MnPS 3 appears at 131.7 and 132.6 for P 2p 1/2 and P 2p 3/2 , respectively, which are relatively at higher binding energies (by ≈0.3 to 0.4 eV) than the case in MnPSe 3 . This corroborates the presence of nonequivalent atoms of the same element, P, in MnPS 3 and MnPSe 3 . Moreover, the additional peak at 138.7 eV in Figure 3e can be attributed to the overlap of a Se Auger emission line and the P 2p core level, which is in agreement with the observation made elsewhere. [24] In order to know the light harvesting capacities of the assynthesized samples, we investigated their optical properties by UV-vis diffuse reflectance spectroscopy from which the absorption spectra are extracted and shown in Figure S8 (Supporting Information). It is obvious that the absorption edge of MnPSe 3 shows a remarkable red shift signifying its enhanced light scavenging property as compared to MnPS 3 . Tauc plots shown in Figure 4a corroborate the fact that the synthesized materials are typical semiconductors with band gap energies (E g ) of 2.00 and 2.39 eV for MnPSe 3 and MnPS 3 , respectively. In order to determine the specific energy levels, we employed ultraviolet photoelectron spectroscopy (UPS) to get the absolute position of VB. By subtracting the width of He I UPS spectra from the system excitation energy of 21.22 eV (Figure 4b), one can see that the VB levels of MnPS 3 and MnPSe 3 are at 1.15 versus NHE (−5.59 eV vs Vacuum) and 0.83 versus NHE (−5.27 eV vs Vacuum), respectively. Having these at hand, the energy level of CB can straightforwardly be calculated using the E g value as 1.24 versus NHE (−3.2 eV vs Vacuum) and 1.17 versus NHE (−3.27 eV vs Vacuum) for MnPS 3 and MnPSe 3 , respectively. Furthermore, we made a Mott-Schottky (M-S) analysis via measuring the space-charge layer capacitance over a scanned range of applied potentials (0.0 to 2.1 V vs saturated calomel electrode (SCE)) at 250 and 500 Hz (see the Experimental Section). Apart from being complementary to the above energy level calculations, M-S plot can give information about the type of semiconductors under investigation. The negative slopes in Figure 4c indicate that both MnPS 3 and MnPSe 3 are p-type semiconductors, in a good agreement with other reports. [14] Meanwhile, the flat band potentials of 0.86 V versus reversible hydrogen electrode (RHE) (for MnPS 3 ) and 0.55 V versus RHE (for MnPSe 3 ) can be taken into account for estimating the energy level of VB; as VB is 0.3 V below the Fermi level for p-type semiconductors. [25] Accordingly, one can see a very close proximity with the results obtained from UPS analysis.
Adv. Funct. Mater. 2018, 28, 1800548 Considering the energetically appropriate levels of CB for photocatalyzing the conversion of H + to H 2 gas, we carried out sun light driven catalysis of water splitting using our obtained MnPX 3 nanosheets. The photocatalytic performances were evaluated in pure water, without cocatalyst or sacrificial agents, under illumination of simulated solar light (AM 1.5G). This experimental condition was followed to unequivocally explore the intrinsic property of the hitherto unexplored MnPX 3 toward photocatalyzing water splitting. It is apparent from Figure 5 that both MnPSe 3 and MnPS 3 exhibit promising photocatalytic activities toward H 2 gas production with a rate of 6.5 and 3.1 µmol h −1 , respectively. Moreover, the samples show appreciable stability in photocatalysis rendering a minor degradation after three cycles (6 h per each cycle). The rate for MnPSe 3 (MnPS 3 ) can reach to 43.5 µmol h −1 (21.2 µmol h −1 ) in the presence of Na 2 S/Na 2 SO 3 as hole scavenger ( Figure S9, Supporting Information). The relatively better performance of MnPSe 3 may be attributed to the following points: (1) The enhanced absorption in visible region that would account for generation of tremendous electrons at its conduction band, (2) The high carrier mobility, as it has been proved recently, [20] which eases the accessibility of electrons to the incoming H + and (3) given the fact that the constituents of the conduction bands are P 2 X 6 4− cluster orbitals, the slight electronegativity difference between S and Se may endow an optimized condition for the H + adsorption and H 2 desorption. However, we could not detect oxygen molecule in our system. This might be associated with the lack of sufficient active sites for promoting the kinetically sluggish OO covalent bond formation in oxygen evolution reaction (OER). [26] Additionally, the energy level of CB is just slightly below the O 2 /H 2 O level that defies the evolution of oxygen and hence limited to only accumulating radical intermediates on the surface. As a matter of concept, we conducted a spin trapping experiment to trace surface hydroxyl radicals by using electron spin resonance spectroscopy (ESR). ESR spectra in Figure 5b show intensity ratio of 1:2:2:1 for both MnPS 3 and MnPSe 3 photocatalysts under the illumination of Hg light, illustrating characteristic features of hydroxyl radicals. [27] The finding corroborates the fact that the photogenerated holes have oxidized the adsorbed water molecules to hydroxyl radicals on the surface of the catalysts. Thus, the rout toward OO covalent bond formation in OER is not completed under our experimental condition. This can be hastened by designing suitable cocatalyst that would assist the coupling of hydroxyl radicals and formation of superoxides to eventually result in the evolution of oxygen gas.

Conclusion
In conclusion, we report here a rationally designed novel route of synthesizing manganese-based transition metal phospho-trichalcogenides. Having optimized the CVD growth conditions, nanosheets of MnPS 3 and MnPSe 3 are successfully grown on flexible CF substrate. Following the theoretically speculated photocatalytic activity of direct band gap MnPX 3 nanosheets, sunlight-driven catalytic activity is measured and promising performances are recorded. This work can ease the subsequent experimental efforts in this emerging family of 2D materials in the quest for the breakthroughs of water splitting catalysis and other optoelectronic devices.

Experimental Section
Synthesis of MnO 2 Nanosheets on Carbon Fiber: Hydrothermal method was followed to synthesize MnO 2 nanosheets with vertically standing architecture. A piece of CF was ultrasonically cleaned with acetone, ethanol, and ultrapure water before use. The cleaned CF substrate was acidified with 0.1 m HNO 3 to functionalize the surface for dense growth of MnO 2 nanosheets. The pretreated substrate was then rinsed in ethanol and water before being placed in Teflon-lined autoclave containing aqueous solution of KMnO 4 (0.1 g in 50 mL of distilled water). Afterward, the content was sealed in iron cover and kept at 80 °C for 12 h. Eventually, the MnO 2 /CF was taken out, rinsed with ethanol and ultrapure water, and dried at 60 °C before it was used as precursor in the subsequent CVD growth.
Synthesis of MnPX 3 Nanosheets: MnPX 3 (X = S or Se) nanosheets were facilely grown in a two zone chemical vapor deposition technique. The front thermal zone was loaded by a weighed amount of mixture of P and S (for MnPS 3 ) or Se (for MnPSe 3 ) powders whereas the hydrothermally grown MnO 2 NS/CF was placed at the back zone. The system was flushed with Ar gas three times and pumped into a vacuum lower than 1 Pa in order to create oxygen free environment. Before starting the growth, the front zone was heated to 110 °C for 10 min to create a thiophosphate/ selenophosphate atmosphere that would take the P and chalcogen vapor together during the reaction to the deposition zone. The temperature of the front zone was set to 280 °C (MnPS 3 ) or 300 °C (MnPSe 3 ) and that of the back zone was raised to 450 °C (MnPS 3 ) or 500 °C (MnPSe 3 ). The reaction was run for 2 h with 100 sccm Ar gas flow, followed by naturally cooling of the furnace to room temperature. The obtained MnPX 3 NS/CF was used as a photocatalyst without any pretreatment.
Characterizations: The morphology and crystal structure of the sample were investigated through AFM, field emission scanning electron microscopy (Hitachi S4800), and transmission electron microscopy (FEI Tecnai F20 TEM). The obtained sample, MnPX 3 NS/CF, was transferred to SiO 2 /Si substrate and ultrathin Cu grid for being visualized by AFM and TEM, respectively. XRD (D/MAX-TTRIII (CBO) diffractometer) was employed to acquire X-ray diffractograms using Cu Kα radiation. Elemental mapping images were obtained from STEM-EDX equipped on TEM-2100F (200 kV). Information about Raman spectra was gathered from confocal microscope-based Raman spectrometer (Renishaw InVia) in ambient air environment with an excitation laser line of 532 nm. XPS spectra were collected from ESCALAB 250 Xi (system of Thermo Scientific) with analysis chamber of 1.5 × 10 −9 mbar and the X-ray spot of 500 µm. The base pressure in the analysis chamber was about 3 × 10 −9 mbar. Room temperature UV-Vis diffuse reflectance spectra were recorded using a Lambda 750 spectrophotometer in the wavelength range of 400-1000 nm. Ultraviolet photoelectron spectroscopy (Thermo Scientific ESCALab 250Xi) was used to estimate the valance band energy of MnPX3 hexagonal nanosheets. The gas discharge lamp was employed with the admitted helium gas (HeI (21.22 eV) emission line). During analysis, the helium pressure inside the chamber was about 2E−8 mbar. The data were collected with −10 V bias.
Electrochemical Test: The flat band potential, as an estimation of band edge position, was measured via a typical three electrode system in 0.1 m Na 2 SO 4 aqueous solution (pH = 6.57). SCE, Pt wire, and the as-grown MnPX 3 /CF (≈1 cm 2 ) were used as reference, counter, and working electrodes, respectively. The required data for Mott-Schottky (M-S) analysis were collected via measuring the spacecharge layer capacitance over a scanned range of applied potentials (0.0 to 2.1 V) at 250 and 500 Hz using Princeton Applied Research Potentiostat/Galvanostat VersaSTAT3 electrochemical analyzer. The recorded potentials were referenced to RHE according to E (vs RHE) = E (vs SCE) + 0.242 V + 0.059 × pH.
Photocatalytic Measurement: The photocatalytic water splitting reaction was carried out under simulated solar energy (AM 1.5G) illumination in a gas closed circulation system with a top-irradiation type reactor (quartz cell). Typically, the grown MnPX 3 /CF (mass loading ≈20 mg) was placed in a reactor containing 100 mL of ultrapure water. Before the reaction, the closed reactor system was degassed until O 2 and N 2 could not be detected. After starting irradiation, the evolving gas was quantified with a thermal conductivity detector TCD-based gas chromatograph (GC7900, Shimadzu, Japan, TCD, argon as a carrier gas and 5 Å molecular sieve column) within each one hour of the illuminations. A base-line was calibrated by using the carrier Ar gas for each test before illumination.
Hydroxyl Radical Detection: Electron spin resonance spectrometer was used to detect water oxidation intermediate during photocatalysis. Since radical oxide moieties are short-lived in nature, 5,5-Dimethyl-1-pyrroline N-oxide (DMPO) as a spin trapping agent was employed to easily detect hydroxyl radicals. In a typical experiment, the photocatalyst was immersed in 200 µL distilled water and irradiated for 10 min followed by the addition of 50 µL DMPO. Eventually, the ESR spectra were collected using Bruker E500 electron spin resonance spectrometer.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.