Porous Single-Crystalline Monolith to Enhance Catalytic Activity and Stability

Engineering the catalytic activity and stability of materials would require the identification of the structural features that can tailor active sites at surfaces. Porous single crystals combine the ordered lattice structures and disordered interconnected pores, and they would therefore provide the advantages of precise structure features to identify and engineer the active sites at surfaces. Herein, we fabricate porous single-crystalline vanadium nitride (VN) at centimeter scale and further dope Fe (Fe0.1V0.9N) and Co (Co0.1V0.9N) in lattice to engineer the active sites at surface. We demonstrate that the active surface is composed of unsaturated coordination of V-N, Fe-N, and Co-N structures which lead to the generation of high-density active sites at the porous single-crystalline monolith surface. The interconnected pores aid the pore-enhanced fluxion to facilitate species diffusion in the porous architectures. In the nonoxidative dehydrogenation of ethane to ethylene, we demonstrate the outstanding performance with ethane conversion of 36% and ethylene selectivity of 99% at 660°C. Remarkably stability as a result of their single-crystalline structure, the monoliths achieve the outstanding performance without degradation being observed even after 200 hours of a continuous operation in a monolithic reactor. This work not only demonstrates the effective structural engineering to simultaneously enhance the stability and overall performance for practically useful catalytic materials but also provide a new route for the element doping of porous single crystals at large scale for the potential application in other fields.


Introduction
Improving catalytic materials' activity and stability would require the identification of the precise structural features that can facilitate the creation and engineering of active sites at surfaces [1][2][3]. The precise structure information at the depth from the top layer to sub-nanometer layer provides the fundamental basis to identify and engineer the coordination structure of the active sites at the surface of catalytic materials. Single-crystalline materials would confine the well-defined structures at the surface layer by the longrange ordering of crystal lattice. The coordination structures of active sites would therefore be created and tailored by controlling the structures of the single-crystalline materials. The preference of single-crystalline materials is mainly focused on the dimension at nanoscale, in the form of nanocrystal, to create and engineer the precise coordination structure of active surface to enhance the catalytic activity. Single-crystalline materials at nanoscale would therefore maintain the high catalytic activity, but they suffer from the thermal instability especially at high operation temperature. Bulk single-crystalline materials at macroscale with higher thermal stability could provide the advantages of long-range ordering of crystal lattice to create and engineer the coordination structure at surface layer, but their surface areas are quite limited to host the chemical reactions. They are rarely used in catalysis and normally demonstrate negligible catalytic activity even though they present high thermal stability in contrast to nanocrystals. Therefore, singlecrystalline materials at macroscale with high porosity provide an alternative to synergistically engineer the catalytic activity and thermal stability.
Porous single-crystalline materials synergistically incorporate ordered crystal lattices and disordered interconnected pores in bulk-porous architecture at macroscale, which therefore demonstrates the combined advantages of porous materials and single-crystalline materials. The long-range ordering of crystal lattice provides the huge potential to identify and engineer the well-defined structure features to create active surface regions [4,5]. Porous single-crystalline materials at macroscale are therefore more thermally stable than single crystal at nanoscale, but the single-crystalline skeleton at nanoscale would demonstrate the comparable catalytic activity in contrast to nanocrystals. Porous single-crystalline materials have the interconnected pores in the porous architectures, which leads to the comparable advantage of porosity in contrast to porous materials. The porous architectures not only provide high surface areas to accommodate the chemical reactions but also facilitate species diffusion through three-dimensional percolation. The interaction of adsorbents with active surface in catalysis could be therefore tailored by the well-defined coordination structures at surface and the enhanced species diffusion in the porous single-crystalline architecture. Porous singlecrystalline materials at macroscale are readily machinable into complex monolithic architectures in the form of catalytic unit and reactor to well suit the current infrastructures in chemical industry.
Transition metal nitride like vanadium nitride (VN) is widely used for catalytic dehydrogenation of alkane to produce olefin. They possess unique physicochemical properties that are due to the presence of nitrogen atoms at the interstitial sites [6][7][8]. The unsaturated V-N coordination structure would produce active sites at the well-defined surfaces to tailor the activity and stability of catalytic materials. Porous single-crystalline vanadium nitride monoliths would therefore demonstrate the advantages to create and engineer the coordination structures of active sites at the surface layers, and the porous architectures would also function in the form of monolithic reactors. In chemical industry, ethylene as an important building block can be produced from the steam cracking of ethane. In shell gas, ethane contributes to~20% in volume, which provides the huge potential to convert ethane into high-value ethylene. The steam cracking process is energy-intensive with an operation temperature higher than 800°C, which inevitably encounters the adverse challenges including side reactions, severe coking, and CO 2 emissions in the process [9][10][11][12]. Nonoxidative dehydrogenation of ethane to ethylene is an effective pathway to maintain high ethylene selectivity, but the ethane conversion is limited and accompanied by undesirable coke formation [13][14][15][16].
Herein, we fabricate the porous single-crystalline vanadium nitride (VN) monolith at 1 cm scale to create the welldefined active surface and further dope Co and Fe in lattice to engineer the coordination structures of the surfaces to enhance the catalytic activity and durability of nonoxidative dehydrogenation of ethane to ethylene. In a monolithic reactor, we demonstrate the exceptionally high ethane conversion and ethylene selectivity as well as the outstanding durability even in a continuous operation of 200 hours in nonoxidative dehydrogenation process at reduced temperatures.

Results and Discussion
We fabricate porous single-crystalline (PSC) VN, Fe 0.1 V 0.9 N (FVN) and Co 0.1 V 0.9 N (CVN) monoliths at 1 cm scale by using a lattice reconstruction strategy. Parent single crystals of BiVO 4 , FeVO 4 , and CoV 2 O 6 are directly treated in ammonia atmosphere and converted into PSC VN, PSC FVN, and PSC CVN monoliths, respectively. The average roughness of the surface of parent single crystals is polished to be below 0.3 nm to sufficiently eliminate the influence of surface instability as shown in Figure S1. Figure S2-S4 present the growth mechanism of PSC monoliths, which indicates the lattice shrinkage from parent single crystals to PSC nitride monoliths leading to the creation of porosity in single crystals. In nitriding process, the removal of Bi/O in conjunction with the formation of V-N bond in the recrystallization process leads to the growth of PSC monoliths. As shown in Figure S5-S7, the energies are less than~6.4 eV for removing Bi, Co, and Fe in (001) BiVO 4 , (0-11) CoV 2 O 6 , and (100) FeVO 4 , while the energies are below~5.0 eV for the replacement of O with N in (001) BiVO 4 , (0-11) CoV 2 O 6 , and (100) FeVO 4 . The reasonable energies would facilitate the growth of PSC monoliths in the nitriding process at high temperatures. In addition, the reasonable lattice mismatch below 25% finally leads to the preferential growth of PSC (220) VN, PSC (200) FVN, and PSC (200) CVN monoliths from the (001) BiVO 4 , (0-11) CoV 2 O 6 , and (100) FeVO 4 , respectively. Figure 1(a) shows the X-ray diffraction (XRD) of PSC (220) VN monolith which confirms that the porous architecture is in a single-crystalline state. Figure 1 (b) shows the scanning electron microscopy (SEM) picture of the PSC (220) VN monolith, which indicates porous microstructure with the pore size of about 37 nm well fit the result from Brunauer-Emmett-Teller (BET) test in Figure S8. The porosity of the PSC VN monoliths grown from BiVO 4 is about 78%, which is mainly dominated by the structure and composition of the parent single crystals. Figure 1(c) shows the scanning transmission electron microscope (STEM) image of PSC (220) VN monolith, which again confirms the interconnected pores (30-40 nm) in the architecture. The inset shows the selected area electron diffraction (SAED) pattern of the porous architectures, which indicates the single-crystalline nature of PSC (220) VN monolith. Figure 1 Figure 1(e) presents ordered atom stacking indicating the singlecrystalline features in the partial enlargement of Cscorrected STEM image. Figure 1(f) shows that the V and N are evenly distributed in the skeleton as validated by the high-angle annular dark-field (HAADF) image and elemental mapping. Figure 1(g) presents the X-ray photoelectron spectroscopy (XPS) of the PSC VN monoliths which indicates that the elements V and N exist stably in the +3 and -3 of chemical states, respectively.   Figure S9. Figure 2(g) shows the HS-LEIS spectra at the surface of PSC FVN and CVN monoliths. The V and Fe are at the top layer and bonded with the beneath-layer N to form unsaturated metal-nitrogen coordination structures. Similarly, the Co and V at the top layer bond with the beneathlayer N at the surface of PSC CVN monolith. A negligible amount of adsorbed oxygen species is also observed at the surface for the PSC FVN and CVN monoliths. Figure S10 and S11 show the visualized structures of the surface of PSC FVN and CVN monoliths, which exhibits the charge density difference diagrams at the unsaturated metalnitrogen coordination structures. Compared with V, the charge density difference around Fe and Co changes into lower state, indicating the smaller electronegativity of Fe and Co in contrast to V in lattice. Figure S12 shows the density of states for VN, FVN, and CVN, which indicates the similar electronic structure even though the Fe and Co are doped into lattice. Figure S13 shows the XPS of PSC FVN and CVN monoliths which indicates the formation of Fe-N and Co-N bonds [17,18]. We further acquire the extended X-ray absorption fine structure spectroscopy (EXAFS) of PSC FVN and CVN monoliths to obtain the X-ray absorption near-edge structure spectroscopy (XANES) and Fourier transforms (FTs) of K edges (Fe and Co). We infer the Fe-N and Fe-V coordination numbers in Table S1 by fitting the EXAFS data in    Figure S16 present the coordination structures at the denoted areas. The coordination structures at regions 1, 2 and 3 are calculated to be 48% VN+52% VN 0.87 , 63% VN+29% VN 0.87 +8% VN 0.83 , and 59% VN+41% VN 0.87 , respectively. The surface around the pores is composed of unsaturated V-N coordination structures. Although the V-N coordination structure dominates the well-defined surfaces, we still observe~30-50% of the highly unsaturated coordination structure including VN 0.87 and VN 0.83 at surfaces. We further visualize the electronic states of regions 1, 2, and 3 at the surfaces in Figures 3(g)-3(i) and show that the welldefined surface is electron-deficient with obvious charge transfer from V atoms to N atoms leading to the high density of active sites. The unsaturated V-N structures could be high-density active sites responsible for the catalytic activity for the activation of C-H bond in ethane dehydrogenation.
We further simulate the surface structures around the pores and present the high-density active sites composed of unsaturated V-N coordination structures in Figures 4 (a) and 4 (b). The charge density difference diagram indicates the electron-deficient surface around the pores in the PSC VN monoliths. In the pores, the coordination structures are composed of 50%-70% VN and 30%-50% VN 0.87 / 10 nm 50 nm 100 nm   Figure 4(c). The surface is free of Brønsted acid sites because of the absence of the characteristic absorption band near 1540 cm -1 for PSC monoliths. The presence of Lewis acid sites at surface is confirmed by the well-resolved characteristic absorption bands in the range of 1440-1460 cm -1 and 1590-1610 cm -1 [19]. We conduct the test of temperatureprogrammed desorption of ammonia (NH 3 -TPD) experiments in Figure 4 (d) to validate the acidity of the active sites and show that the desorption peak of NH 3 at 150-350°C and 350-500°C can be attributed to the weak and moderate acid sites, respectively. We further show that the doping of Fe and Co at the site of V in lattice increases the acidity of the active sites as validated by the characteristic peak at higher temperatures. We calculate the total density of acid sites when we assume that one NH 3 molecule adsorbs on one acid site. PSC VN, FVN, and CVN monoliths therefore demonstrate the density of acidic sites of 396, 410, and 426 μmol g -1 , respectively. In Figure 4(e) and Figure S17, it is calculated that the theoretical value of the acid site is 11.21 nm -2 for all the PSC nitride samples, which is well consistent with the density of metal atoms at the top layer of the porous single crystal and is~5.5 times higher than that of the polycrystalline nitride samples. We conduct the test of in situ FTIR to clarify the activation of ethane as shown in Figure 4(f). We see the red shift of the peak indicating the high energy state of ethane at higher operation temperatures. Strong adsorption peaks can be observed in the infrared spectrum of 3000-2600 cm -1 that is associated with the C-H bond of the CH 3 group in ethane molecules [20]. The bands at 2968 and 2983 cm -1 designated as C-H bond asymmetric stretching vibration of     Figure 4(h), which indicates the effective chemical interaction between ethane molecules and active sites. The bond lengths of C-C and C-H change by +0.003 and +0.004 Å, respectively. The bond angle between C-C and C-H is also changing to the range of 110.95-111.87°. In the adsorption process, C and H atoms accept the electrons from the V atoms with the adsorption energy of ethane at -0.85 eV, which is favorable to the process of ethane chemisorption at the surface. These results indicate the effective activation of ethane molecules at the high-density active sites at the electron-deficient surface in the PSC VN monolith. We further calculate the transition states of ethane dehydrogenation as shown in Figure 4(i) and present that the pathway of ethane dehydrogenation: C 2 H 6 * → TS1 → C 2 H 5 * → TS2 → C 2 H 4 * . The energy barrier of C 2 H 5 formation is 1.36 eV with a reaction energy of 0.95 eV, and the barrier of C 2 H 4 formation is 1.20 eV with a reaction energy of 0.74 eV. The potential barriers are less than 1.40 eV, indicating that the energy required for reactions is reasonable and the reactions are easy to proceed at the active surfaces. 7 Research of ethane with PSC VN, FVN, and CVN monoliths at 600-660°C. The ethane conversion generally increases versus operation temperature and finally reaches about 29%, 34%, and 36% for the PSC VN, FVN, and CVN monoliths at 660°C, respectively. The ethylene selectivity still maintains at as high as 97%-99% even at the high ethane conversion for all the PSC monoliths at 600-660°C. It should be noted that the ethane conversion generally approaches the thermodynamic equilibrium which indicates the effective activation of ethane in the PSC nitride monoliths. For comparison, we prepare porous polycrystalline (PPC) VN monoliths with a pore size of~40 nm and show that the polycrystalline framework has amorphous layer around the pores in Figure S8 and S18. Figure S19 shows the nonoxidative ethane dehydrogenation using PPC monoliths, which indicates that the catalytic performance is about 5 times lower than that of the PSC monoliths at 660°C. The well-defined surfaces composed of the unsaturated metal-nitrogen coordination structures create the high-density active sites at the clean surfaces and activate the dehydrogenation of ethane to avoid the side reaction. The moderate doping of nonprecious metal Fe and Co elements also effectively engineer the acidity of active sites and further enhance the ethane conversion and ethylene selectivity.
In Figures 5(c)-5(e), the durability test is carried out with PSC nitride monoliths at 660°C for a continuous operation of 200 hours, which shows that the ethane conversion and ethylene selectivity are generally not attenuated. The carbon balance is maintained at 95%-99% during the long-term operation. Figure 5(f) and Figure S20 show the XRD of the PSC nitride monoliths after the long-term stability test, which confirms the absence of impurity phases and carbon species. In Figure 5(g) and Figure S21, the test of Raman spectroscopy reveals the absence of carbon deposition within the detection limit. The HS-LEISS in Figure 5(h) and Figure S22 shows a weak carbon signal by 3 KeV He + ion scattering, which indicates that there are trace amounts of carbon at the top layer of the PSC monolith after the stability test. Although the performance degradation is not observed, the formation of carbon from over-cracking of ethane is still present at a trace amount. In Figure 5(i) and Figure S23, the STEM and SEM present the uniform microstructures with a pore size of~30 nm and well-defined surface structures around the pores even after stability test. The effective structural engineering to tailor the coordination structures at the active surface simultaneously enhances the stability and activity of the PSC monoliths.

Conclusions
In conclusion, we have fabricated porous single-crystalline VN and Co-/Fe-doped VN monoliths at 1 cm scale and then identified and engineered the structural features to create well-defined active sites at surfaces to simultaneously enhance the catalytic activity and stability. The monoliths demonstrate outstanding catalytic performance with ethane conversion of 36% and ethylene selectivity of 99% at 660°C in the nonoxidative dehydrogenation of ethane. The well-defined surfaces composed of unsaturated metal-nitrogen coordination structures contribute to the effective activation of ethane, while the long-range ordering of lattice confines the active surface to deliver enhanced durability. We therefore show the outstanding performance of nonoxidative dehydrogenation of ethane even after a continuous operation for 200 hours. The current work will open new pathway not only to realize the enormous potential of catalytic materials but also to enable dramatic catalytic reaction kinetics improvements across a range of research fields.

Characterization of Porous Single-Crystalline Monoliths.
Atomic force microscopy (AFM, size Icon, Brook) is used to evaluate the topographic information at the surface of parent crystals. X-ray diffraction (XRD, miniflex 600, Rigaku) is used to identify the formation and orientation of the crystal in the range of 10-80°at a scan rate of 15°m in -1 (λ =1.5418 Å). Field emission scanning electron microscope (FE-SEM, SU-8010) with the acceleration voltage of 10 kV is used to observe the microscopic morphology of the PSC monoliths. Surface area and pores are analyzed by N 2 adsorption and desorption on a porosity analyzer (BET, Tristar II 3020). The focused-ion-beam (FIB) nano-tomography (Helios 650, Zeiss Auriga) is used to prepare samples with the thickness lower than 100 nm. FIB nanoscale tomography is performed to sample slice for transmission electron microscope (TEM) characterization. Cs-TEM (FEI Titan3 G2 60-300) is used to analyze the crystal lattice structure and elemental mapping. Selected area electron diffraction (SAED) pattern is executed to clarify the crystal orientation. Raman spectrum is obtained by characterizing the samples with the 532 nm wavelength excitation (LabRAM HR Evolution, Horiba Jobin Yvon) at room temperature.
The characterization of the top layer of the PSC monoliths is performed on a high-sensitivity low-energy ion scattering spectrometer (Qtac100, ION-TOF) equipped with a dual-ring analyzer using 3000 eV He + with 6000 pA current and 5000 eV Ne + with a 3000 pA current as ion sources. X- 8 Research ray photoelectron spectroscopy (XPS) is performed with a ESCALAB 250Xi photoelectron spectrometer equipped with an Al Kα source. The elemental content data for the samples are collected using ICP-OES by UTLIMA-2 (JY, France). The in situ FTIR spectrum is obtained by characterizing the PSC monoliths in the diffuse reflectance mode of the Bruker Optic TENSOR infrared spectrometer equipped with a mercury cadmium telluride (MCT) detector at a resolution of 4 cm -1 over 32 scans. The PSC monoliths are firstly heated to 350°C in He gas flow for pretreatment. The test is performed after adsorbing pyridine at room temperature and then removing the physically adsorbed pyridine in a vacuum. NH 3 -TPD is measured on AutoChem II 2920 (Automatic Catalyst Characterization System). The PSC monoliths are pretreated in He gas flow at 150°C for 1 hour and then reduced to 50°C in a 10% NH 3 /He (V/V) gas flow for 1 hour. The processed PSC monoliths are purged at a temperature of 100°C and kept at the same flow rate for another 1.5 hours. The temperature is increased from 100 to 800°C at 10°C min -1 and kept at 800°C for 10 min.

XAFS Measurements and
Analysis. The X-ray absorption find structure spectra (Fe and Co K edge) are executed by the 1W1B of Beijing Synchrotron Radiation Facility (BSRF). The operating voltage of the BSRF storage ring is 2.5 GeV, and the average current is 250 mA. The EXAFS data are obtained through Si (111) double-crystal monochromator in transmission/fluorescence mode in ionization chamber at the condition of room temperature and atmospheric pressure. The acquired EXAFS data are analyzed by the standard procedures with the ATHENA module in the IFEFFIT software packages. The k 3 -weighted EXAFS spectra are obtained by subtracting the post-edge background from the overall absorption and then normalizing with respect to the edge-jump step. Subsequently, k 3 -weighted χ(k) data of K-edge (Fe and Co) are Fourier transformed to real (R) space using a Hanning windows (dk = 1:0 Å -1 ) to separate the EXAFS contributions from different coordination shells. The least-square curve parameter fitting is executed by the ARTEMIS module of IFEFFIT software packages to obtain the quantitative structural parameters around central atoms [21,22]. The density of Lewis acid sites at surface is calculated according to where n is the density of top layer V atoms of VN at different facet, N is the number of V atom at the top layer of unit cell, a and b are the cell parameters of unit cell, and α is the angle between a and b. We adopt the three strongest X-raydiffraction facets of VN to calculate the average density of V atom at top layer of (200), (220), and (222) The formula of selectivity (Si, %) is below, where Ci is the concentration (vol.%) of the i-th carbon-containing component at the reactor outlet, and ni is a coefficient, which takes into account the number of carbon atoms in the products; selectivity to the reaction products (ethylene and carbon oxides) is calculated using The carbon balance (Y, %) is determined by 4.5. Theoretical Calculation. Density functional theory (DFT) calculations are executed with Vienna Ab Initio Simulation Package (VASP) [23]. The generalized gradient approximation of Perdew-Burke-Ernzerhof (GGA-PBE) is used to represent the exchange-correlation function [24]. And the projector-augmented wave (PAW) method is considered to describe the interaction between the core and valence electrons with a plane wave cut-off energy of 500 eV. The DFT-D3 method [25] and spin-polarization 9 Research are employed for the calculations with the energy and residual forces converged to 10 -5 eV and 0.02 eV/Å, respectively. A unit cell of cubic VN with the structural parameters of a = b = c = 4:14 Å is calculated on an 8 × 8 × 8 k -point grid. To avoid the interaction between the slabs, the vacuum region is set at the thickness of 24 Å. A p (3 × 3) superstructure of VN (200) surface is calculated with four layers where the bottom two layers of atoms are frozen. A 3 × 3 × 1 k-point grid is used to sample the Brillouin zone for VN (200) surface. The possible transition state (TS) is performed by climbing image nudged elastic band (CI-NEB) method [26]. We obtain the adsorption energy of C 2 H 6 using where E total is the total energy of VN (200) surface with adsorbed C 2 H 6 , E C2H6 is the energy of C 2 H 6 in free phase, and E slab is the energy of the model without any adsorption.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare they have no competing interests.
atoms, and blue balls are N atoms. Figure Figure S20: comparison of crystal structure before and after stability test. (a-c) The XRD spectra of the catalysts before the stability test. (d-f) The XRD spectra of the catalysts after the stability test. Figure S21: characterization of Raman spectroscopy. (a-c) The Raman spectra of the catalysts before the stability test. (d-f) The Raman spectra of the catalysts after the stability test. Figure