Adsorption of Transition Metals on Black Phosphorene: a First-Principles Study

Black phosphorene is a novel two-dimensional material which has unique properties and wide applications. Using first-principles calculations, we investigated the adsorption behavior of 12 different transition metals (TMs; Fe, Co, Ni, Cu, Ru, Rh, Pd, Ag, Os, Ir, Pt, and Au) on phosphorene. Our results showed that all of the adsorption systems have a large binding energy. The Fe-, Co-, and Au-phosphorene systems display magnetic states with magnetic moments of 2, 1, and 0.96 μB, respectively, which means that these systems are magnetic semiconductors. Adsorption of oxygen molecules on TM-phosphorene was also investigated. Interestingly, all the O2-(TM-phosphorene) systems, except O2-(Pd-phosphorene), can elongate the O–O bond, which is critical to their application as catalysts in the oxidation of CO. We also found that the adsorption of O2 molecules enables the O2-(Fe-, Ni-, Cu-, Ir-, Rh-, Ag-, and Au-phosphorene) systems to become magnetic semiconductors, and it allows O2-(Co-phosphorene) to display half-metallic state. Our results are expected to have important implications for phosphorene-based catalysis and spintronics.

For two-dimensional (2D) materials, adsorption is usually selected as the approach to induce magnetism for specific applications. Previously, Cao et al. [27] showed that the electronic and magnetic properties of graphene can be effectively modulated by adatoms of Fe, Co, Ni, and Cu. Kaloni et al. [28] demonstrated that magnetic moments can be induced in Ti-, V-, Cr-, Mn-, Fe-, and Co-decorated silicene systems using first-principles calculations. Ersan et al. [29] found that b-Arsenene displayed spin-polarized characters after adsorption of H, B, C, P, Ge, As, and Sb atoms. Furthermore, w-Arsenene can attain net magnetic moments with the adatoms of H, B, N, P, Cl, Ti, As, and Sb. For black phosphorene, Kulish et al. [30] predicted that Ag-, Au-, Ti-, V-, Cr-, Mn-, Fe-, and Co-phosphorene are rather stable, and a diverse range of magnetic moments can be induced in theoretical calculations. Moreover, the properties of different types of charge carriers can also be tuned by adsorbing different atoms on phosphorene. Ding and Wang [31] used the first-principles calculations to systematically illustrate the structural, electronic, and magnetic properties of atoms adsorbed on phosphorene. They noted that adatoms can introduce magnetism in phosphorene, with P, Co, and Au adatoms inducing stable magnetic properties. Hu and Hong [32] used the first-principles calculations to demonstrate the magnetic properties of metal adatoms on phosphorene; they showed that magnetism can be obtained in phosphorene by adsorbing Cr, Fe, Co, or Au atoms on its surface. Furthermore, they predicted that the Fe-phosphorene adsorption system will be a promising dilute magnetic semiconducting material. Thus, the adsorption of transition metals (TMs) on black phosphorene can be expected to effectively tune the magnetic properties of the material.
Although the above investigations studied the adsorption behavior of transition metals on black phosphorene, some issues remain unresolved. For instance, previous studies mainly focused on the properties of 3d TMs adsorbed on phosphorene. How will 4d and 5d TMs engineer the properties of phosphorene? In addition, noble metals absorbed on phosphorene can also be used as single-atom catalysts. Li et al. [33] suggested that silicene with adsorbed Au can be a high-activity catalyst with low catalytic energy barriers for the oxidization of CO. Can a noble metal absorbed on phosphorene also a good candidate for the oxidization of CO? To answer these questions, we present in this paper the results of a detailed first-principles study on the structural, magnetic, and electronic properties of 12 different types of transition metal atoms adsorbed on black phosphorene. We selected elemental Fe, Co, and Ni, which are ferromagnetic metals in their bulk phase; elemental Cu, which is diamagnetic; and the noble metals Ru, Rh, Pd, Ag, Os, Ir, Pt, and Au, which are very effective for the oxidation of CO [19,[34][35][36][37][38][39][40][41][42][43][44][45]. We found that phosphorene forms strong bonds with all 12 metals, and all of the TM-phosphorene systems are rather robust. The electronic and magnetic properties of phosphorene can be effectively tuned by the adatoms. Moreover, we also found that most TM-phosphorene adsorption systems are good candidates for the catalyst in the oxidation of CO. The results of this investigation can be used for fundamental studies of phosphorene, and they can also widen its potential application in many important fields.

Methods/Experimental
Our calculations were based on spin-polarized density functional theory (DFT), and they were performed using the Vienna Ab Initio Simulation Package (VASP) [46,47] and the generalized gradient approximation (GGA) of the Perdew-Burke-Ernzerhof (PBE) functional [48][49][50]. The DFT-D3 method of Grimme [51] was used to calculate the van der Waals interaction. An energy cutoff of 400 eV with a plane-wave basis set was employed. In the calculations, the atoms were relaxed until the total energy converged to 1 × 10 −5 eV and the residual force on each atom was less than 0.01 eV/Å. A large supercell (4 × 3) along the zigzag and armchair directions was used to avoid interactions between neighboring unit cells. The lattice constants were set to a = 13.20 Å and b = 13.74 Å. We applied a vacuum space of 20 Å in the z direction to minimize the interactions between adjacent interlayers. During the optimization, a Monkhorst-Pack [52] k-point grid of 3 × 3 × 1 was adopted, and a k-point grid of 7 × 7 × 1 was used for the total energy calculations.

Results and Discussion
We first explored the structural properties of pristine phosphorene. Figure 1a shows the illustrations of the top and side views of the crystal structure. It can be seen that the phosphorene monolayer consists of two atomic planes, and the unit cell of phosphorene consists of four P atoms. The phosphorene monolayer has a tetragonal lattice with equilibrium lattice constants a = 3.30 Å and b = 4.58 Å. The length of the P-P bond in the horizontal direction (l 1 ) is 2.22 Å, while the length in the other direction (l 2 ) is 2.26 Å. The pristine phosphorene has a direct bandgap of 0.89 eV (Fig. 1b), with both the conduction band minimum (CBM) and the valence band maximum (VBM) located at the Г point. The lattice constant and the bandgap we obtained highly agree with the values obtained in previous research studies [30][31][32]53].
A typical adatom is always adsorbed at either one of three positions: above a hollow site (H), on a bridge (B) between two phosphorus atoms, and on top of a phosphorus atom (T). We calculated the adsorption energy of an adatom on phosphorene to examine the stability of the adsorption systems using the relationship: where E TM is the energy of an isolated metal atom, E phosphorene is the total energy of the pristine phosphorene layer, and E TM-phosphorene is the total energy of the adsorption system. Based on this equation, a larger adsorption energy indicates a more stable structure. We found that all the metal atoms studied in our work prefer to stay on the H site of phosphorene. The calculated adsorption energies of metal atoms adsorbed on the H site of phosphorene, shown in Table 1, vary from 2 to 6 eV. The bond length of TM-phosphorene (d TM-P ) was demonstrated to be short, in the range of 2.11-2.43 Å. Bader charge analysis [54][55][56] shows that 0.16, 0.16, 0.07, 0.17, 0.32, 0.33, and 0.16|e| are transferred from the Ru, Rh, Pd, Os, Ir, Pt, and Au metal atoms, respectively, to phosphorene in the (4d-TM)-phosphorene and (5d-TM)phosphorene adsorption systems. All these results denote the formation of chemical bonds between the TM adatom and phosphorene. In addition, these results are close to recent studies [30][31][32].
As shown in Table 1, the Ni-, Cu-, Ru-, Rh-, Pd-, Ag-, Os-, Ir-, and Pt-phosphorene systems exhibit nonmagnetic states, while the Fe-, Co-, and Au-phosphorene systems have the magnetic moments of 2, 1, and 0.96 μ B , Table 1 Calculated minimum bond length of TM-phosphorene (d TM-P ), adsorption energy (E ad ), total magnetic moment (M total ), and charge transferred from TM adatom to phosphorene for a single TM atom adsorbed at the most stable adsorption site on phosphorene   (Fig. 3b), at a certain angle from the surface. Meanwhile, the two neighboring O atoms around the TM adatom are not equivalent. The results are displayed in Table 2. The adsorption energy (E ad ) of O 2 on an O 2 -(TM-phosphorene) system was calculated as: where E O 2 −TM−phosphorene , E TM-phosphorene , and E O 2 are the total energies of the O 2 -(TM-phosphorene) system, the TM-phosphorene system, and the O 2 molecule, respectively. As shown in Table 2   In order to obtain more insight into the underlying mechanism of the high activity of these systems, we selected O 2 -(Pt-phosphorene) as an example and investigated its local density of states (LDOS). Figure 4a shows the LDOS projected onto d orbitals of Pt in the Pt-phosphorene system, d orbitals of Pt in the O 2 -(Pt-phosphorene) system, the O-O bond in the O 2 -(Pt-phosphorene) system, and the gas phase O 2 . In the upper panel of Fig. 4a, one peak can be seen at E F − 0.6 eV, which originates from the partially occupied d orbital of Pt in the Pt-phosphorene system. These states should be responsible for the high activity of the Pt-phosphorene system. After the adsorption of an O 2 molecule, the LDOS projected onto d orbitals of Pt below the Fermi level is downshifted after the adsorption of the O 2 molecule owing to the charge transfer, and the states above the Fermi level is also substantially increased. Meanwhile, the LDOS projected onto the adsorbed O 2 molecule indicates that the O 2 2π * orbitals (lowest unoccupied molecular orbital, LUMO) are becoming partially occupied, which has downshifted from its gas value of E F + 2 eV to E F − 0.1 eV. For clarification, the charge density difference of the O 2 -(Pt-phosphorene) system is also presented.
The charge density difference is defined as follows: where ρ T , ρ molecule , and ρ absorbed are the total charges on the O 2 -(Pt-phosphorene) system, O 2 molecule, and the Ptphosphorene system, respectively. As shown in Fig. 4b, the large yellow region localized on the O 2 molecule indicates that there is a significant electron transfer from Ptphosphorene to O 2 , which also indicates the strong orbital hybridization between O 2 and the Pt-phosphorene system.    Table 3. To better comprehend how the adsorption of a gas molecule affects the electronic structure of the O 2 -(TM-phosphorene) system, the electronic band structures of each system was calculated, and the results are shown in Fig. 5. First, we discovered that a flat band occurs around the Fermi level (E F ) after the ad-

Conclusions
We investigated the structural, electronic, and magnetic properties of different TM-phosphorene systems. All the adatoms were found to prefer to occupy the hollow site on phosphorene. The considerable adsorption energy reveals that all of the TM-phosphorene adsorption systems are rather robust, indicating that phosphorene forms strong bonds with all 12 types of TM adatoms. Furthermore, we found that doping with Fe, Co, and Au can result in magnetic semiconducting properties in monolayered phosphorene, with total magnetic moments of 2, 1, and 0.96 μ B , respectively.
In addition, we also examined the properties of an O 2 molecule adsorbed on the TM-phosphorene system. It was very encouraging to find that all of the O 2 -(TM-phosphorene) systems, except for O 2 -(Pd-phosphorene), display good catalytic activity for the oxidation of row; the corresponding band structure of each system is shown in the bottom row. In the top row, a plot of the spin-polarized charge density with a charge density iso-surface value of 0.002 e/ Å 3 is superimposed on the top and side views of the crystal structure of pristine phosphorene; the yellow and cyan regions correspond to up and down spins, respectively. In the plots of band structures, the black and the red lines denote spin-up and spin-down channels, respectively; the Fermi level is set to zero, and it is indicated by the gray dashed line