Xinle
Lu
ab,
Liaoxin
Sun
b,
Bing
Fu
c,
Shoutian
Sun
*a and
Xiang
Ye
*a
aDepartment of Physics, Shanghai Normal University, Shanghai 200234, P. R. China. E-mail: stsun@shnu.edu.cn; yexiang@shnu.edu.cn
bState Key Laboratory of Infrared Physics, Chinese Academy of Sciences, Shanghai 200083, P. R. China
cFuzhou Medical College of Nanchang University, Fuzhou 344000, P. R. China
First published on 7th September 2021
The effect of adsorbed adatoms on the structural stability and electronic properties of monolayer N2P6 have been systematically studied via first-principles simulation methods. It is found that pristine N2P6 is an indirect 0.21 eV band gap semiconductor, with a pleated honeycomb-like structure similar to phosphorene. The calculation results show that adsorbed adatoms can modify the properties of monolayer N2P6 effectively. The degree of local distortion strongly depends on the electronegativity and size of adatoms, also the adsorption energy ranges from 0.3 to 5.8 eV depending on the species of adatoms. The electronic properties show metallic behavior with several adsorbed metal atoms (Li, Na, Al, K, Cu, Ni, and Zn) and some non-metal atoms (H, F, P, and Cl), while adsorbed O, S, Ca, and Si atoms still remain semiconductors. The systems of Ni and Zn adatoms show ferromagnetic behavior, and adsorbed Mg exhibits a half-metallic character. Our theoretical studies indicate that N2P6 possesses potential application in the field of gas sensors.
Graphene, as a single-layer honeycomb lattice of carbon atoms, has been widely used in various applications including solar cells,15 electrodes,16 wearable devices,17 and sensors,18,19 owing to its high transparency, high thermal conductivity and high specific surface area. However, the zero bandgap of graphene20 with a smaller switching current ratio prohibits its applications as a logic switch in semiconductors. Several graphene-like materials3,21 and their allotropes have been synthesized experimentally. It has been found that defects and adsorption of adatoms on top of graphene-like C2N, C3N, C3N4, and C4N3 can modify their electrical and magnetic properties, which could be useful for application in gas sensors.22–25
Although a 2D transition metal sulfide (such as MoS22) has a high current switching ratio, its carrier mobility is much lower than that of graphene, which limits the application of transition metal sulfides in electronics.26 In 2014, phosphorene was successfully exfoliated from the bulk black phosphorus (BP) phase using mechanical exfoliation.27 It is a direct semiconductor with excellent electrochemical performance. And as an electrode material, phosphorene has an ultra-high specific capacity of 2596 mA h g−1 in lithium-ion batteries, which is much higher than that of graphite.28,29 Owing to phosphorene's “puckered” honeycomb structure, it possesses a much higher surface to volume ratio and lower out-of-plane electrical conductance, while its chemical stability is lower than those of graphene and TMDCs.30 Adsorption of gas molecules on phosphorene has been extensively investigated and it has been demonstrated that phosphorene could be a superior gas sensor.31,32 Abbas et al. found that the phosphorene based FET sensors have excellent sensitivity to 5 ppb NO2.33 In particular, researchers have recently shown the superior selectivity of green phosphorene towards the decomposition of numerous SF6 gases.34 However, black phosphorene can only show strong physical adsorption to nitrogen-containing gas molecules, such as NO, NO2 and NH3, which inhibits its application for gas detection. Subsequently, more efforts have been made to improve the adsorption ability of black phosphorene, including surface decoration.25,35 In addition, binary phosphorus nitrides show potential in electronic devices and optoelectronics, drawing the attention of researchers. Ma et al. predicted three 2D phosphorous nitride (PN) single-layer sheets (named α-, β- and γ-PN) with superior mobility based on first-principles methods.36 Dong et al. designed three 2D materials (named P2N2, P3N and δ-P3B) exhibiting extraordinary stability. Among them, P3N with 3% strain could become a good photocatalyst for water splitting.37 What's more, most 2D materials of group V elements were synthesized at high pressure and high temperature, such as the BP allotrope of nitrogen (bp-N), α-P3N5 and γ-P3N5, have excellent thermal, mechanical and chemical stability.38,39
Thus, based on the excellent properties of phosphorene in gas sensing and the structural stability of binary phosphorus nitrides, we have predicted a kind of nitrogen and phosphorus compound N2P6. Electronic properties and crystal orbital Hamilton population (COHP) were studied via density functional theory (DFT) calculations.40–43 Pristine N2P6 is a pleated honeycomb-like 2D semiconductor with an indirect band gap of 0.21 eV. The adsorption of 22 different atoms including alkali metal (Li, Na, and K), alkaline earth metal (Be, Mg, and Ca), and 3d transition-metal (Fe, Co, Ni, Cu, and Zn) elements, and nonmetallic atoms (H, B, C, N, O, F, Si, P, S and Cl) and Al on N2P6 have been studied. It is found that some adatoms (B, Be, C, N, Fe, and Co) modify the geometry of the substrate, and other adatoms can become adsorbed at four stable adsorption sites of N2P6, except Li and Zn adatoms that show a small deviation from the special adsorption sites. The results also show that H, Na, Mg, Al, K and Ca prefer to stay on the top of N atom of N2P6, while non-metallic adatoms O, F, S and Cl prefer to stay on the top of P atom. Other atoms are adsorbed on the top of the P6 ring (Si, Ni and Cu) and the P–P bridge (P) site. The electronic properties reveal that different species of adatoms can modulate the bandgap of the system effectively which could give rise to the donor and acceptor states. Even Ni and Zn adatoms make N2P6 attain its magnetic properties. Furthermore, Ken et al. verified the formation of new PxNy monolayers by laser heating under 67 and 70 GPa pressure, which indicates the possibility of synthesis of the N2P6 monolayer in the future.44 We believe that our results will help to understand the properties of adsorption N2P6 and promote its application in the field of gas sensors.
Due to the different electronegativities (N: 3.04; P: 2.19), a charge transfer occurs from P to N atoms. As seen in Fig. 1(a), there is a higher charge density around N atoms compared to P atoms. Each N atom obtains about 1.74 electrons from its neighbor P atoms according to Bader charge analysis, which is consistent with the electronegativity model and charge density difference as shown in Fig. S2 (ESI†). According to our results, the P atoms are positively charged, whereas the N atoms are negatively charged, resulting in a strong bonding character of N–P bonds. Moreover, the total charge density revealed that spatial configuration of N atoms is tetrahedral, verifying that it is the sp3 nonequivalent hybridization of N and P atoms, which induces the folded two-dimensional honeycomb-like structure.
The projected band structure of N2P6 is shown in Fig. 2. The results show that the N2P6 is an indirect semiconductor (bandgap: 0.21 eV), with the valence band maximum (VBM) and the conduction band minimum (CBM) located at M and Γ points, respectively. It can be seen that the VBM is predominantly contributed by the pz orbitals of N and P atoms, while the CBM is mainly composed of the pz orbital of P atoms. For elements of group V monolayers or binary phosphorus nitrides (P3N5), the PBE_GGA functionals commonly underestimate the size of the bandgap.55 Hence, we adopted Heyd–Scuseria–Ernzerhof (HSE06) exchange–correlation functionals to predict reliable electronic structures. The estimated band gap is 1.05 eV as shown in Fig. S3 (ESI†).
Fig. 2 Electronic structure with orbital resolved band structure, the corresponding PDOS and –COHP results of the N2P6 unit cell. The Ef is set to zero as shown by the dashed line. |
To achieve an in-depth understanding of this structure, the partial density of states (PDOS) and the COHP have been calculated and analyzed. The PDOS and –COHP are depicted in Fig. 2, confirming that this structure is relatively stable and could be synthesized experimentally. Notice that N_pz and P_pz dominate in the PDOS around the Fermi level; thus we focus on –COHP of N–P and P–P bonds to analyze the bonding character. The coupling of N/P_px,y,z orbitals are the main source in the formation of covalent N–P bonds. For optimized N2P6, there is three-fold rotational symmetry Ĉ3 at the position of the N atom, and all of the coupling between N and P is exactly the same. Thus, a typical N1–P1 (red arrow in Fig. 1a) bond was selected to perform COHP analysis. As the -COHP images in Fig. 2 show, the N/P_pz interaction (ondine area) produce part of anti-bonding components below the Fermi energy, and almost of the occupied states are bonding components for the coupling of N/P_py (pale orange area). Two different P–P bonds are observed (blue arrow in Fig. 1a), one (P2–P3), perpendicular to a-axis, is that P/P_py (rose area) interaction which is stronger than other orbitals; other P–P bonds (P2–P4) intensities mainly come from the px, pz orbital coupling. Their –COHP indicates that the bonding of these two kinds of P–P is stable. On the other hand, the N–P (ICOHP: Integral Crystal orbital Hamilton population = −3) bonding strength is stronger than the P–P (ICOHP = −2.3) one. The strong interaction between N_pz and P_pz induces splitting of the energy band near the Fermi level.
Fig. 3 (a) A schematic illustration of top and side views of the embedded single atom, and (b) adsorption sites of the atoms on monolayer N2P6 as presented in Table 1. |
Notice that adsorbed B, Be, Co, Fe, and N atoms on N2P6 break the substrate structure due to the strong interaction between adatoms and N2P6. For other systems, we calculated the ICOHP value of the system and found the results to be negative as shown in Table 2. This indicates that all bonds are stable as ICOHP values for all of the N–P, P–P bonds in the adsorption system, approximately the same as that of pristine N2P6. Hence, the Ad–N2P6 systems are stable. We used the biggest change (Δl = |lp–p − lp–p(sub0)|, lp–p(sub0) = 3.183 Å) of the distance (lp–p) between the P atom and its third nearest atom to describe the in-plane local distortion, and out-plane buckling of N2P6 is defined by the height change (Δh = |h − hsub0|, hsub0 = 1.338 Å, hsub0 is the height of pristine N2P6) when compared against pristine N2P6. Fig. 4 shows the tendency of charge transfer (CT) and distortion magnitude (Q) which is described by in-plane distortion Δl and out-plane distortion Δh, that is (Fig. 4(b)).
The results show tendency to have high consistency in structural distortion and charge transfer of adatoms. To better understand the structural distortion, we analyze the variation of structural parameters, including bond length, height, bond angle and buckling for different adatoms adsorbed N2P6 (Fig. S4, ESI†). Adsorptions on the TN and HPP site mainly exhibit in-plane distortion and have similar degree of in-plane and out-of-plane distortion to CT, which further verifies the relevance between charge transfer and distortion. As for the TP site, there are small changes in the distortion magnitude. Hence, the variation of the relative angle (Δθ = θ − θ0, θ0 = 88.89°) for each system has been studied, which is defined as the maximum variation of the angle θ made up by adatoms nearest the P atom and its first and second P–P bond, and we found that adsorption behavior mainly cause alteration of N/P atomic position (bottom panel of Fig. S4, ESI†). However, focusing on the area closest to the adsorbed atom, adatoms have more impact on dNP and dPP at TP sites than TN and TPP sites. This indicates that the distortion by TP adsorption is more localized than other adsorption sites. The length of N–P bonds is smaller than the substrate while the P–P bond is longer at the stable TN site (light blue area), the main reason being the charge transfer between the N and adatoms elongates N–P bond length. Similarly, the length of N–P, P–P bonds decreased on TP, HPP sites (light green, light red area), and adsorbed O, F, S, and Cl atoms mainly cause out plane bulking, but retain the atomic structure to a great extent compared to TN, HPP adsorption sites as shown in the top panel of Fig. S4 (ESI†).
Then, we defined a physical quantity, charge transfer capability which is obtained by the charge transfer value and the maximum electron number the adatom can gain or lose, to verify the accuracy of charge transfer (Take the H atom as an example; it can lose at most one electron, if the actual charge given is 0.418, the value of . Similarly, there is 2e located in the outermost orbit of Mg atom, if the actual charge given is 1.483, then we could get ) and use it to compare electronegativities. For TN and HPP adsorption sites, the curve trend of charge transfer capability shows the opposite behavior of electronegativity due to adatom electronegativity being lower than those of N/P atoms (3.04/2.19). And the trend of charge transfer capability is similar to the electronegativity for TP site adsorption. Obviously, it proves that the charge transfer is dominated by electronegativity. In general, the extent of structural distortion depends on the electronegativity of adatoms. The adsorption energy is defined as Ea = EAd/N2P6 − EAd − EN2P6, where EAd/N2P6 is the total energy of the structure with adatom adsorption on N2P6, EN2P6 is the total energy of pristine N2P6, and EAd is the total energy of an isolated adatom in a vacuum (Table 1). The calculation results show that adsorbed atom systems are stable (Ea < 0). The adsorption energies of adatoms located at the stable TN, TP, HPP, and BPP sites ranges from −5.9 to −1.9 eV, which can be considered as chemisorption.
Adatom | Site | d TC | d NP | d PP | h | θ | l P–P | C T | E a (eV) | E g (eV) | M (μB) |
---|---|---|---|---|---|---|---|---|---|---|---|
Al | TN | 2.095 | 1.852 | 2.262 | 1.559 | 84.683° | 3.124 | −0.877 | −2.962 | M | 0 |
Mg | TN | 2.076 | 1.859 | 2.250 | 2.085 | 80.490° | 2.910 | −1.482 | −1.929 | HM | 0 |
H | TN | 1.039 | 1.885 | 2.253 | 1.342 | 90.947° | 3.212 | −0.418 | −2.664 | M | 0 |
Ca | TN | 2.293 | 1.844 | 2.250 | 1.943 | 82.794° | 2.978 | −1.456 | −2.959 | 0.21 | 0 |
Na | TN | 2.366 | 1.823 | 2.259 | 1.595 | 86.872° | 3.104 | −0.851 | −2.648 | M | 0 |
K | TN | 2.799 | 1.859 | 2.259 | 1.540 | 87.684° | 3.130 | −0.866 | −4.086 | M | 0 |
Cl | TP | 2.066 | 1.671 | 2.207 | 1.373 | 97.570° | 3.321 | 0.428 | −2.547 | M | 0 |
S | TP | 1.951 | 1.728 | 2.249 | 1.438 | 94.542° | 3.304 | 0.603 | −3.602 | 0.304 | 0 |
O | TP | 1.496 | 1.716 | 2.248 | 1.384 | 95.241° | 3.321 | 1.286 | −5.852 | 0.259 | 0 |
F | TP | 1.608 | 1.653 | 2.200 | 1.428 | 99.821° | 3.366 | 0.603 | −4.347 | M | 0 |
Cu | HPP | 2.309 | 1.763 | 2.254 | 1.632 | 94.208° | 3.057 | −0.300 | −2.858 | M | 0 |
Si | HPP | 2.419 | 1.757 | 2.223 | 1.868 | 101.10° | 2.899 | −0.763 | −3.778 | 0.265 | 0 |
Ni | HPP | 2.183 | 1.777 | 2.264 | 1.704 | 95.342° | 2.965 | −0.178 | −4.187 | M | 0.590 |
P | BPP | 2.185 | 1.715 | 2.248 | 2.169 | 74.583° | 2.737 | −0.036 | −2.992 | M | 0 |
Li | 2.036 | 1.845 | 2.256 | 1.603 | 88.885° | 3.160 | −0.883 | −3.071 | M | 0 | |
Zn | 2.266 | 1.842 | 2.248 | 1.576 | 86.929° | 3.085 | −0.544 | −0.344 | M | 0.494 |
The structural, electronic, and magnetic parameters of adsorption structures involving TN, TP, HPP and BPP adsorption sites are listed in Table 1. This includes the bond length between adatom and its nearest atom (dTC), the bond length of N–P atoms (dNP), the bond length of P–P atoms (dPP), the height of adsorption structure, the charge transfer (CT) between the embedded atom and N2P6 monolayer, the adsorption energy of adatom (Ea), and magnetic parameters are measured by the magnetic moment per supercell (μ). An electronic state is specified as metal (M), half-metal (HM), or semiconductor (SC), and correspond to the band gap of the adsorption system (Eg).
Adatom | ICOHP | Adatom | ICOHP | |
---|---|---|---|---|
TN | H–N | −3.56789 | Al–N | −1.51874 |
Mg–N | −0.25946 | Ca–N | −0.21766 | |
Na–N | −0.18104 | K–N | −0.09594 | |
TP | O–P | −5.03522 | F–P | −3.44282 |
S–P | −4.09341 | Cl–P | −2.88108 | |
HPP | Si–P | −1.805 | ||
Zn–P | −0.965 | Li–P | −0.967 | |
BPP | P–P | −2.637 |
The band structure and DOS at the stable TN site adsorption are shown in Fig. 6. These new structures show metallic behavior after adsorbing H, Na, Al, and K atoms. The calculated Bader (CT) demonstrates H, Na, Al and K adatoms transfer 0.418, 0.851, 0.877 and 0.866 electrons to N2P6, respectively. It can be seen Fermi energy shifts up into the conduction band. The energy band dispersion of N2P6 of TN site adsorption is weakly modified by the adatoms compared with pristine N2P6. From the PDOS and the fatband in Fig. 6, it can be seen the energy bands right below and above Ef are mainly contributed by the hybridization of N_s and P_pz. However, the Mg–N2P6 and Ca–N2P6 conduction band are strongly modified by adatoms because the s orbitals of adatoms is located at CBM, as well as the Fermi energy of systems shifts up. The system transforms into semi-metal after embedding Mg, with VBM and CBM located at K and Γ point, respectively. The Ca–N2P6 system still stays an indirect semiconductor with band gap of 0.21 eV, with VBM and CBM located at K and M point, respectively. In addition, the PDOS main peaks contributed by the metal adatoms locate from conduction band to valence band depending on the atom electronegativity. It implies that the chemical bond between metal adatoms and their nearest N atoms changes from ionic bond (K–N) to covalent bond (Al–N), which is consistent with the ICOHP results.
Fig. 6 Orbital resolved band structure, DOS and PDOS of adsorbed H, Na, Mg, Al, K and Ca adatoms at the TN site. The Ef is set to zero as shown by the red dashed–dotted line. |
Similarly, we study the non-metal adsorption of O, F, S and Cl adatoms on TP site. 1.50–2.07 Å of dTC implies that the bonding of adatoms and neighboring P atom is chemical bonding. Meanwhile, relative value of ICOHP (Table 2) also confirms the strong interaction between adatoms and N2P6, which proves that the bond of Ad–P should be covalent bonding. As can be seen in Fig. 7, the interaction O_pz and N_pz shows the bonding component below Fermi level. This indicates that the adsorbed O atom makes the structure more stable for the O–N2P6 system. In addition, the O (S, F, and Cl)-N2P6 structures transform to p-type semiconductors after adsorbing adatoms.
Fig. 8 shows the fatband structure, DOS and PDOS of O, F, S and Cl decorated adsorbed systems. The O (S) adatom interact with neighboring P atom of N2P6via sp3-hybridization, resulting in no in-plane distortion of the structure, which is consistent with conclusion of structural change. The interaction between O/S and N2P6 are mainly comprised of the hybridization of O/S_p with P_p orbitals of the nearest atoms. Interestingly, the energy overlap ranges from −2 eV to −3 eV and −0.5 e V to 1.5 eV for the corresponding O and S adsorption system, respectively. There is no influence on electronic states near Ef because the coupling is far away from VBM/CBM. Thereby the O (S)–N2P6 system is a direct semiconductor with a bandgap of 0.259 (0.3) eV, where the VBM/CBM is located at the Γ point. On the contrary, the interaction of F/Cl with substrate eliminates the semiconducting band gap of N2P6, and achieves metallic properties. The reason can be attributed to the much higher electronegativity of F/Cl atoms.
Fig. 8 Orbital resolved band structure, DOS and PDOS of adsorbed O, F, S and Cl adatoms at the TP site. The Ef is set to zero shown by the red dashed–dotted line. |
Si, Ni and Cu adatoms have similar electronegativity. As shown in Fig. 9, the DOS distribution displayed in Fig. 9 focuses on understanding the electronic properties of adsorption. The band structure and DOS shift down relatively compared to pristine N2P6, and the electronic properties are totally different for Si, Ni and Cu adsorption. For example, it is quite different from the DOS before and after Si adsorption, the band structure moves down 0.5 eV after Si adsorption with respect to pristine N2P6, and reforms an indirect band gap of 0.3 eV. This proves that there is strong interaction between Si and the substrate, which reflects that the system becomes more stable after Si exposure. The corresponding ICOHP values (Table 2) and –COHP (Fig. 7) also imply the same conclusion. Moreover, the PDOS reveals that the valence electronic orbital of adatom show sp3 behavior, forming a tripod-like structure with the nearest three P atoms, and the partial hybridization of s and pxy is obvious for P atoms in N2P6. As a result, not only does the length of Si and P bond increase from 2.22 to 2.41, but also the CBM stems from the hybridization of N_s (black line, Fig. 9) with P_pz orbitals (grey grid, Fig. 9), and the VBM originates from the hybridization of P_sp and N_s, pxy orbitals. More importantly, the partial hybridization behavior of P atoms may lead to reopening of a band gap in the adsorption structure as bonding and antibonding.
Fig. 9 Orbital resolved band structure, DOS and PDOS of adsorbed Si, Ni and Cu adatoms at the HPP site. The zero of energy is set to Ef, shown by the red dashed–dotted line. |
Unlike the original N2P6, the embedding of magnetic Ni atom gives rise to some localized states, destroys the spin degeneracy of the system, and modifies the electronic and magnetic properties. The Fermi energy shift proves that Ni atoms serve as an electron donor, agrees well with charge transfer results. Energy bands split into spin up and spin down channels. The spin down (orange dotted line) channel appears the localized states which also become metallic character, while spin up (black solid line) exhibits a metal, inducing a ferromagnetic metallic with 0.59 μB magnetic moment. The DOS is shown in Fig. 9, it can be seen spin-splitting appears below Ef, with the d electronic states of Ni mainly existing below the Ef.
The Cu–N2P6 exhibits a metal character. The semiconductor to metal transition caused by charge transfer of adatoms, and adatom serves as the electron donor and transfers 0.3 electrons to the substrate, which makes Ef move up 0.5 eV (Fig. 9). Though 3d orbital states of the Cu atom has high energy values, the calculated PDOS shows 3d-orbital of Cu adatom located below −1.5 eV, which could hardly modify electron states of N2P6 around Ef.
For the P–N2P6, P adatom located on the BPP site, 2.185 Å of dTC is in conformity with the principle of P–P bond length, indicating a system of chemisorption. As shown in Fig. 7, the –COHP of P adatom and its nearest atom is similar to those of metal adatoms, but there is covalent bonding between them based on the data of ICOHP. The P–N2P6 system exhibits metal nature form the band structure, with P adatom transform 0.299e to its two nearest P atoms on N2P6 (Fig. S5, ESI†). P adatoms interact through sp3-hybridization and form two σ bonds to two neighboring P atoms (Fig. S5, ESI†).
The Li, Zn adatoms as well as the previous TN–N2P6, serve as electron donors transferring 0.883e and 0.544e to pristine N2P6, respectively. From top and side views of the structure of Li, Zn–N2P6 as shown in Fig. 10(b), we found Li adsorption of non-special site relatively increases the degree of out-plane distortion while the bucking of in-plane is very small. Zn adsorption causes distortion both in out-plane and in-plane of the structure, which results in the breaking of the system.
As shown in (Fig. 10), it can be seen the Fermi energy shifts toward low energy adsorbed Li atom from the band structure and DOS, which is similar to Na–N2P6. However, different from Ni adsorption, the ground state of Zn adsorbed by N2P6 destroys the spin degeneracy of the system which is nonmagnetic. Spin-splitting can be found in the vicinity of Ef of the main 4s-orbital peaks for Zn adatom. On the other hand, we found impurity level is introduced at the Ef level making Zn–N2P6 metal ferromagnetic. From the PDOS in Fig. 10(a), there are some localized states located at Fermi level for N and P atoms of N2P6. Metallic states originate from the Zn_4s spin up with N and P_sp spin up orbital states. The DOS of Zn adatom located below Ef illustrates the adsorption system is stable. As seen from –COHP in Fig. 5, bonding between Zn_s and N_pz is the key factor which induces spin-splitting of the band structure.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1cp03211a |
This journal is © the Owner Societies 2021 |