3.1 Li and P functionalization on g-C3N4
The optimized structure of pristine g-C3N4 monolayer is shown in Fig. 1(a), and a cell constant value of a = 7.134 Å = b is obtained, which is consistent with previously simulated (7.14 Å) as well as experimental (6.810 Å) outcomes [31, 32]. According to the symmetry of the g-C3N4, there are two distinct carbon atoms (C1 and C2) and three distinct nitrogen atoms (N1, N2, and N3) (Fig. 1a) [33]. There are five substitutional sites (C1, C2, N1, N2, and N3) and two interstitial sites (I1 and I2), as shown in Fig. 1(a), for the loading of metals and non-metals in heptazine-based g-C3N4 [18]. The large cavity (I2 site) is the steadiest position for metal loading, and the P atom at I1/I2 site is the most reliable configuration compared to other doping sites [34, 35]. Therefore, we have considered a large cavity for the Li atom and the I1 site for the P atom. In the case of single Li atom adsorption on the g-C3N4 monolayer, the Li atom is positioned in the center of the membrane plane's pore, with the adsorption energy of − 4.36 eV, very close to the result reported in the literature [36]. However, three Li in the large pore orient directly above (2.04 Å above the monolayer plane) the cavity with a slight shift of internal nitrogen towards the Li atoms after optimization (Fig. 1(c)) with adsorption energy of -3.61 eV [37]. The adsorption energy (Eads = -3.61 eV) of Li on the g-C3N4 monolayer is considerably more than the cohesive energy (Ecoh = -1.63 eV) of Li [13]. This adsorption energy value shows an even dispersal of Li dopants without cluster formation. As shown in Fig. 1(c), the length between the Li atom and the adjacent N atoms is ~ 1.89 Å, while the distance between two Li atoms is about 3.52 Å. In the case of Li and P functionalized g-C3N4, we have considered a g-C3N4 monolayer having three Li atoms and one P atom in the empty cavity at a distance of 2 Å from the monolayer plane. After the geometry optimization, the P atom sits 2.37 Å above the monolayer plane, and the space between the P and the neighboring C and N atoms are 1.79 Å and 1.83 Å, respectively, as shown in Fig. 1(d). The high adsorption energy value (-3.94 eV) ensures the stability of the Li and P functionalized g-C3N4 monolayer. This configuration is carefully chosen to guarantee judicious space for the adsorption of H2 all over the Li and P on the two sides of the monolayer.
The electronic band structure, along with the partial density of states (PDOS), is also used to study the functionalization of Li and P into the g-C3N4 monolayer. The band structure indicates the semiconducting nature of the g-C3N4 monolayer having a bandgap of 1.1 eV (Fig. 2(a)), which is less than the reported experimental value of 2.7 eV due to the well-identified fact that GGA-PBE underrates the bandgap energy [38]. Nonetheless, while the band gap calculated from GGA-PBE is not close to the experimental value, it gives vital insight into the impacts of element doping on g-C3N4, particularly on partial DOS, projected DOS, electronic band structure, and optical properties [39, 40]. Hence, as discussed in the manuscript, we have used GGA-PBE to calculate the properties of all the doped systems. After the Li addition to the g-C3N4 monolayer, the band structure changes significantly due to the allocation of extra electrons from the Li to N, causing the Li-doped g-C3N4 monolayer to become metallic [37]. After adding Li and P to the g-C3N4 monolayer, its metallic nature diminishes with a feeble bandgap of 0.036 eV, as shown in Fig. 2(c). The partial density of states (PDOS) of valence s orbital of the Li atom and p orbital of P, C, and N atoms (close to the Li and P atoms) are shown in Fig. 2(d)-(f). In a pure g-C3N4 monolayer, the valence band is principally contributed by the 2p orbital of the N atom, while the conduction band is conquered by the 2p orbital of the C atom (Fig. 2(d)). Over the incorporation of Li atoms in the g-C3N4 monolayer, mid-band states appear near the Fermi level due to the transfer of additional electrons from Li to N, as revealed in Fig. 2(e). The binding of Li to the monolayer is indicated by the substantial overlap between the 2s orbital of Li and the 2p orbital of N near the Fermi level. A similar trend can also be observed in the PDOS of Li and P added g-C3N4 monolayer, as visible in Fig. 2(f).
The frontiers MOs of the pristine-(g-C3N4)8, (g-C3N4)8Li3, and (g-C3N4)8Li3P monolayers are given in Fig. 3 to show the charge localization and delocalization. In the pristine g-C3N4 monolayer, there is a localization of photogenerated \({e}^{-}/{h}^{+}\) couples in a specific heptazine unit, which results in an extreme recombination proportion of \({e}^{-}/{h}^{+}\) couples. This charge localization explains the poor photocatalytic performance of pristine g-C3N4. In (g-C3N4)8Li3 and (g-C3N4)8Li3P monolayers repositioning of HOMO and LUMO is observed, which indicates the delocalization of photogenerated \({e}^{-}/{h}^{+}\) pairs. This delocalization of \({e}^{-}/{h}^{+}\) pairs in the Li and P functionalized g-C3N4 monolayers result in the high photocatalytic performance related to the pristine g-C3N4 monolayer.
Next, we have calculated the electron localization function (ELF) plots of the pristine-(g-C3N4)8, (g-C3N4)8Li3, and (g-C3N4)8Li3P monolayers as shown in Fig. 1(b) and Fig. S2. It has been noticed from the ELF plots that the C-N bonds are covalent, while Li-N and P-N are ionic.
As displayed in Fig. 4, we have also calculated the optical properties, for example, refractive index, reflectivity, optical conductivity, and optical absorption of pure, Li functionalized, and Li and P functionalized graphitic carbon nitride. There is a substantial perfection in the reflectivity in the infrared area after adding Li and P atoms to the pure graphitic carbon nitride, as revealed in Fig. 4(b). Optical conductivity is considered a crucial means for studying electronic circumstances in materials. Figure 4(c) shows robust growth in the optical conductivity in the infrared as well as visible area after adding Li and P atoms. Li-functionalized graphitic carbon nitride shows better optical absorption in the infrared region. However, Li and P functionalized graphitic carbon nitride has the best absorption in the visible and ultraviolet region compared to pristine as well as Li functionalized g-C3N4 monolayers, as shown in Fig. 4(d).
3.2 Adsorption of H2 on the Li and P decorated g-C3N4 monolayer
After confirming the stability of Li and P functionalized graphitic carbon nitride monolayer, we have explored the interaction of the H2 molecule with Li and P decorated g-C3N4 monolayer by placing H2 molecules over Li and P atoms in the supercell. The optimized structures of different configurations of Li and P functionalized graphitic carbon nitride with the maximum quantity of H2 adsorbed only on one side of the monolayer are given in Fig. 5. After geometry optimization, the adsorption energy, bond length, average Li-H2 distance, and average P-H2 distance are noted in Table 1. At first, a single H2 molecule is placed near the individual Li and P atoms, and the structure is optimized. We progressively increase the quantity of H2 molecules, and the structure is re-optimized after each H2 molecule addition (Fig. 5(d), Fig. 5(e), and Fig. 5(f)). It is found that the highest number of H2 adsorbed per Li and P atoms in the instance of (g-C3N4)8Li3P is three, as shown in Fig. 5(f); the fourth H2 is kept away from the monolayers. A reasonable fraction of the distance is preserved between the H2 molecules to bypass annoying repulsions.
The adsorption energy of the first H2 molecule is evaluated using Eq. 2 and is come up with − 0.131 eV considering GGA-vdW for (g-C3N4)8Li3P (Fig. 5(d)). For the second and the third H2 molecule, the adsorption energies are − 0.093 and − 0.089 eV, respectively, for (g-C3N4)8Li3P. Even more importantly, according to Table I, the adsorption energies are doubled upon, including dispersion corrections. Hence van der Waals forces play a significant role in the binding. Also, as given in Table 1, the adsorption energies reduce thru the rise of H2, suggesting that the system can adsorb an inadequate amount of H2. It has been noticed that the distance between two hydrogen atoms (H-H) is elongated than the isolated H2 (0.75 Å). This elongation is due to the polarization between the Li, P, and H2, as shown in Table 1.
Further, it is observed that the average Li-H2 and P-H2 distances increase with the intensification of adsorbed H2; however, the H-H bond length shrinkages. This decrease in the H-H bond span implies that the adsorption strength decreases with the intensification of the adsorbed H2. This is confirmed by the adsorption energies of different configurations, as given in Table 1. It is observed that the computed adsorption energy is diminished from − 0.131 eV for single adsorbed H2 to -0.065 eV per H2 for the six adsorbed H2 on the Li and P functionalized g-C3N4, which may be attributable to the steric repulsion between the adsorbed H2 molecules [14]. Even though the typical adsorption energy for each H2 decreases as the quantity of H2 increases, the adsorption energy for each H2 is typical for solid physisorption in the six H2 adsorbed Li and P functionalized g-C3N4 [12].
Table 1
The adsorption energy for each H2 (GGA, GGA + vdW), average Li-H2 distances, average P-H2 distances, and typical H-H bond lengths on Li and P decorated carbon nitride monolayers.
System | Total Adsorption energy [Eads (H2) in eV] | Adsorption energy per H2 [Eads /H2 in eV] | Li – H2 distance (Å) | P – H2 distance (Å) | H – H bond length (Å) |
GGA + vdW | GGA | GGA + vdW |
(g-C3N4)8Li3 + 3H2 | -0.363 | -0.059 | -0.121 | 2.07 | - | 0.756 |
(g-C3N4)8Li3 + 6H2 | -0.642 | -0.056 | -0.107 | 2.28 | - | 0.755 |
(g-C3N4)8Li3 + 9H2 | -0.945 | -0.045 | -0.105 | 2.66 | - | 0.752 |
(g-C3N4)8Li3P + 4H2 | -0.524 | -0.082 | -0.131 | 2.18 | 3.07 | 0.759 |
(g-C3N4)8Li3P + 8H2 | -0.744 | -0.043 | -0.093 | 2.22 | 3.08 | 0.757 |
(g-C3N4)8Li3P+12H2 | -1.068 | -0.037 | -0.089 | 2.27 | 3.20 | 0.754 |
(g-C3N4)8Li3P+16H2 | -2.128 | -0.080 | -0.133 | 2.30 | 3.34 | 0.754 |
(g-C3N4)8Li3P+20H2 | -1.740 | -0.043 | -0.087 | 2.32 | 3.42 | 0.752 |
(g-C3N4)8Li3P + 24H2 | -1.560 | -0.034 | -0.065 | 2.33 | 3.50 | 0.751 |
The graph of volumetric capacity vs. gravimetric capacity is shown in Fig. 7. The ultimate H2 storage capacity of (g-C3N4)8Li3 (Fig. 5(c)) and (g-C3N4)8Li3P (Fig. 5(f)) are found 2.34 (9H2) and 2.98 wt % (12H2) respectively, with the adsorption of H2 to just a single side of the monolayer. Finally, we have allowed the adsorption of H2 on both sides of the monolayers (Fig. 6). By the adsorption of H2 on both sides, we have obtained a gravimetric and volumetric capacity of 5.78 wt% hydrogen and 0.0275 Kg hydrogen/L, respectively, for the (g-C3N4)8Li3P configuration (Fig. 7). These values are pretty close to the U.S. Department of Energy (DOE) [7].
3.3 Nature of interaction between H2 and Li and P functionalized g-C3N4 monolayer
To determine the nature of the Li and P functionalization binding on the g-C3N4 monolayer and the adsorption of hydrogen molecules on the Li and P functionalized g-C3N4 monolayer qualitatively, we have plotted the total electron density. Here is no electron density at the interface area between the Li atom and g-C3N4 monolayer and amid H2 and the Li atom (Fig. 8 (a) and (b)). These electron density plots specify that the Li functionalization on the g-C3N4 and the H2 adsorption on the Li functionalized g-C3N4 do not form a covalent bond along the g-C3N4, indicating the physical nature of the adsorption. The same type of bonding nature is also observed in the case of Li and P functionalization on the g-C3N4 monolayer, as shown in Fig. 8 (c) and (d), which is also confirmed previously by the adsorption energy calculation as given in Table 1.
The CDD plots of the systems, as mentioned earlier, are shown in Fig. 9. As can be seen from Fig. 9 (a) and Fig. 9 (c), few charges exit on the topmost of Li and P atoms, while most of the charges are mainly concentrated near the g-C3N4 monolayer, which indicates some charges transfer from Li and P atoms to the g-C3N4 monolayer. These partially charged Li (P) ions and the g-C3N4 monolayer would yield a local electric field. These local electric fields polarized the hydrogen molecules and bound them thru the polarization phenomenon [41]. Hence the H2 gets polarized with the electron density gathering on the side adjacent to the g-C3N4 monolayer and the depletion on the side, aside from the g-C3N4 monolayer, as shown in Fig. 9 (b) and Fig. 9 (d). This polarization is the reason behind the elongation of the H-H bond in the H2 adsorbed on the Li and P functionalized g-C3N4 compared to the H-H bond in the isolated H2. Hence, we can say that the H2 adsorption on the Li and P functionalized g-C3N4 features dipole-dipole interactions [14].
We have calculated the charge transfer by Bader charge analysis as given in Table 2 [42]. Individual Li drops nearly 0.90 e¯ to the monolayer in Li atoms added g-C3N4 monolayer, while each Li atom loses almost 0.90 e¯ and the P atom loses approximately 0.80 e¯ in case of Li and P atoms added g-C3N4 monolayer. This charge transfer suggests that the bonding between Li and g-C3N4 monolayer and between P and g-C3N4 monolayer is ionic, proved previously by the ELF plots as shown in Fig. S2. These + ve charged Li and P ions yield local electric fields that polarize H2 molecules and thus enhance the adsorption [43,44]. Hence, finding a system in which the metal ion is remained positively charged is the key to molecular hydrogen adsorption [45]. Since no charge transfer takes place in the case of the physisorption mechanism, the amount of hydrogen that can be stored on the Li and P functionalized g-C3N4 monolayer is limited mainly by steric hindrance.
Table 2
Bader charges of C and N atoms before and after Li and P functionalization and H2 adsorption.
Configuration | Bader charges (e−) |
Before (Li/P) functionalization | After (Li/P) functionalization | After H2 adsorption |
C | N | Li | P | C | N | Li | P | C | N | Li | P | H |
(g-C3N4)8Li3 | 1.56 | -1.17 | - | - | 1.45 | -1.16 | 0.80 | - | 1.44 | -1.16 | 0.79 | - | -0.003 |
(g-C3N4)8Li3P | 1.56 | -1.17 | - | - | 1.30 | -1.17 | 0.80 | 0.90 | 1.31 | -1.17 | 0.89 | 0.79 | -0.004 |
To understand the nature of the H2 adsorption on the Li and P functionalized g-C3N4, we have plotted the densities of states for the hydrogen molecule, Li atoms, P atom, and the g-C3N4 of the H2 adsorbed Li and P functionalized g-C3N4. The Fermi level is fixed to zero in these plots. As shown in Fig. 6(d), there is no apparent hybridization between the Li and P functionalized g-C3N4 monolayer and the adsorbed hydrogen molecule. Therefore, the hydrogen adsorption mechanism in Li and P functionalized g-C3N4 monolayer is reasonably different from the Kubas interaction, which generally occurs in transition metal-doped materials [46]. In Kubas interaction, stronger hybridization exists between the d orbital of the functionalized transition metal and the adsorbed hydrogen σ orbital.