CLiAl2E and CLi2AlE (E = P, As, Sb, Bi): Planar Tetracoordinate Carbon Clusters with 16 and 14 Valence Electrons.

The strategy to remove the lone pairs of ligands combined with the bonding similarity between Li and Al have been utilized to design new planar tetracoordinate carbon (ptC) species C 2v CLiAl2E and CLi2AlE based on ptC global minima CAl3E (E = P, As, Sb, Bi) clusters. The explorations of potential energy surfaces and high-level CCSD(T) calculations indicate that these planar tetracoordinate carbon (ptC) species with 16 and 14 valence electrons (ve) are the global minima except for CLiAl2P. Bonding analyses reveal that there is one π and three σ bonds between C and ligands, one delocalized σ bond between the peripheral ligands, and three/two lone pairs for CLiAl2E and CLi2AlE (E = P, As, Sb, Bi). Especially, the C=E double bonds are crucial for the stabilities of these ptC clusters. The ptC core is governed by 2π + 6σ bonding, which conforms to the 8-electron counting. Born-Oppenheimer molecular dynamics (BOMD) simulations reveal that CLiAl2E and CLi2AlE (E = P, As, Sb, Bi) clusters are robust against isomerization and decomposition. The results obtained in this work complete the series of ptC CLi n Al3-n E (E = P, As, Sb, Bi; n = 0-3) systems and 18ve, 16ve, 14ve, and 12ve counting.

Introduction

Planar tetracoordinate carbon (ptC) structure was first put forward in a transition state in 1968 by Monkhrost.[1] Two years later, based on analyzing the electronic structure of planar methane, Hoffmann and co-workers creatively proposed strategies to stabilize ptC structure.[2] In 1976, Schleyer and Pople computationally characterized the first ptC molecule 1,1-dilithiumcyclopropane (C3H3Li2).[3] Since then, considerable and continuous efforts have been devoted to design, prediction, and production of planar tetra-, penta-, hexa-, and heptacoordinate carbon (ptC, ppC, phC, and p7C) species[4−9] for decades, due to their unusual structures, peculiar bonding properties, and potential application prospect. Several ptC clusters including NaAl4C–, CAl3Si–, CAl3Ge–, and CAl4H– have been observed in the gas photoelectron spectroscopy (PES) experiments,[10−12] which are isoelectronic with previously reported cis-CSi2Al2 and trans-CSi2Al2 by Schleyer and Boldyrev.

Two strategies including “electronic” and “mechanical”, can be used to stabilize the ptC species. The mechanical approach employs the solely steric forces (small-ring strain and/or an annulene perimeter or cylindrical cage or tube) to stabilize the ptCs.[2,4−8] Compared with mechanical strategy, electronic strategy has the merit of being uncomplicated and easily applicable. Based on CSi2Al2, CSi2Ga2, and CGe2Al2, Schleyer, Boldyrev, and Simons proposed that the presence of 18 valence electrons (ve) is crucial for favoring ptC over corresponding tetrahedral structures.[13,14] As an important rule not a law, 18ve counting is a nice building way to suggest the new ptC/ppC species, both in computational designs and gas-phase experimental works, as exemplified by a series of CAl42–,[10−16] CAl5+,[17−21] and CBe5[4−25] based clusters. As proposed by Keese,[6] the electron counting rule for planar hypercoordinate carbons may depend strongly on the chemical surrounding. Actually, to date, ptC structures with fewer than 18 valence electrons are rarely studied. It should be noted that ptC CAl4– and CAl3E (E = Si, Ge) with 17ve were observed in photoelectron spectroscopy experiment by Wang as early as 1999.[11,26] In 2018, 16ve CBe4Au4 with a quasi-ptC was theoretically predicted by us.[27] Using the ionic strategy, Ding and co-workers designed the first pentaatomic ptC species CAl3Tl with 16ve.[28] 15ve ptC-CB4+ was also predicted as the global minimum.[29] 14ve ptC species CSi2H2 and CGe2H2 were only the local minima on the potential energy surfaces.[30] In 2015, 12ve CLi3E (E = P, As, Sb, Bi) and CLi3E+ (E = O, S, Se, Te, Po) were reported as the ptC species.[31] These findings continue to challenge the 18ve counting of ptC species. Thus, 18ve counting is by no means a prerequisite for the ptC species.

Usually, the electrons of ptC species are used to form the C–ligand (C–L) bonding, ligand–ligand (L–L) bonding, and the lone pairs (LPs) of ligands. The LPs of ligands seem less important for the stability of ptC, compared with the first two kinds of bonding. Some of LPs may be removed to change the electrons of the whole ptC system. Can we design the new ptC species by removing the LPs of ligands? The answer seems to be yes. Based on the similarity between Li and Al, we herein design and characterize the ptC species C2 CLiAl2E and CLi2AlE (E = P, As, Sb, Bi) with 16ve and 14ve, respectively. To our knowledge, CLiAl2E are the third series of 16ve ptC GMs, while CLi2AlE (E = P, As, Sb, Bi) are the first series of ptC GMs with 14ve. The current results complete the series of ptC CLiAl3–E (E = P, As, Sb, Bi; n = 0–3) systems with 18ve, 16ve, 14ve, and 12ve counting.

Computational Methods

Global minimum (GM) structural searches were performed for CLiAl2E and CLi2AlE (E = P, As, Sb and Bi) clusters, using the coalescence kick (CK)[32−34] algorithm at the B3LYP[35,36] /Lanl2DZ level. More than 2000 stationary points were probed for each cluster (1000 singlets vs 1000 triplets). The top low-lying isomers were then fully optimized using the B3LYP approach and the aug-cc-pVTZ basis set (aug-cc-pVTZ-pp for Sb and Bi).[37,38] Vibrational frequencies were calculated at the same level to ensure that the reported structures are true minima on the potential energy surfaces. Single point CCSD(T)[39,40] calculations were performed for the global minima and three low-lying isomers of CLi2AlE and CLiAl2E (E = P, As, Sb, Bi) cluster geometries to further evaluate their relative energies. Natural bond orbital (NBO)[35] analyses were performed at B3LYP/C, Li, Al, P, As/aug-cc-pVTZ/Sb, and Bi/aug-cc-pVTZ-pp level to obtain Wiberg bond indices (WBIs) and natural atomic charges. Chemical bonding was elucidated via canonical molecular orbitals (CMOs) and adaptive natural density partitioning (AdNDP).[41] Nucleus independent chemical shifts[42] and isochemical shielding surfaces[43] in the Z direction (NICS and ICSS) were calculated to assess π/σ aromaticity for these ptC species. Orbital compositions were analyzed using the Multiwfn program.[44] All electronic structure calculations were done using the Gaussian 09 package.[45] Molecular structures, CMOs, isochemical shielding surface (ICSS), and AdNDP bonding patterns were visualized using the CYLview,[46] Multiwfn,[44] and GaussView 5.0[47] programs.

Results

Design of CLiAl2E and CLi2AlE (E = P, As, Sb, Bi)

Scheme shows that the strategy to design CLiAl2E and CLi2AlE (E = P, As, Sb, Bi). CAl42– (D4) has been regarded as the prototype of 18ve ptC. Based on the isoelectronic relationship between Al2– and E, 18ve ptC-CAl3E (E = P, As, Sb, Bi) were designed by Ding and Merino in 2015, which are the global minima on the potential energy surfaces. Using three Li atoms to replace three Al atoms of 18ve CAl3E, 12ve ptC-CLi3E (E = P, As, Sb, Bi) can be formed. The structures of CAl3E (18ve) and CLi3E (12ve) are similar, with the central ptC and three periphery Al/Li. In other words, each Al atom in CAl3E can be replaced by Li atom. In fact, only one electron of Al atom is used for the bonding in these species, because two valence electrons are used to form the LP. Thus, there is little effect on the stability to remove the LPs of metal ligands for these ptC clusters. Replacing one/two Al atom(s) with Li in 18ve ptC-CAl3E, the 16ve ptC-CLi2AlE and 14ve ptC-CLi2AlE (E = P, As, Sb, Bi) can be designed. Thus, 18ve ptC-CAl3E, 16ve ptC-CLiAl2E, 14ve CLi2AlE, and 12ve CLi3E (E = P, As, Sb, Bi) formed a perfect ptC-CLiAl3–E (E = P, As, Sb, Bi; n = 0–3) systems. There are two types of Al atoms in CAl3E: one is bridging Al (adjacent with E atom) and the other is terminal Al. If one Al atom of CAl3E is being replaced with Li atom, our calculations indicate that the terminal Al atom prefers to be replaced (except for E = P). In contrast, if two Al atoms of CAl3E are being replaced, the bridging Al atoms possess high preference. However, there are still some questions to be answered. Are 16ve ptC-CLi2AlE and 14ve ptC-CLi2AlE (E = P, As, Sb, Bi) global minima on the potential energy surfaces? Are they dynamically stable? What are their bonding properties? Are they aromatic? We will answer these questions in the following discussion.

Scheme 1

Basic Idea of Designing 16ve-CLiAl2E and 14ve-CLi2AlE (E = P, As, Sb, Bi)

Planar Tetracoordinate Cabon (ptC) CLiAl2E and CLi2AlE (E = P, As, Sb, Bi) Clusters

The extensive computational structural searches of the potential energy surface and high-level ab initio calculations suggest that each global minimum of CLiAl2E and CLi2AlE (E = P, As, Sb, Bi) has a ptC in the center. The optimized structures of GMs 1–8 (except for 1) along with their bond distances are illustrated in Figure . Three low-lying isomers are also depictured in Figure S1 (Supporting Information) for comparison. As shown in Figures and S1, 1–8 possess the perfectly C2 planar structures, with the central ptC and two periphery symmetric Al/Li ligands. 16ve ptC-GMs 2–4 are 0.74, 2.91, and 3.39 kcal·mol–1 more stable than their closest competitors at the single point CCSD(T) level, while 1 is the second lowest structure, being 0.15 kcal·mol–1 higher in energy than 1b. The relative energy (ΔE) of bridging Li vs terminal Li structures of the CLiAl2E (E = P, As, Sb, Bi) series is shown in Figure . From P to Bi, the geometric radius increases gradually in the order 1.07 < 1.19 < 1.39 < 1.48 Å,[48] while the electronegativity decreases in the order 2.19 > 2.18 > 2.05 > 2.02. The relatively small geometric radius and large electronegativity make P atom possible to share Li with the C atom. 5–8 are the GMs of the CLi2AlE (E = P, As, Sb, Bi) with 14ve, which are 4.91, 6.52, 6.26, and 5.90 kcal·mol–1 more stable than their second isomers at the single point CCSD(T) level. It should be noted that the second lowest isomers of CLiAl2E and CLi2AlE (E = P, As, Sb, Bi) can be obtained by exchanging positions of Al and Li in 1–8, which also are ptC species with C symmetry.

Figure 1

Optimized ptC global minimum structures 1–8 of C2 CLiAl2E and CLi2AlE (E = P, As, Sb, Bi) at B3LYP/C, Li, Al, P, As/aug-cc-pVTZ/Sb, Bi/aug-cc-pVTZ-pp levels. Bond distances are in angstroms.

Figure 2

Relative energy (ΔE) of bridging Li vs terminal Li structures of the CLiAl2E (E = P, As, Sb, Bi) series at the single point CCSD(T) level, with zero point energy (ZEP) corrections at B3LYP level. The CAl2LiE with terminal Li is used as a reference.

Optimized ptC globpan class="Chemical">al minimum structures 1–8 of C2 CLiAl2E and CLi2AlE (E = P, As, Sb, Bi) at B3LYP/C, Li, Al, P, As/aug-cc-pVTZ/Sb, Bi/aug-cc-pVTZ-pp levels. Bond distances are in angstroms.

Relative energy (ΔE) of bridging Li vs terminpan class="Chemical">al Li structures of the CLiAl2E (E = P, As, Sb, Bi) series at the single point CCSD(T) level, with zero point energy (ZEP) corrections at B3LYP level. The CAl2LiE with terminal Li is used as a reference.

The bond distances, Wiberg bond indices (WBIs), and natural atomic charges can help us to explore the bonding characters of 1–8, which are shown in Figure and Table . The stabilities of ptC benefit from C–ligand and ligand–ligand bonding. The C–ligand bonding includes C–E, C–Al, and C–Li bonding, while ligand–ligand bonding includes E–Al, E–Li, and Al–Li bonding in 1–8. With the increase in the size of E atoms, the changes of the corresponding bond distances have the almost same trends. The C–E bond distances in 1–8, in turn, are 1.71, 1.85, 2.07, 2.18, 1.69, 1.82, 2.05, and 2.16 Å, which are close to the C=E double bond lengths (1.69, 1.82, 2.04, and 2.15 Å for E = P, As, Sb, Bi).[49] The WBIC–E are in the range of 1.54–1.92 in 1–8, which is consistent with the conclusions from the analyses of bond distances. Thus, there is the C=E double bond in 1–8, respectively. It should be noted that the C–E bonding gradually decreased, with the electronegativity of E atom gradually increasing in these ptC species. For reference, a single C–Al bond has an upper bound of 2.01 Å and that of C–Li is 2.08 Å, based on the covalent atomic radii. The C–Al bond distances are in the range of 2.01–2.06 Å in 1–4, being close to the single bond length. The C–Al bonding in 5–8 are strong than that of 1–4, as the C–Al bond distances (from 1.90 to 1.94 Å). The WBIC–Al are in the range of 0.42–0.49 in 1–4, indicating that the C–Al bonding is close to half of the covalent single bond, while that of 5–8 are from 0.73 to 0.79, being close to a covalent single bond. The C–Li bond distances are from 1.90 to 2.00 Å in 1–8. There is a little covalent bonding between C and Li in 1–8, due to the large differences in their electronegativity. The E–Al bond distances are from 2.47 to 2.89 Å in 1–4, while the E–Li bond distances are from 2.33 to 2.70 Å in 5–8. The WBIE–Al in 1–4 and WBIE–Li in 5–8 are in the range of 0.43–0.47 and 0.30–0.33, respectively, suggesting that there are partial covalent bonding between them. The long distances between terminal Al/Li atoms and the bridging Li/Al, indicate that there are weak bonding between them. The WBIAl–Li(terminal) of 1–4 are only 0.10–0.16 and WBILi–Al (terminal) of 5–8 are only 0.09–0.10, respectively.

Table 1

Lowest Frequencies (vmin, cm–1), Natural Atomic Charges (|e|), Wiberg Bond Indices, and Highest Occupied Molecular Orbital–Lowest Unoccupied Molecular Orbital (HOMO–LUMO) Gap (Gap/eV) of 1–8 at B3LYP Level

cluster	vmin	q(C)	q(E)	q(Al)	q(Li)	WBIC–E	WBIC–Li	WBIC–Al	WBIE–Al/Li	WBIAl–Li	gap
1	55	–1.94	0.03	0.57	0.80	1.84	0.08	0.41	0.43	0.10	2.62
2	55	–2.05	0.11	0.57	0.79	1.76	0.08	0.42	0.44	0.12	2.32
3	85	–2.22	0.35	0.55	0.77	1.59	0.08	0.47	0.47	0.15	1.91
4	82	–2.23	0.38	0.54	0.76	1.54	0.08	0.49	0.47	0.16	1.74
5	118	–1.83	–0.05	0.50	0.69	1.92	0.15	0.73	0.30	0.10	2.10
6	94	–1.91	–0.01	0.54	0.69	1.84	0.15	0.74	0.31	0.09	2.04
7	70	–2.09	0.13	0.59	0.68	1.69	0.14	0.76	0.33	0.10	1.87
8	59	–2.10	0.13	0.60	0.69	1.64	0.15	0.79	0.33	0.09	1.80

The mixed covalent/ionic nature of C–Al, C–E, C–Li, E–Al, E–Li, and Al–Li bonding, as also reflected from the NBO charges. As shown in Table , the central ptCs in 1–8 carry the negative charges (−1.83 to −2.10 |e|). There are obvious electron transfer from the Al and Li atoms to C. E atoms carry a small amount of positive charges in 1–4 (0.03–0.38 |e|) and 7, 8 (0.13 |e|), while that of 5 and 6 carry a little negative charges (−0.01 and −0.05 |e|). Al atoms carry 0.50–0.60 |e| positive charges, while Li atoms carry 0.68–0.80 |e| positive charges. The positive charges of terminal Al/Li are more than that of the bridging Al/Li atoms in these ptC species. These C–ligand and ligand–ligand bonding together contribute to the stabilities of these ptC species. In addition, the relatively large HOMO–LUMO (gaps) (2.62, 2.32, 1.91, and 1.74 eV) further support the stabilities of 1–8. Notably, it decreases gradually from E = P to Bi, implying a decrease in stability according to the principle of maximum hardness.

Molecular Dynamics

The kinetic stability is equally important for the experimental realization. Born–Oppenheimer molecular dynamics (BOMD) simulations were performed at B3LYP/C, Al, Li, P, As/6-31G(d)/Sb, and Bi/Lanl2DZ level for 50 ps at room temperature (298 K) to investigate the dynamic stabilities of 1–8. The kinetic stability can be evaluated by examining the structural evolution during the simulation, which is measured by the root-mean-square deviation (RMSD). As shown in Figure S2, the average RMSDs of 1–8 are in the range of 0.15–0.22 Å, indicating they possess good kinetic stabilities.

Discussion

Chemical Bonding in CLiAl2E and CLi2AlE (E = P, As, Sb, Bi) Clusters

Why are these 16ve and 14ve ptC clusters stable? To understand the bonding nature of CLiAl2E and CLi2AlE (E = P, As, Sb, Bi), molecular orbital analyses were performed, aided with orbital composition calculations. The valence canonical molecular orbitals (CMOs) of 1 and 5 are plotted in Figure . The valence CMOs of 2, 3, 4 and 6, 7, 8 have the same patterns with 1 and 5, respectively. For clarity, here, only the valence CMOs of 1 and 5 will be discussed. As shown in Figure a, eight occupied CMOs of 1 can be divided into four subsystems: (i) Subset 1 has three CMOs: HOMO, HOMO – 2, and HOMO – 6, which are largely Al/P s based and can be readily located as three 1c–2e lone pairs; (ii) Subset 2 is a π framework and involve only the HOMO – 3, which is primarily on the CAl2P core; (iii) Subset 3 is a σ sextet (including HOMO – 4, HOMO – 5, and HOMO – 7), which is delocalized with components from C center; (iv) Subset 4 is σ delocalized subsystem (including only the HOMO – 1), which is primarily on the peripheral LiAl2P ring (the contribution of C in HOMO – 1 is only 9%). One π and three σ bonds around ptC center, in combination with one periphery Al–P–Al–Li delocalized bonds can make ptC CLiAl2P stable. The CMOs of 5 are basically similar to 1, except for one LP of Al, which also has three π and σ subsystems. It has little impact on the bonding properties to replace one Al atom with one Li atom in 16ve CLiAl2P. Thus, both 16ve CLiAl2P and 14ve CLi2AlP have three π and σ subsystems (2π, 6σ, and 2σ). If these bondings are all delocalized, the 2π, 6σ, and 2σ electron counting conform to the Hückel (4n + 2) aromatic rule, where n = 0, 1, 0, respectively. The π-bonding orbital (HOMO – 3 in 1, HOMO – 2 in 5) is predominantly located on the C–P bond, the contributions of Al and Li are only 5 and 3%. Thus, the π bonds are basically localized in nature. The C–P σ bond is close to the single covalent bond, as the WBIC–P (1.87 in 1 and 1.92 in 5) indicated. Thus, there are C=E double bonds in ptC CLiAl2E and CLi2AlE (E = P, As, Sb, Bi) clusters, which have some negative impact on the electronic delocalization.

Figure 3

Canonical molecular orbitals (CMOs) of (a) CLiAl2P (1) and (b) CLi2AlP (5).

Canonical molecular orpan class="Chemical">bitals (CMOs) of (a) CLiAl2P (1) and (b) CLi2AlP (5).

To probe the π/σ aromaticity for these ptC species, NICS values in the Z direction were calculated on the concerned points, including the centers of the triangles and the points located at 1 Å above the centers of the triangles and above the ptCs, respectively. The NICS (1) and NICS (0) values for 1–8 clusters are depictured in Figure S3. Most of NICS (1) and NICS (0) with negative values, indicating each of 1–8 has a certain degree of π/σ aromaticity. According to the data, 1–4 are more aromatic than 5–8. However, evaluating NICS of several points seems to be inadequate. The magnetic criterion isochemical shielding surface (ICSS) calculation is handled in a three-dimensional grid of lattice points and direction and anisotropy effects can be quantified in a more straightforward way. Contrary to the NICS approach, positive ICSS values indicate diatropic ring currents and aromaticity, while negative values indicate paratropic ring currents and antiaromaticity. Here, we used ICSS (1) and ICSS (0), the shielding tensor component perpendicular to the studied molecular planes at 1 and 0 Å above them, to characterize π and σ aromaticity of representative clusters 1 and 5. To more intuitively observe the aromaticity, the color-filled maps of ICSS (1) and ICSS (0) of 1 and 5 are plotted in Table S4, compared with 18ve-CAl3P and 12ve-CLi3P. As shown in Table S4, going down from 18ve-CAl3P to 12ve-CLi3P, both π and σ electrons delocalization decrease. Al atom can partly participate in the π and σ electrons delocalization, while the Li atom has little contribution to the electron delocalization. Al atoms have more contribution to the σ electrons delocalization in comparison with the π electrons delocalization. The π electrons mainly located around the C/P, and a few of them delocalized. In 18ve-CAl3P, the π electrons are a little delocalized, while the σ electrons are fully delocalized. Replacing all Al with Li atoms, the 12ve-CLi3P formed, whose π and σ electrons are localized. The π and σ electrons delocalization of 16ve-CLiAl2P and 14ve-CLi2AlP are intermediate between CAl3P and CLi3P. The π and σ electrons delocalization are beneficial but not required for these ptC species. Thus, these ptC 1–8 clusters are stable, benefiting from one π and three σ bonds around ptC, and one periphery L–L σ bond.

Electron Counting to Stabilize ptC Species vs the Unnecessary Lone Pairs of Ligands

As we all know, 18ve counting is originally proposed to stabilize the ptC, which refers to the total valence of a ptC system including bonding electrons and nonbonding electrons. The ptC is always stabilized in a specific chemical environment, which is dominated by the ligands. The C–ligand and ligand–ligand bondings are important for the stabilities of ptC species, while some LPs of ligands seem to be unnecessary. Here, we take 18ve-CAl3P, 16ve-CLiAl2P, 14ve-CLi2AlP, and 12ve-CLi3P as the examples to discuss the relations between the electron counting and the unnecessary lone pairs of ligands.

The AdNDP bonding patterns of 18ve-CAl3P, 16ve-CLiAl2P, 14ve-CLi2AlP, and 12ve-CLi3P are depictured in Figure . We analyze 16ve-CLiAl2P first. Each Al/P atom has one LP, thus, there are three LPs. There is one delocalized σ bond on the periphery LiAlPAl ring. The residual eight electrons can correspond to C–L bonding: one 2c–2e C–P π bond, one C–P σ bond, and two delocalized 3c–2e (Al–C–Al) σ bonds. All ON values are close to the expected 2.00 |e|. The AdNDP conclusion accords with the above CMO analyses. Although there are some differences in compositions, geometries, and numbers of valence electrons, these ptC species have two aspects in common: their ptC core is held together by the same 2π/6σ frameworks, which collectively follow the octet rule; in addition, there is one 4c–2e delocalized σ bond in each ptC species. Thus, the 2π/6σ frameworks and peripheral 4c–2e L–L σ bonds are crucial for the stabilities of these ptC species. Going down from 18ve-CAl3P to 12ve-CLi3P, the peripheral lone pairs (LPs) are diminishing in number. As the first column shown, the LPs of Al atoms gradually decrease from three to zero from top to bottom, while the LP of P is still intact. Thus, the LPs of Al atoms seem to be unnecessary. In these ptC species, only one electron of each Al atom is used to form chemical bonding with its neighbors, except for an LP. As shown in Figure , the electron localization function (ELF) color-filled map for (a) CAl3P, (b) CLiAl2P, (c) CLi2AlP, and (d) CLi3P further confirmed the above conclusion. The good stabilities of them suggest that the LPs of Al atoms in these species are less important, even unnecessary and can be removed. Thus, removing the unnecessary lone pairs of ligands should be a new strategy, which can help us design more new ptC species with the variable electron counting.

Figure 4

AdNDP bonding patterns of 16ve-CLiAl2P (1) and 14ve-CLi2AlP (5) in comparison with 18ve-CAl3P and 12ve-CAl3P, with the occupation numbers (ONs).

Figure 5

Electron localization function (ELF) color-filled map for (a) CAl3P, (b) CLiAl2P, (c) CLi2AlP, and (d) CLi3P.

AdNDP bonding patpan class="Chemical">terns of 16ve-CLiAl2P (1) and 14ve-CLi2AlP (5) in comparison with 18ve-CAl3P and 12ve-CAl3P, with the occupation numbers (ONs).

Electron localization function (ELF) color-filled map for (a) pan class="Chemical">CAl3P, (b) CLiAl2P, (c) CLi2AlP, and (d) CLi3P.

Necessary LP of E (E = P, As, Sb, Bi)

To design the new ptCs by our strategy, it is needed to distinguish the unnecessary and necessary LPs in the ptC clusters. There are also some necessary LPs, which are important for the stability of ptCs. Usually, these necessary LPs come from the nonmetal atom with the relatively large electronegativity. There is one LP of E in the 18ve-CAl3P, 16ve-CLiAl2P, 14ve-CLi2AlP, and 12ve-CLi3P clusters, which is necessary to stabilize the ptC and cannot be removed. There are strong competitions between ptC, tetrahedral carbon (thC), and tri coordinate carbon (trC) isomers. Using the Al atom to replace the P atom in ptC CAl3P, CLiAl2P, CLi2AlP, and CLi3P clusters, we got the 16ve-CAl4, 14ve-CLiAl2P, 12ve-CLi2AlP, and 10ve-CLi3Al, which prefer to possess the thC or trC structures. Thus, these LPs are necessary for maintaining the stabilities of ptCs. Why are they important? There are electrostatic repulsion among the four positively charged Al ligands, which make that thC is more favored than ptC for 16ve-CAl4. The LP of P atom in 18ve-CAl3P, can help to attract the two adjacent Al atoms due to electrostatic interactions between the LP and the ligands. In addition, three 3p electrons of P can contribute to form stronger bonding with the central ptC and other ligands. Thus, the LP of E is important for the stability of 18ve-CAl3P, 16ve-CLiAl2P, 14ve-CLi2AlP, and 12ve-CLi3P clusters.

Summary

In summary, we have certificated that it is possible to replace the Al atoms in previously reported ptC-containing global minima C2 18ve-CAl3E (E = P, As, Sb, Bi) with the Li to design the new ptC species CLiAl2E (E = P, As, Sb, Bi) and CLi2AlE (E = P, As, Sb, Bi). These 16ve and 14ve ptC species are the global minima (except for CLiAL2P) with the peculiar C=E double bonds. In nature, it is the new strategy that help us to the design these ptC species, which is removing the unnecessary lone pairs of ligands. For a ptC cluster, one π and three σ bonds for the C–ligand bonding, and one σ bond for the ligand–ligand bonding are important. The 18ve counting and fully delocalized (2π and 6σ) are not required.