On the Nature of Bonding in Synthetic Charged Molecular Alloy

Sep 26, 2018 - However, their chemical bonding has seldom been elucidated using modern quantum chemistry tools. Herein, we report on in-depth chemical...
0 downloads 0 Views 5MB Size
Download PDF
This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Article Cite This: ACS Omega 2018, 3, 11958−11965

http://pubs.acs.org/journal/acsodf

On the Nature of Bonding in Synthetic Charged Molecular Alloy [P7ZnP7]4− Cluster and Its Relevant [P7]3− Zintl Ion Xue-Rui You and Hua-Jin Zhai* Nanocluster Laboratory, Institute of Molecular Science, Shanxi University, Taiyuan 030006, China

Downloaded via 5.188.216.218 on October 2, 2018 at 13:34:20 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Charged molecular alloys and Zintl ions are of interest in synthetic chemistry. However, their chemical bonding has seldom been elucidated using modern quantum chemistry tools. Herein, we report on in-depth chemical bonding analyses for a charged molecular alloy C2 [P7ZnP7]4− cluster and its relevant Zintl ion C3v [P7]3− ligand, making use of electronic structure calculations at PBE0/def2-TZVP level, natural bond orbital and orbital composition analyses, canonical molecular orbitals, and adaptive natural density partitioning (AdNDP). The computational data show that C3v [P7]3− Zintl ion has three isolated, negatively charged, bridging P sites. Such charges are largely P 3p lone-pairs in nature, but they also participate in secondary P−P bonding along the bridging sites. C2 [P7ZnP7]4− cluster is formulated as [P7]2−[Zn]0[P7]2−, in which [P7]2− ligands maintain the structural and bonding integrity of [P7]3− Zintl ion despite their difference in charge state. Two [P7]2− ligands collectively bind with Zn center via four bridging P sites, resulting in a quasi-tetrahedral ZnP4 core with the eight-electron counting. This bonding picture can alternatively be rationalized using the superatom concept. The Zn−P bonds are weak with a bond order of around 0.5, because the P centers have partial nonbonding 3p character, akin to 3p2 lone-pairs albeit with a lower occupation number.

1. INTRODUCTION During the past decades, synthetic chemistry has witnessed fantastic and fast-growing developments in Zintl ions, endohedral Zintl ions, cage compounds, intermetalloid clusters, and charged molecular alloys.1−9 These compounds have unique molecular structures, intriguing (and yet largely unknown) chemical bonding, and unusual chemical and physical properties, with promising applications in materials science, catalysis, and so forth. Among them, the so-called “charged molecular alloys” are interesting, which use Zintl ions or relevant post-transition metal or metalloid species as ligands to construct compounds with a metallic core.4−6 These alloy clusters normally assume elongated dumbbell or “polymeric” chain geometries, in which the Zintl ion ligands maintain their chemical integrity. However, chemical bonding in such compounds has seldom been elucidated in full detail, mainly because the systems are quite large and complex so that analyses of bonding via canonical molecular orbitals (CMOs) are challenging. Indeed, this situation appears to be true even for naked Zintl ions. Group 15 elements form stable homoatomic cage-like Zintl ions, including heptapnictide trianions [E7]3− (E = P, As, Sb),10 which have rich chemistry as reported in the earliest and most groundbreaking reactivity studies on Zintl ions.5 The [E7]3− cages have nortricylane-like structures and are obtained from dissolution of A3E7 alloys (A = alkali metal; E = P, As, Sb) in polar nonprotic solvents (ethylenediamine or liquid ammonia).1 A series of charged molecular alloys, or binary intermetalloid clusters, were synthesized with intact [E7]3− ligands. These compounds can be sorted to two categories. In © 2018 American Chemical Society

category one, transition-metal atoms serve as links that connect [E7]3− anions. Here, the metal core is believed to be in the form of an M2+ center or a Cu22+ and Au22+ dumbbell, such as in [(P7)Zn(P7)]4−, [(P7)Cd(P7)]4−, [(P7)Cu2(P7)]4−,11 and [(As7)Au2(As7)]4− compounds.12 Alternatively, the compounds may be viewed as sandwich-type clusters. In the gas phase, sandwich hetero-decked clusters were computed extensively by Yang and co-workers, which follow a versatile scheme for assembly and stabilization.13−26 The other category is that the metal core fuses with [E7]3− ligands, leading to breakage of one E−E bond in each [E7]3− ligand. Examples are [(As7)Pd2(As7)]4− and [(Bi7)Ge4(Bi7)]4−.27,28 It should be noted that the exact nature of bonding between metal centers and Zintl ion ligands in these compounds remains to be understood, in light of limited knowledge in the literature. In this contribution, we aim at understanding chemical bonding of a synthetic charged molecular alloy cluster, C2 [P7ZnP7]4−, using quantum chemical calculations. This cluster was crystallized in the form of [K(2,2,2-crypt)]4[ZnP14]·6py (py = pyridine) in 2010 by Goicoechea and co-workers,11 which consists of a bridging Zn center and two cage-like [P7] ligands, assuming an overall dumbbell shape. The cluster is presumably formulated as [P7]3−[Zn]2+[P7]3−. However, there has been no computational study on this system so far, to our knowledge. Our geometric optimizations and electronic structure calculations were carried out using density functional Received: July 27, 2018 Accepted: September 10, 2018 Published: September 26, 2018 11958

DOI: 10.1021/acsomega.8b01790 ACS Omega 2018, 3, 11958−11965

ACS Omega

Article

theory (DFT)29 at the PBE0 level.30 To further ensure computational reliability for a multiply charged anion, conductor-like polarizable continuum mode (C-PCM) calculations31−34 were performed to take into account solvation effects. Chemical bonding was elucidated using an array of technical tools, including in particular the CMO (that is, Kohn-Sham orbital) analysis and adaptive natural density partitioning (AdNDP).35 AdNDP turns out to be a powerful method to reach an elegant bonding picture for C2 [P7ZnP7]4− cluster, as well as related naked C3v [P7]3− Zintl ion. The bonding picture of the latter is refined with respect to existing knowledge. It is shown herein that C2 [P7ZnP7]4− cluster features an eight-electron quasi-tetrahedral ZnP4 core, which holds together the Zn center and two P7 cage ligands with a formula of [P7]2−[Zn]0[P7]2−. The two cage-like ligands inherit the structure and bonding of C3v [P7]3− Zintl ion despite the difference in charge state. Overall, the eight-electron bonding picture of C2 [P7ZnP7]4− cluster renders it a “superatom”,36−38 which underlies its stability as a synthetic charged molecular alloy.11 Furthermore, the Zn−P bonding is far weaker than anticipated for an eight-electron tetrahedral system, with a calculated bond order of around 0.5. This observation is ascribed to the polar nature of the ZnP4 core, in which P centers collectively contribute six electrons for Zn−P bonds with Zn contributing the remaining two. In such a case, the P centers maintain partial nonbonding 3p character and only a portion of around two electrons (out of six) from four P centers actually participate in the Zn−P bonds. This bonding situation is intriguing.

dumbbell shape. At the core of the cluster is a ZnP4 structural block. Approximately, the Zn center coordinates to four P atoms in a tetrahedral fashion: two P atoms from each [P7] cage. Herein, cage-like [P7] ligands possess similar structural parameters with respect to naked [P7]3− Zintl ion,10 although the ligands in C2 [P7ZnP7]4− do not appear to be exactly in the charge state of [P7]3− (vide infra). Optimized C3v (1A1) structure of [P7]3− Zintl ion at PBE0/ def2-TZVP level is shown in Figure 2. The P sites are sorted

Figure 2. Structure of C3v (1A1) [P7]3− Zintl ion at PBE0/def2-TZVP level. Seven P atoms are sorted into three types: P(i) for one apex atom at the top, P(ii) for three bridging P atoms on the waist, and P(iii) for the basal three-membered ring. Bond distances (in Å; black color) and natural charges (in |e|; blue color) are shown.

into three subgroups according to their coordination environments. P(i) represents one apical P atom at top of the cage. Three P(ii) sites are located on the waist of the cage, whereas three P(iii) sites form the base triangle. Specifically, P(i) coordinates to three P(ii) atoms in a triangular pyramid; P(ii) sites are dicoordinated, each bridging P(i) and one P(iii) site; and P(iii) sites are tricoordinated. Overall, the [P7]3− Zintl cage has three exactly equivalent, distorted P(i)P(ii)2P(iii)2 pentagons and one equilateral P(iii)3 triangle. In dumbbell C2 [P7ZnP7]4− cluster (Figure 1), the [P7] cages maintain their characteristic pentagonal and triangular surfaces, for which the P10P12P14 and P11P13P15 triangles are the bases, respectively. The P−P distances in C2 [P7ZnP7]4− and C3v [P7]3− are 2.14−2.26 and 2.14−2.29 Å, respectively (Figure S2, Supporting Information). These numbers are relatively uniform. According to the recommended covalent radii by Pyykkö,40 typical single P−P and double PP bonds are around 2.22 and 2.04 Å, respectively. Therefore, all P−P links in C2 [P7ZnP7]4− cluster and C3v [P7]3− Zintl ion are roughly single bonds, making their bonding pictures rather simple. In contrast, the Zn−P links in C2 [P7ZnP7]4− cluster have calculated distances of 2.48 and 2.50 Å (Figure 1), which are to be compared with the Zn−P single bond of 2.29 Å,40,41 suggesting that the Zn−P links in [P7ZnP7]4− are far weaker than single bond. Note that the present Zn−P distances are also longer than the mean Zn−P distance (2.41 Å) from the Cambridge Structural Database.42 2.2. Wiberg Bond Indices and Natural Atomic Charges. The above bond assignments for C2 [P7ZnP7]4− cluster and C3v [P7]3− Zintl ion are confirmed by Wiberg bond indices (WBIs) from the natural bond orbital (NBO) analysis, which gives bond orders of 0.85−1.16 and 0.80−1.20, respectively, for the P−P links. These WBIs are close to single

2. RESULTS 2.1. Cluster Structure of [P7ZnP7]4−. We initially obtained the coordinates of [P7ZnP7]4− cluster from crystal data reported in the Goicoechea paper (Figure S1, Supporting Information).11 Full reoptimizations of the cluster were subsequently carried out at the PBE0/def2-TZVP30,39 level. The ultimate optimized structure has C2 (1A) symmetry, as illustrated in Figure 1. All calculated structural parameters of

Figure 1. Optimized structure of C2 (1A) [P7ZnP7]4− cluster using the C-PCM calculations at PBE0/def2-TZVP level. Selected bond distances (in Å; black color in boldface) are compared with experimental crystal data from ref 11 (black color, in parentheses). Also presented are WBIs (red color) and natural atomic charges (in | e|; blue color) via NBO analysis.

[P7ZnP7]4− cluster are in line with experimental data. For example, Zn−P bond distances are calculated to be 2.50 and 2.48 Å, which are virtually identical to crystal data of 2.50 versus 2.47 Å, thus validating the computational method. The C2 [P7ZnP7]4− cluster can be viewed as two [P7] cages interconnected via a Zn bridge, which features an overall 11959

DOI: 10.1021/acsomega.8b01790 ACS Omega 2018, 3, 11958−11965

ACS Omega

Article

Figure 3. Pictures of CMOs of C3v [P7]3− Zintl ion calculated at PBE0/def2-TZVP level. (a) P 3s2 lone-pairs. (b) CMOs that are approximated to three bridging P 3p2 lone-pairs; see Table 1 and the text for further details. (c) Two-center two-electron (2c-2e) Lewis P−P σ single bonds.

bonds, with the P(ii)−P(iii) bonds in [P7]3− Zintl ion being moderately beyond single bonds (WBIs: 1.20). For Zn−P links in C2 [P7ZnP7]4−, the calculated WBIs are 0.48/0.49 (Figure 1), which are indeed markedly smaller than one. As for the natural charges, bridging P(ii) sites in C3v [P7]3− Zintl ion carry the majority of negative charges (−0.64 |e| per atom). The P(i) site has the least of charges (−0.19 |e|), and the P(iii) sites each have −0.30 |e|. As a zero-order picture, one can state that the charges in [P7]3− Zintl ion are located on three bridging sites (64% in total). In other words, the Zintl ion has three P− bridging sites and hence three sites with 3p2 lone-pair characters. These largely localized charges are maintained in C2 [P7ZnP7]4− cluster, in which similar P centers each carry a charge of −0.51 to −0.58 |e| (Figure 1). In contrast, apical P site in C2 [P7ZnP7]4− has a charge of −0.11 | e| and the base P sites each carry −0.17/−0.18 |e|. Note that the Zn center is only moderately charged (+0.49 |e|; Figure 1).

bottom P3 fragment are in a combination of bonding and nonbonding/antibonding, which further recombine as three P 3s2 lone-pairs. Similarly, the four bonds in top P4 fragment can be localized as four P 3s2 lone-pairs, although the analysis is slightly more complex. Among these four bonds for the P4 fragment, the bottom one is completely bonding between all four P centers, the middle pair are nonbonding/antibonding between three bridging P sites, while the top one (that is, HOMO−9; Figure 3) is antibonding between the apical P atom and three bridging P atoms. The bottom and top bonds can recombine first, which recover a P 3s2 lone-pair for the apical atom, as well as a bond (formal and completely bonding) for three bridging P atoms. The latter bond then recombines with the middle pair to generate three P 3s2 lonepairs for the bridging sites. A set of nine CMOs composed of P 3p AOs are shown in Figure 3c. Among these, HOMO−7 and HOMO−3/ HOMO−3′ are constructive combination between top P4 and bottom P3 fragments, whereas HOMO−5 and HOMO− 4/HOMO−4′ are their destructive counterparts. They collectively recover three bonds for top P4 and three for bottom P3, each in a typical bonding/nonbonding/antibonding combination, which further recombine as two-center twoelectron (2c-2e) P−P σ bonds (six in total). The remaining CMOs, HOMO−2/HOMO−6/HOMO−6′, are responsible for three P−P σ bonds on the waist, which connect the P4 and P3 fragments. The above Lewis elements (lone-pairs and 2c-2e σ bonds) use 32 electrons. The three CMOs in Figure 3b appear to be tricky, which are mainly composed of 3p AOs from three P(ii) centers by 72− 79% (Table 1), oriented tangentially to the P(ii)P(ii)P(ii)

3. DISCUSSION 3.1. On the Nature of Bonding in C3v [P7]3− Zintl Ion. The [P7]3− Zintl ion is an important ligand in synthetic chemistry,5,10 which has gained accumulating knowledge in the literature. However, a detailed bonding analysis based on CMOs is nonexistent for the species. Given the importance of this ligand and the fact that relevant charged molecular alloys are well-defined molecular systems, we believe it is valuable to perform a thorough bonding analysis for C3v [P7]3− Zintl ion (Figure 2). The C3v [P7]3− Zintl ion is a 38-electron system with seven P centers and nine P−P links, whose occupied CMOs are depicted in Figure 3. Element P has an electronic configuration of 3s23p3, for which 3s2 electrons normally form lone-pairs. The set of CMOs derived from P 3s atomic orbitals (AOs) are shown in Figure 3a. These seven CMOs represent constructive/destructive combinations between top (triangular pyramid) P4 and bottom (triangular) P3 segments. Specifically, HOMO−12/HOMO−11, HOMO−10′/HOMO−8, and HOMO−10/HOMO−8′ are constructive/destructive pairs, which follow the building principles of CMOs43 and collectively recover three bonds for the top P4, as well as three for bottom P3. In addition, HOMO−9 is attributed to top P4 fragment, being completely antibonding between top P atom and three bridging P sites. The resultant three bonds for

Table 1. Composition Analysis for Selected CMOs of C3v [P7]3− Zintl Ion at PBE0/def2-TZVP Level CMOa HOMO HOMO−1 HOMO−1′

P(i) (%)b

P(ii)3 (%)b

P(iii)3 (%)b

s

s

p

s

p

0.8 0.8

79.1 71.8 71.8

0.1

20.2 24.1 24.0

p 1.5 1.5

a

CMOs as depicted in Figure 3b. The three CMOs are routinely viewed as P 3p lone-pairs on bridging P(ii) sites, which are only an approximation. bThe P sites are labeled in Figure 2.

11960

DOI: 10.1021/acsomega.8b01790 ACS Omega 2018, 3, 11958−11965

ACS Omega

Article

Zintl ion each formally carry a negative charge and yet a P 3p2 lone-pair herein is not completely pure, which also partially participates in P(ii)−P(iii) bonding, moderately boosting the bridging P(ii)−P(iii) links to beyond single bonds (by a bond order of 0.2). This understanding updates the existing bonding picture of C3v [P7]3− Zintl ion,5,10 which is relevant to its reactivities in synthetic chemistry. We believe that the same is true for other relevant ligands, such as [As7]3− and [Sb7]3−. 3.2. Charged Molecular Alloy C2 [P7ZnP7]4− Cluster: An Eight-Electron “ZnP4” Core. Having understood the structure and bonding of naked C3v [P7]3− Zintl ion, those of C2 [P7ZnP7]4− as a charged molecular alloy cluster become simpler. Note that with C-PCM calculations to account for solvation effects, all occupied CMOs of C2 [P7ZnP7]4− are highly negative in energy eigenvalues, of which the HOMO is completely binding by about 3.71 eV, indicating a robust electronic system despite multiple charges. In addition, we have analyzed the wavefunction stability for [P7ZnP7]4−, and the result indicates that the wavefunction is stable under the perturbations considered. The C2 [P7ZnP7]4− cluster represents an alloy complex between a Zn center and two cage-like P7 ligands. The latter are not exactly [P7]3−; rather they appear to be in the [P7]2− charge state.44 Nonetheless, because the frontier CMOs of C3v [P7]3− Zintl ion are located on the P(ii) bridges and primarily lone-pairs in nature (Figures 3 and 4), the skeleton of P7 cage is robust and does not depend sensitively on its charge state. A couple of observations testify to the structural/chemical integrity of P7 cages in C2 [P7ZnP7]4− cluster. First, the overall structural pattern of one apical P(i) site, three P(ii) bridges, and a base P(iii)3 triangle is maintained. Specifically, P2 and P3 are apical sites, whereas P10P12P14 and P11P13P15 are triangular bases. Second, P−P distances in C3v [P7]3− are the shortest for bridging P(ii) sites and longest for triangular P(iii)3 base: 2.21 Å for P(i)−P(ii), 2.14 Å for P(ii)−P(iii), and 2.29 Å for P(iii)−P(iii). The corresponding P−P distances in C2 [P7ZnP7]4− are 2.16−2.20, 2.14−2.17, and 2.24−2.26 Å, respectively, which reflect the subtle trend of C3v [P7]3− Zintl ion, except for slight distortions due to lower symmetry of the charged molecular alloy cluster. Third, extra charges in C2 [P7ZnP7]4− cluster are also quite localized on the bridging sites (−0.51 to −0.58 |e| per atom), as compared to apical (−0.11 | e|) and base (−0.17 to −0.18 |e| per atom) sites. Quantitatively, about 71% of negative charges on a P7 cage ligand in C2 [P7ZnP7]4− are located on three bridging sites, compared with 64% in naked C3v [P7]3− Zintl ion. This observation further demonstrates that the bridging P sites are electrophilic, although the resultant 3p lone-pairs are not necessarily pure (Figure 4b and Table 1). In effect, each P7 cage in C2 [P7ZnP7]4− collectively carries a charge of −2.25 |e|, which deviates from C3v [P7]3− Zintl ion. As an 86-electron system, CMOs of C2 [P7ZnP7]4− cluster are shown in Figures S3 and S4 (Supporting Information), as well as Figure 6a. Briefly, each specific CMO of C3v [P7]3− Zintl ion forms a pair of CMOs in C2 [P7ZnP7]4− cluster via constructive/destructive couplings, which can in turn recover two identical bonds, one on each P7 ligand. Thus, all 32 CMOs in Figures S3b and S4 (Supporting Information) are P 3s2 lone-pairs or 2c-2e P−P σ bonds, whereas two CMOs in Figure S3c (Supporting Information) represent P 3p2 lonepairs on P8 and P9 sites. Furthermore, the five CMOs in Figure S3a (Supporting Information) are primarily Zn 3d10 lone-pairs. These Lewis elements consume 78 electrons.

triangle. These CMOs also represent a bonding/nonbonding/ antibonding combination, readily localized as three P 3p2 lonepairs (to a zeroth order approximation), each on one bridge. Such a picture is routinely discussed in synthetic literature,5,10 which can be elegantly borne out from AdNDP analysis (Figure 4). Occupation numbers (ONs) are a quantitative

Figure 4. Bonding pattern of C3v [P7]3− Zintl ion on the basis of AdNDP analysis. ONs are shown. (a) Seven P 3s2 lone-pairs. (b) Three bridging P 3p2 lone-pairs. (c) Nine 2c-2e P−P σ bonds.

indicator on how an AdNDP bond is close to the bonding essence in a molecular system. All ONs for 3s2 lone-pairs and 2c-2e σ bonds are nearly perfect (1.91−1.95 and 1.96−1.99 |e|, respectively), whereas those for three 3p2 lone-pairs on bridging P(ii) sites are 1.70 |e|. The latter ON values are less than ideal, suggesting that this treatment is approximate. Although we have clearly presented above that three 3p2 lonepairs on the P(ii) sites are only a zeroth order approximation (see also Table 1), one reviewer still questions their lower ON values (Figure 4b). To fully accommodate the reviewer, we offer here an alternative, refined AdNDP scheme for them (Figure 5), wherein the three 3p2 lone-pairs are refined as three

Figure 5. A refined AdNDP bonding scheme of C3v [P7]3−. Three P(ii) 3p2 lone-pairs are expanded to three P(ii)−P(iii) π bonds (from a local point-of-view), with enhanced ON values from 1.70 to 1.82 |e|. Note that this scheme is intrinsically the same as that in Figure 4b (see text).

P(ii)−P(iii) π “bonds” with slightly enhanced ONs of 1.82 |e|. It should be stressed that the two schemes do not differ fundamentally, because even for the three P(ii)−P(iii) π “bonds”, the relative contributions of P(ii) versus P(iii) are greater than 9 to 1. In other words, the effective π bond order is negligible (less than 0.2) despite being called a π “bond”. Indeed, the corresponding CMOs contain secondary (but nonnegligible) components from P(iii) sites, by 20−24% (Figure 3b and Table 1), which contribute to extra P(ii)− P(iii) bridging bonding. As a consequence, P(ii)−P(iii) distances are slightly shorter than the P(i)−P(ii) or P(iii)− P(iii) ones: 2.14 versus 2.21/2.29 Å (Figure 2). The same trend is true for their WBIs: 1.20 versus 1.00/0.80. Therefore, we should conclude that the bridging P(ii) sites in C3v [P7]3− 11961

DOI: 10.1021/acsomega.8b01790 ACS Omega 2018, 3, 11958−11965

ACS Omega

Article

Figure 6. Pictures of CMOs of (a) C2 [P7ZnP7]4− cluster that are responsible for Zn−P bonding in the ZnP4 core. Also shown for comparison are those of (b) model D2d [ZnP4H8]2− cluster, (c) Td CH4, and (d) Ne atom.

Figure 7. AdNDP pattern of C2 [P7ZnP7]4− cluster. ONs are shown. (a) Five Zn 3d10 lone-pairs. (b) Fourteen P 3s2 lone-pairs. (c) Two P 3p2 lonepairs on P8/P9 sites. (d) Eighteen 2c-2e P−P σ bonds. (e) Four 2c-2e Zn−P σ single bonds.

The key to chemical bonding in C2 [P7ZnP7]4− cluster lies in the ZnP4 core, which involves only eight electrons. Their CMOs are illustrated in Figure 6a, which consist of one completely bonding HOMO−23 between all four Zn−P links and three bonding CMOs each connecting two Zn−P links. The latter three CMOs are roughly equivalent to each other, in the spirit of triple degeneracy in the Td system. The four CMOs can formally be localized as four Zn−P σ bonds according to the building principles, rendering a classical eightelectron system with quasi-tetrahedral arrangement for C2 [P7ZnP7]4− cluster. In summary, the P7 cages in C2 [P7ZnP7]4− cluster appear to be in formal charge state of [P7]2−, in which each ligand uses two P(ii) 3p based orbitals to interact with Zn, forming an eight-electron bonding system. Here each ligand offers three electrons via two bridging P sites, with the Zn center balancing the remaining two electrons. The Zn center carries only a positive charge of +0.49 |e| (Figure 1), which deviates markedly from the anticipated ionic picture of [P7]3−[Zn]2+[P7]3−. Overall, the bonding between cage-like P7 ligands and Zn center in C2 [P7ZnP7]4− cluster is quite “localized” in nature. In C2 [P7ZnP7]4− cluster, an isolated Zn center interacts with two cage-like P7 ligands. Each P7 ligand offers bonding via two isolated bridging P sites, which have essentially nonbonding 3p electrons. These four P sites in the ZnP4 core coordinate to Zn center in a quasi-tetrahedral fashion with bond angles ∠PZnP of 82.8/122.4/124.9° because of geometric constraints. For comparison, an ideal Td system has bond angles of 109.5°. The AdNDP scheme (Figure 7) firmly confirms the above bonding picture for C2

[P7ZnP7]4−; see below in Section 3.3 for further discussion on the essence of eight-electron ZnP4 bonding. One reviewer suggests that we further confirm the [P7]2− charge state of cage ligands in C2 [P7ZnP7]4− cluster using additional computational tools. We stress that NBO analysis is widely considered to be reliable for such a molecular system, and therefore natural atomic charge of +0.49 |e| on Zn center indicates that the cage ligands are roughly on [P7]2− charge state,44 rather than [P7]3−. We also reiterate that the [P7]2− or [P7]3− charge state does not make a difference in terms of structure, stability, and bonding of cage ligands (Figures 4 and 5; Table 1). To accommodate the reviewer, we have run extended charge decomposition analysis (ECDA).45 We start with [P7]3−[Zn]2+[P7]3− in the ECDA calculation, and the analysis shows that 1.48 |e| are donated from two cage ligands to the Zn center. Thus, ECDA leads to a resultant configuration of [P7]2.26−[Zn]+0.52[P7]2.26−, which is identical to the NBO data (see Figure 1) and closely in line with our proposed bonding picture of [P7]2−[Zn]0[P7]2−. 3.3. Superatomic C2 [P7ZnP7]4− Cluster? The quasitetrahedral, eight-electron counting is a critical bonding feature for C2 [P7ZnP7]4− cluster, which has been reached based on indepth bonding analyses. We believe that this magic electron counting underlies the stability of C2 [P7ZnP7]4− as a synthetic charged molecular alloy cluster.11 Superatoms or superatomic clusters are a popular concept in cluster science,36−38 which may be exploited to elucidate the structure and bonding of C2 [P7ZnP7]4− cluster. To this end, we analyzed a model D2d [ZnP4H8]2− cluster (Figure S5a, Supporting Information); the latter is considered isoelectronic to C2 [P7ZnP7]4− cluster in 11962

DOI: 10.1021/acsomega.8b01790 ACS Omega 2018, 3, 11958−11965

ACS Omega

Article

Figure 8. AdNDP schemes for model D2d [ZnP4H8]2− cluster. ONs are shown. (a) The 5c-2e ZnP4 σ scheme. (b) Four 2c-2e Lewis Zn−P σ “single” bonds. (c) A scheme that separates “nonbonding” vs “bonding” components, whose combination recovers scheme (b), as illustrated at the right side. Scheme (c) reflects the essence of bonding in the system, which also works for C2 [P7ZnP7]4− cluster.

terms of bonding in the ZnP4 core. The D2d [ZnP4H8]2− cluster is a true minimum. Remarkably, it is virtually identical to C2 [P7ZnP7]4− cluster in Zn−P distances, WBIs, and nature charges (Figure S5 in Supporting Information versus Figure 1). We can therefore use the simpler D2d [ZnP4H8]2− cluster to mimic chemical bonding of C2 [P7ZnP7]4− cluster and to shed further light on the bonding essence in the latter system. CMO analysis for D2d [ZnP4H8]2− cluster is relatively straightforward (Figure S5, Supporting Information). This is a 42-electron system. Five Zn 3d10 lone-pairs are clearly identified, as are four P 3s2 lone-pairs and eight 2c-2e P−H σ bonds, which use 34 electrons. The remaining eight electrons are responsible for Zn−P bonding, whose CMOs (Figure 6b) show one-to-one correspondence to those of C2 [P7ZnP7]4− cluster (Figure 6a), except that each CMO of [ZnP4H8]2− has uniform contributions from all four Zn−P links due to its higher symmetry. Furthermore, the bonding patterns are analogous to those in CH4 (Figure 6c) and even Ne atom (Figure 6d), justifying an argument that C2 [P7ZnP7]4− cluster can be considered a superatom,36−38 whose eight-electron quasi-tetrahedral ZnP4 core underlies its electronic and structural stability. We shall now make a further comment on the nature of eight-electron Zn−P bonding in C2 [P7ZnP7]4− cluster, with the aid of model D2d [ZnP4H8]2− cluster. Owing to the polar nature of Zn−P bonding and the fact that two P7 cages contribute six electrons for the eight-electron system, in contrast to two electrons only for Zn center, the four Zn−P bonds in C2 [P7ZnP7]4− cluster is anticipated to be polar as well. On the basis of this understanding, the recovery of four 2c-2e Zn−P σ bonds in AdNDP analysis (Figure 7e) can be an overestimation for the system. Note that the ON values amount to 1.80−1.86 |e| for such bonds, which are close to ideal and markedly higher than their WBIs (0.48/0.49; Figure 1). One way to reconcile this discrepancy is to take a closer

look at the shape of such AdNDP bonds, as illustrated in Figure 8 using model D2d [ZnP4H8]2− cluster. Four 2c-2e Zn−P σ bonds with ideal ONs (1.99 |e|) are shown in Figure 8b. However, the shape of each bond is nonsymmetric for Zn versus P centers. Alternatively, one can first isolate the nonbonding P 3p component in AdNDP. Such a component is in the spirit of a lone-pair, but with a much lower ON (0.91 |e|; Figure 8c, top row). Subsequently, one analyzes the two-center Zn−P σ component, yielding an ON of 0.87 |e| (Figure 8c, bottom row). The latter represents the genuine covalent Zn−P bonding, in line with their low WBIs (0.48/0.49). The inset in Figure 8 shows that schemes (b) and (c) are equivalent to each other, that is, addition of nonbonding and two-center components generates the 2c-2e Zn−P σ “single” bonds. We reiterate that such “single” bonds are rather weak despite their ideal ONs in AdNDP analysis.

4. CONCLUSIONS We report on in-depth chemical bonding analyses for a synthetic charged molecular alloy C2 [P7ZnP7]4− cluster. Our effort makes use of electronic structure calculations, NBO analysis, and orbital composition analysis. To elucidate the nature of bonding, we have performed extensive CMO analyses and AdNDP. The latter approach offers an elegant bonding picture, which was seldom exploited in synthetic works in the literature. Structural, electronic, and bonding properties of cage-like C3v [P7]3− Zintl ion are also examined, showing that three isolated bridging sites carry the majority of negative charges. Such charges have considerable P 3p2 lonepair characters, but also participate in secondary P−P bonding along bridging sites. This bonding picture refines existing knowledge on naked [P7]3− Zintl ion. Charged molecular alloy C2 [P7ZnP7]4− cluster can be approximately formulated as [P7]2−[Zn]0[P7]2−, in which [P7]2− ligands inherit the structure and bonding of Zintl ion except for a difference in 11963

DOI: 10.1021/acsomega.8b01790 ACS Omega 2018, 3, 11958−11965

ACS Omega charge state. Each [P7]2− ligand interacts with the Zn center via two bridging P sites, forming a quasi-tetrahedral ZnP4 core. The eight-electron counting underlies the stability of this charged molecular alloy, which can alternatively be considered as a superatom or superatomic cluster. The Zn−P bonding in C2 [P7ZnP7]4− cluster is weaker than anticipated for an eightelectron tetrahedral configuration because the P centers maintain partial nonbonding 3p character, which is in the spirit of lone-pair albeit with a much lower ON. As a consequence, the effective Zn−P bond order is around 0.5.

■

REFERENCES

(1) Ahlrichs, R.; Fenske, D.; Fromm, K.; Krautscheid, H.; Krautscheid, U.; Treutler, O. Zintl anions as starting compounds for the synthesis of polynuclear transition metal complexes. Chem. Eur J. 1996, 2, 238−244. (2) Fässler, T. F.; Hoffmann, S. D. Endohedral Zintl ions: intermetalloid clusters. Angew. Chem., Int. Ed. 2004, 43, 6242−6247. (3) Sevov, S. C.; Goicoechea, J. M. Chemistry of deltahedral Zintl ions. Organometallics 2006, 25, 5678−5692. (4) Fässler, T. F. Relationships between soluble Zintl anions, ligandstabilized cage compounds, and intermetalloid clusters of tetrel (Si− Pb) and pentel (P−Bi) elements. Struct. Bond. 2011, 140, 91−131. (5) Scharfe, S.; Kraus, F.; Stegmaier, S.; Schier, A.; Fässler, T. F. Zintl ions, cage compounds, and intermetalloid clusters of group 14 and group 15 elements. Angew. Chem., Int. Ed. 2011, 50, 3630−3670. (6) Weinert, B.; Dehnen, S. Binary and ternary intermetalloid clusters. Struct. Bond. 2017, 174, 99−134. (7) Corbett, J. D. Polyatomic Zintl anions of the post-transition elements. Chem. Rev. 1985, 85, 383−397. (8) Corbett, J. D. Polyanionic clusters and networks of the early pelement metals in the solid state: beyond the Zintl boundary. Angew. Chem., Int. Ed. 2000, 39, 670−690. (9) Fässler, T. F. Homoatomic polyhedra as structural modules in chemistry: what binds fullerenes and homonuclear Zintl ions? Angew. Chem., Int. Ed. 2001, 40, 4161−4165. (10) Turbervill, R. S. P.; Goicoechea, J. M. From clusters to unorthodox pnictogen sources: solution-phase reactivity of [E7]3− (E = P−Sb) anions. Chem. Rev. 2014, 114, 10807−10828. (11) Knapp, C.; Zhou, B.; Denning, M. S.; Rees, N. H.; Goicoechea, J. M. Reactivity studies of group 15 Zintl ions towards homoleptic post-transition metal organometallics: a “bottom-up” approach to bimetallic molecular clusters. Dalton Trans. 2010, 39, 426−436. (12) Chaki, N. K.; Mandal, S.; Reber, A. C.; Qian, M.; Saavedra, H. M.; Weiss, P. S.; Khanna, S. N.; Sen, A. Controlling band gap energies in cluster-assembled ionic solids through internal electric fields. ACS Nano 2010, 4, 5813−5818. (13) Yang, L.-M.; Ding, Y.-H.; Sun, C.-C. Design of the sandwichlike compounds based on the all-metal aromatic unit Al3−. ChemPhysChem 2006, 7, 2478−2482. (14) Yang, L.-M.; Ding, Y.-H.; Sun, C.-C. Design of sandwichlike complexes based on the planar tetracoordinate carbon unit CAl42−. J. Am. Chem. Soc. 2007, 129, 658−665. (15) Yang, L.-M.; Ding, Y.-H.; Sun, C.-C. Assembly and stabilization of a planar tetracoordinated carbon radical CAl3Si: a way to design spin-based molecular materials. J. Am. Chem. Soc. 2007, 129, 1900− 1901. (16) Yang, L.-M.; Ding, Y.-H.; Sun, C.-C. Sandwich-like compounds based on the all-metal aromatic unit Al42− and the main-group metals M (M = Li, Na, K, Be, Mg, Ca). Chem.Eur. J. 2007, 13, 2546−2555. (17) Yang, L.-M.; Wang, J.; Ding, Y.-H.; Sun, C.-C. Theoretical study on the assembly and stabilization of a silicon-doped all-metal aromatic unit SiAl3−. Organometallics 2007, 26, 4449−4455. (18) Yang, L.-M.; Wang, J.; Ding, Y.-H.; Sun, C.-C. Investigation of the typical triangular structure B3 in boron chemistry: insight into bare all-boron clusters used as ligands or building blocks. J. Phys. Chem. A 2007, 111, 9122−9129. (19) Yang, L.-M.; Ding, Y.-H.; Sun, C.-C. Theoretical study on the assembly and stabilization of a magic cluster Al4N−. J. Phys. Chem. A 2007, 111, 10675−10681. (20) Yang, L.-M.; Ding, Y.-H.; Tian, W. Q.; Sun, C.-C. Planar carbon radical’s assembly and stabilization, a way to design spin-based molecular materials. Phys. Chem. Chem. Phys. 2007, 9, 5304−5314.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsomega.8b01790. Cartesian coordinates of C2 [P7ZnP7]4− cluster at the PBE0/def2-TZVP level; view of crystal structure [K(2,2,2-crypt)]4[ZnP14]·6py; optimized structure of [P7ZnP7]4− cluster using C-PCM calculations at the PBE0 level, with bond distances for all P−P and Zn−P links; pictures of CMOs of [P7ZnP7]4− cluster associated to lone-pairs and 2c-2e P−P bonds; and optimized structure of model D2d [ZnP4H8]2− cluster at PBE0 level and pictures of selected CMOs (PDF)

■

ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China (21873058, 21573138) and the Sanjin Scholar Distinguished Professors Program.

5. METHODS SECTION Geometric structures of [P7ZnP7]4− cluster were constructed based on experimental crystal data11 and subsequently reoptimized using DFT29 at the PBE0/def2-TZVP level.30,39 C-PCM calculations31−34 were carried out to take into account the solvation effects, whose performance has been tested recently in charged molecular alloy compounds.27 The NBO analysis was performed to obtain WBIs and natural atomic charges, which serve as a crude measure of the P−P and Zn−P bonding, as well as charge distribution in the system. To elucidate chemical bonding, CMOs were fully analyzed following the basic building principles, which is a sophisticated process considering that [P7ZnP7]4− cluster is an 86-electron system. CMO analyses were aided with orbital composition calculations; the latter offers quantitative data on how a CMO is composed of specific AOs. To simplify CMO analyses, the [P7]3− Zintl ion was examined first. The bonding pictures of [P7]3− Zintl ion and [P7ZnP7]4− cluster were independently borne out from AdNDP analyses, performed using the AdNDP program.35 The CMOs and AdNDP data were visualized using the GaussView46 and Molekel47 programs. All electronic structure calculations were carried out using the Gaussian 09 package.48 Orbital composition analysis and ECDA45 were performed using the Multiwfn program.49

■

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Hua-Jin Zhai: 0000-0002-1592-0534 Notes

The authors declare no competing financial interest. 11964

DOI: 10.1021/acsomega.8b01790 ACS Omega 2018, 3, 11958−11965

ACS Omega

Article

(21) Yang, L.-M.; Ding, Y.-H.; Sun, C.-C. The Si-doped planar tetracoordinate carbon (ptC) unit CAl3Si− could be used as a building block or inorganic ligand during cluster-assembly. Theor. Chem. Acc. 2008, 119, 335−342. (22) Yang, L.-M.; He, H.-P.; Ding, Y.-H.; Sun, C.-C. Achieving stable hypercarbon CB62− based cluster-assembled complexes: a general strategy. Organometallics 2008, 27, 1727−1735. (23) Yang, L.-M.; Shao, C.-B.; Ding, Y.-H.; Sun, C.-C. Clusterassembled compounds comprising an all-metal subunit Li3Al4−. Phys. Chem. Chem. Phys. 2008, 10, 2020−2025. (24) Yang, L.-M.; Wang, J.; Ding, Y.-H.; Sun, C.-C. Sandwich-like compounds based on bare all-boron cluster B62−. Phys. Chem. Chem. Phys. 2008, 10, 2316−2320. (25) Yang, L.-M.; Li, X.-P.; Ding, Y.-H.; Sun, C.-C. Theoretical study on a class of organometallic complexes based on all-metal aromatic Ga3− through sandwiching stabilization. Eur. J. Inorg. Chem. 2008, 2099−2106. (26) Yang, L.-M.; Li, X.-P.; Ding, Y.-H.; Sun, C.-C. CSi2Ga2: a neutral planar tetracoordinate carbon (ptC) building block. J. Mol. Model. 2009, 15, 97−104. (27) You, X.-R.; Feng, L.-Y.; Li, R.; Zhai, H.-J. Chemical bonding and σ-aromaticity in charged molecular alloys: [Pd2As14]4− and [Au2Sb14]4− clusters. Sci. Rep. 2017, 7, 791. (28) Wilson, R. J.; Dehnen, S. (Ge4Bi14)4−: A case of “element segregation” on the molecular level. Angew. Chem., Int. Ed. 2017, 56, 3098−3102. (29) Hohenberg, P.; Kohn, W. Inhomogeneous electron gas. Phys. Rev. 1964, 136, B864−B871. (30) Adamo, C.; Barone, V. Toward reliable density functional methods without adjustable parameters: the PBE0 model. J. Chem. Phys. 1999, 110, 6158−6170. (31) Klamt, A.; Schüürmann, G. COSMO: a new approach to dielectric screening in solvents with explicit expressions for the screening energy and its gradient. J. Chem. Soc., Perkin Trans. 2 1993, 799−805. (32) Andzelm, J.; Kölmel, C.; Klamt, A. Incorporation of solvent effects into density functional calculations of molecular energies and geometries. J. Chem. Phys. 1995, 103, 9312−9320. (33) Barone, V.; Cossi, M. Quantum calculation of molecular energies and energy gradients in solution by a conductor solvent model. J. Phys. Chem. A 1998, 102, 1995−2001. (34) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. Energies, structures, and electronic properties of molecules in solution with the C-PCM solvation model. J. Comput. Chem. 2003, 24, 669−681. (35) Zubarev, D. Y.; Boldyrev, A. I. Developing paradigms of chemical bonding: adaptive natural density partitioning. Phys. Chem. Chem. Phys. 2008, 10, 5207−5217. (36) Castleman, A. W., Jr.; Khanna, S. N. Clusters, superatoms, and building blocks of new materials. J. Phys. Chem. C 2009, 113, 2664− 2675. (37) Claridge, S. A.; Castleman, A. W., Jr.; Khanna, S. N.; Murray, C. B.; Sen, A.; Weiss, P. S. Cluster-assembled materials. ACS Nano 2009, 3, 244−255. (38) Jena, P.; Sun, Q. Super atomic clusters: design rules and potential for building blocks of materials. Chem. Rev. 2018, 118, 5755−5870. (39) Weigend, F.; Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: design and assessment of accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297−3305. (40) Pyykkö, P. Additive covalent radii for single-, double-, and triple-bonded molecules and tetrahedrally bonded crystals: a summary. J. Phys. Chem. A 2015, 119, 2326−2337. (41) Covalent radii give an upper bound of bond distances for covalent bonds only. For the Zn−P bonds in C2 [P7ZnP7]4− cluster, the recommended bond distances40 can be unreliable due to the polar nature. (42) Data taken from CSD, version 5.30 (November 2008).

(43) Sergeeva, A. P.; Averkiev, B. B.; Boldyrev, A. I. All-transition metal aromaticity and antiaromaticity. Struct. Bond. 2010, 136, 275− 305. (44) We mention two facts here: (i) HOMO and HOMO−1 of C2 [P7ZnP7]4− cluster, in which the extra charges occupy, are all based on cage P7 ligands (Figure S3c, Supporting Information). (ii) There is little net charge transfer from Zn center to cage P7 ligands (Figure 1), suggesting that it is appropriate to formulate the charged molecular alloy cluster as [P7]2−[Zn]0[P7]2−. (45) Dapprich, S.; Frenking, G. Investigation of donor-acceptor interactions: a charge decomposition analysis using fragment molecular orbitals. J. Phys. Chem. 1995, 99, 9352−9362. (46) Dennington, R.; Keith, T.; Millam, J. GaussView, version 5.0.9; Semichem, Inc.: Shawnee Mission, KS, 2009. (47) Varetto, U. Molekel 5.4.0.8; Swiss National Supercomputing Center: Manno, Switzerland, 2009. (48) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, Revision D.01; Gaussian, Inc.: Wallingford, CT, 2009. (49) Lu, T.; Chen, F. Multiwfn: a multifunctional wavefunction analyzer. J. Comput. Chem. 2012, 33, 580−592.

11965

DOI: 10.1021/acsomega.8b01790 ACS Omega 2018, 3, 11958−11965