Electron-Deficient Multicenter Bonding in Phase Change Materials: A Chance for Reconciliation.

In the last few years, a controversy has been raised regarding the nature of the chemical bonding present in phase change materials (PCMs), many of which are minerals such as galena (PbS), clausthalite (PbSe), and altaite (PbTe). Two opposite bonding models have claimed to be able to explain the extraordinary properties of PCMs in the last decade: the hypervalent (electron-rich multicenter) bonding model and the metavalent (electron-deficient) bonding model. In this context, a third bonding model, the electron-deficient multicenter bonding model, has been recently added. In this work, we comment on the pros and cons of the hypervalent and metavalent bonding models and briefly review the three approaches. We suggest that both hypervalent and metavalent bonding models can be reconciled with the third way, which considers that PCMs are governed by electron-deficient multicenter bonds. To help supporters of the metavalent and hypervalent bonding model to change their minds, we have commented on the chemical bonding in GeSe and SnSe under pressure and in several polyiodides with different sizes and geometries.

Chemical bonding in phase-change chalcogenides.

Almost all phase-change memory materials (PCM) contain chalcogen atoms, and their chemical bonds have been denoted both as 'electron-deficient' [sometimes referred to as 'metavalent'] and 'electron-rich' ['hypervalent', multicentre]. The latter involve lone-pair electrons. We have performed calculations that can discriminate unambiguously between these two classes of bond and have shown that PCM have electron-rich, 3c-4e ('hypervalent') bonds. Plots of charge transferred between (ET) and shared with (ES) neighbouring atoms cannot on their own distinguish between 'metavalent' and 'hypervalent' bonds, both of which involve single-electron bonds. PCM do not exhibit 'metavalent' bonding and are not electron-deficient; the bonding is electron-rich of the 'hypervalent' or multicentre type.

Metavalent or Hypervalent Bonding: Is There a Chance for Reconciliation?

A family of solids including crystalline phase change materials such as GeTe and Sb2 Te3 , topological insulators like Bi2 Se3, and halide perovskites such as CsPbI3 possesses an unconventional property portfolio that seems incompatible with ionic, metallic, or covalent bonding. Instead, evidence is found for a bonding mechanism characterized by half-filled p-bands and a competition between electron localization and delocalization. Different bonding concepts have recently been suggested based on quantum chemical bonding descriptors which either define the bonds in these solids as electron-deficient (metavalent) or electron-rich (hypervalent). This disagreement raises concerns about the accuracy of quantum-chemical bonding descriptors is showed. Here independent of the approach chosen, electron-deficient bonds govern the materials mentioned above is showed. A detailed analysis of bonding in electron-rich XeF2 and electron-deficient GeTe shows that in both cases p-electrons govern bonding, while s-electrons only play a minor role. Yet, the properties of the electron-deficient crystals are very different from molecular crystals of electron-rich XeF2 or electron-deficient B2 H6 . The unique properties of phase change materials and related solids can be attributed to an extended system of half-filled bonds, providing further arguments as to why a distinct nomenclature such as metavalent bonding is adequate and appropriate for these solids.

The Myth of "Metavalency" in Phase-Change Materials.

Phase-change memory materials (PCMs) have unusual properties and important applications, and recent efforts to find improved materials have focused on their bonding mechanisms. "Metavalent bonding" or "metavalency," intermediate between "metallic" and "covalent" bonding and comprising single-electron bonds, has been proposed as a fundamentally new mechanism that is relevant both here and for halide perovskite materials. However, it is shown that PCMs, which violate the octet rule, have two types of covalent bond: two-center, two-electron (2c-2e) bonds, and electron-rich, multicenter bonds (3c-4e bonds, hyperbonds) involving lone-pair electrons. The latter have bond orders less than one and are examples of the century-old concept of "partial" bonds.

Undistorted linear Bi chains with hypervalent bonding in La3TiBi5 from single-crystal X-ray diffraction.

The crystal structure of the lanthanum titanium bismuthide La3TiBi5 (Pearson code hP18, Wyckoff sequence b d g2) has been established from single-crystal X-ray diffraction data and analyzed in detail using first-principles calculations. There are no anomalies pertaining to the atomic displacement parameter of the Ti site, previously reported based on a powder X-ray diffraction analysis of this compound. The anionic substructure contains columns of face-sharing TiBi6 octahedra and linear Bi chains. Due to a significant La(5d) and Bi(6p) orbital mixing, a perfectly one-dimensional character of the Bi chains is not realised, while a three-dimensional electronic structure is established instead. The latter fact explains the stability of the polyanionic pnictide units against Peierls distortions. The hypervalent bonding in the Bi chains is reflected in a rather long Bi-Bi distance of 3.2264 (4) Šand a typical pattern of bonding and antibonding interactions, as revealed by electronic structure calculations.

The solvent effect on two competing reaction mechanisms involving hypervalent iodine reagents (λ(3)-iodanes): facing the limit of the stationary quantum chemical approach.

Trifluoromethylation of acetonitrile with 3,3-dimethyl-1-(trifluoromethyl)-1λ(3),2- benziodoxol is assumed to occur via reductive elimination (RE) of the electrophilic CF3-ligand and MeCN bound to the hypervalent iodine. Computations in gas phase showed that the reaction might also occur via an SN2 mechanism. There is a substantial solvent effect present for both reaction mechanisms, and their energies of activation are very sensitive toward the solvent model used (implicit, microsolvation, and cluster-continuum). With polarizable continuum model-based methods, the SN2 mechanism becomes less favorable. Applying the cluster-continuum model, using a shell of solvent molecules derived from ab initio molecular dynamics (AIMD) simulations, the gap between the two activation barriers ( ΔΔG‡) is lowered to a few kcal mol(-1) and also shows that the activation entropies (ΔS‡) and volumes (ΔV‡) for the two mechanisms differ substantially. A quantitative assessment of ΔΔG‡ will therefore only be possible using AIMD. A natural bond orbital-analysis gives further insight into the activation of the CF3-reagent by protonation.