Two-Dimensional Antimony Oxide
Stefan Wolff, Roland Gillen, Mhamed Assebban, Gonzalo Abellán, Janina Maultzsch
TTwo-Dimensional Antimony Oxide
Stefan Wolff, ∗ Roland Gillen, Mhamed Assebban,
2, 3
Gonzalo Abell´an,
2, 3 and Janina Maultzsch Department of Physics, Chair of Experimental Physics,Friedrich-Alexander-Universit¨at Erlangen-N¨urnberg (FAU), Staudtstr. 7, 91058 Erlangen, Germany Instituto de Ciencia Molecular (ICMol), Universidad de Valencia,Catedr´atico Jos´e Beltr´an 2, 46980, Paterna, Valencia, Spain. Department of Chemistry and Pharmacy & Joint Institute of Advanced Materials and Processes (ZMP),Friedrich-Alexander-Universit¨at Erlangen-N¨urnberg (FAU), Dr.-Mack-Straße 81, 90762, F¨urth, Germany. (Dated: March 6, 2020)Two-dimensional (2D) antimony, so-called antimonene, can form antimonene oxide when exposedto air. We present different types of single- and few-layer antimony oxide structures, based ondensity functional theory (DFT) calculations. Depending on stoichiometry and bonding type, thesenovel 2D layers have different structural stability and electronic properties, ranging from topologicalinsulators to semiconductors with direct and indirect band gaps between 2.0 and 4.9 eV. We discusstheir vibrational properties and Raman spectra for experimental identification of the predictedstructures.
The exfoliation of a single layer of graphene from bulkgraphite unleashed a new field in physics and chemistryfocusing on the investigation of two-dimensional (2D)layered crystals [1, 2]. Over the past years an increasingnumber of 2D materials with vastly different propertieshave been discovered. Group-15 elements, also known aspnictogens, are suitable to form monoelemental 2D lay-ered materials, which are promising candidates for a vari-ety of applications in the field of plasmonics [3], for sens-ing [4], electronic [5–7], and optoelectronic [8] devices.Antimony is one of these elements and can form layeredstructures called antimonene. Recently, few-layer anti-monene was realized experimentally by different meth-ods such as epitaxial growth [9–11] or exfoliation [12–15].The electronic and vibrational properties of monolayerantimonene have been investigated theoretically, with apredicted band gap of about 2.4 eV [13, 16–22]. Anti-monene also appears reactive to air, however, in contrastto black phosphorus [23–29], it seems to form new sta-ble structures after oxidation. An oxidation process mayeven be favorable for tailoring the electronic propertiessince the electronic band structure depends on the degreeof oxidation. Additionally, antimonene seems to be thefirst elemental 2D material that forms a stable 2D oxidenaturally. Because of its similarities to other 2D pnicto-gens, such processes may also occur for related materi-als. Oxidation could then be used to further increase thequality of such monoelemental materials by, e.g., theirencapsulation between oxidized layers.However, the actual structure of oxidized antimonenelayers and their physical properties are unknown [30].In Ref. [22], for instance, the theoretical predictions arebased on a monolayer antimonene with Sb=O doublebonds perpendicular to the antimonene plane. Takinginto account that bulk antimony oxides exist in severaldifferent compositions ( α -Sb O , β -Sb O , Sb O , andmixtures thereof) [31, 32] and display polymorphism,other structures may exist, where the oxygen atoms are bound to at least two Sb atoms and are incorporated intothe antimonene planes. Few-layer antimonene preparedby exfoliation is likely to undergo oxidation [12–15, 30],but knowledge about oxidized few-layer Sb is missing. Aselectronic properties of oxidized antimonene are expectedto depend crucially on the structure and the bonding be-tween oxygen and antimony atoms, a precise knowledgeof the atomic structure is essential for developing this newmaterial. The ability to control the oxidation process canthen be used to tailor the electronic band structure.In this Letter we present two-dimensional antimonyoxide single- and few-layer structures with properties de-pending on bonding type and stoichiometry. Based ondensity functional theory (DFT) calculations, we showthat our proposed novel 2D antimony oxide structuresare semiconducting with direct and indirect band gapsbetween 2.0 and 4.9 eV. Furthermore, we present theirvibrational modes for experimental identification. Weexpect that semimetallic few-layer antimonene can natu-rally form heterostructures with semiconducting oxidizedlayers.Antimony oxide mono- (1L) and bilayer (2L) struc-tures with Sb=O double bonds, perpendicular to theplane, inspired by the fully oxidized antimonene mono-layer proposed in Ref. [22], are shown in Figs. 1(a)-1(d).These structures are here referred to as type (I) . Oursimulations of the 1L phonon spectrum (see below andFig. S12) strongly suggests that the type (I) structuresare metastable and likely transition into another morestable configuration. No stable system with more thantwo layers is found.We introduce more stable antimonene oxide structureswith different stoichiometries, where the oxygen atom isbound to at least two antimony atoms (Sb-O-Sb). We callthese structures type (II) , see Figs. 1(e)-1(j). They areobtained by a frozen-phonon approach by displacing theatoms in a unit cell of type (I) antimonene oxide accord-ing to a calculated phonon mode with negative frequency a r X i v : . [ c ond - m a t . m t r l - s c i ] M a r (a) 1L, type (I), side view (b) 1L, type (I), top view(c) 2L, type (I), side view (d) 2L, type (I), top view(e) Sb O, 3L, type (II) (f) Sb O, 1L, type (II)(g) Sb O , 3L, type (II) (h) Sb O , 1L, type (II)(i) Sb O , 3L, type (II) (j) Sb O , 1L, type (II) FIG. 1. Type (I) antimonene oxide structures with one layerin side view (a) and top view (b), and two layers in (c) and (d).Type (II) antimonene oxide heterostructures with differentstoichiometry of the oxidized layers: (e) and (f) Sb O; (g)and (h) Sb O ; and (i) and (j) Sb O . Type (II) structuresin top view show the oxidized layer only. Oxygen (antimony)atoms are shown in red (gray). Labels a and b on the figuresindicate the in-plane lattice vectors. For details about thestructural parameters, see Tables S1 and S2. found in trilayer type (I) antimonene oxide. Using thisapproach on a trilayer system consisting of a non-oxidizedantimonene layer, sandwiched by the bilayer shown inFig. 1(c), results in the structure of reduced symmetryshown in Figs. 1(e) and 1(f). Corresponding to the unitcell of the top/bottom layer, we call this structure Sb O.By increasing the amount of oxygen in the outer layers,such that the number of oxygen atoms matches the num-ber of antimony atoms (Sb O ), the structure changeseven further: “chains” of alternating oxygen and anti-mony atoms are formed, which are each connected bythree bonds to their neighbouring atoms. The chains are connected to each other by an additional bond betweentwo antimony atoms, see Figs. 1(g) and 1(h).We increase the oxygen concentration in the outerlayers to three oxygen atoms per two antimony atoms(Sb O ), see Figs. 1(i) and 1(j). The chain structure ofalternating oxygen and antimony atoms is maintained.The additional oxygen atom is replacing the Sb-Sb bond,forming an Sb-O-Sb bond. As indicated in Fig. 1, the dif-ferent structures are labeled by the amount of antimonyand oxygen atoms per unit cell in a single outer layer.In few-layer antimonene the interlayer bonds have asignificant covalent contribution to the otherwise nonco-valent van der Waals interaction, as shown in previousexperimental and theoretical work. Therefore we assumethat the inner, nonoxidized layer will indeed be affectedby a change of the lattice parameters induced by the oxi-dation of the outer layers. The trilayer structures shownin Figs. 1(e), 1(g), 1(i) thus provide a qualitative struc-tural picture in comparison to the original type (I) struc-tures. However, the small unit cells used in our calcu-lations combined with a significant change of the latticevectors of the oxidized layers with increasing oxidationinduce a strain of up to 17% in the inner, nonoxidizedlayers. We refer to Sec. 5 of the Supplemental Materialfor a detailed discussion. We note that for the Sb O monolayer, the structure depicted in Fig. 1(j) is not fullydynamically stable and relaxes into a slightly distortedgeometry if a larger unit cell is used (Fig. S4). Becauseof the small energy difference (5 meV per Sb O formulaunit), we will further use the idealized Sb O layer shownin Figs. 1(i) and 1(j) for reasons of convenience. For realsamples, we expect the formation of an amorphouslikeantimonene oxide capping layer.We now turn to the properties of the individual an-timonene oxide monolayers with different degrees of ox-idation, depicted in Figs. 1(f), 1(h), 1(i), and bilayersthereof. The atomic positions and lattice vectors of allisolated structures were optimized. The phase transfor-mation from type (I) to type (II) lowers the total energiesof 1L and 2L structures by roughly 3.9 and 2.7 eV, respec-tively, indicating that type (I) structures are metastableat best and transform into type (II)-like arrangements.To further confirm this, we performed three sets of molec-ular dynamics (MD) simulations at a temperature of300 K. In the first set (Fig. S5), we started with thehexagonal primitive cell of monolayer antimonene andadded three oxygen atoms close to the antimonene layer.The resulting equilibrated structure closely resembles theSb O structure of Fig. 1(j). In the second set (Fig. S6),a 4 × × type (I) E g : 84 cm − A g : 823 cm − type (II)Sb O 316 cm −
587 cm − type (II)Sb O
579 cm −
592 cm − type (II)Sb O
391 cm −
762 cm − FIG. 2. Exemplary display of phonon modes of the monolayerantimonene oxide structures investigated here. The type (II)Sb O and Sb O on the left-hand side are shown from thetop. All other structures are shown from the side. Greenarrows indicate the displacements of the atoms and are tologarithmic scale. is incorporated into the antimonene layer and chains ofSb-O bonds are formed. The well-known bulk antimonyoxides with Sb O and Sb O stoichiometries have nocommon structural motif with the 2D antimonene oxidelayers presented here.For an equal number of atoms in a given type (II)structure, the total energy decreases by roughly 1.3 eVper oxygen atom for an increasing amount of oxygen.This has been verified by calculating the energy of an O molecule and adding or subtracting the energy, respec-tively. The emergence of 2D antimonene oxide with ahigher degree of oxidation than Sb O , however, is ratherunlikely: Any further increase of the oxygen concentra-tion did not result in stable 2D layers, since the numberof antimony atoms which oxygen atoms can bind to islimited.In order to provide a guideline for identifying the differ-ent antimonene oxide structures experimentally, e.g., byRaman spectroscopy, we present their vibrational prop-erties in the following. The type (I) structures belong, like pristine anti-monene, to the D d symmetry group; therefore the vibra-tional modes include modes with E g and A g symmetry(Fig. 2). All twelve modes of type (I) monolayers (fouratoms per unit cell, see also Fig. S8) fall into two regions,one below 180 cm − and one at around 820 cm − , seeFig. 3. The latter corresponds to stretching of the Sb=Obonds and is indicative of a type (I) structure. The fre-quency range below 180 cm − comprises the vibrationswithin the Sb layers; in addition there are rigid-layer vi-brations in the case of the bilayer system.The calculation of the phonon dispersion of the type(I) monolayer structure (Fig. S12) results in negative fre-quencies of the acoustic branches over a large region ofthe Brillouin zone. This indicates that such structuresare not stable experimentally.The highest phonon frequency in the type (II) Sb Ostructure is at about 590 cm − and is a mode with anout-of-plane component, see Fig. 2. A second char-acteristic mode of the type (II) Sb O layer, at about315 cm − , is dominated by a motion of the oxygen atoms(Fig. 2 and 3). For the displacement patterns of all vi-brational modes, see Fig. S9. The symmetry of thesestructures is C for 1L and C i for 2L and 3L.The monolayer type (II) Sb O structure [Figs. 1(g)and 1(h)] can be further symmetrized such that it corre-sponds to the C symmetry group. The frequency of theout-of-plane mode decreases to about 530 cm − and themode becomes Raman inactive. Two additional modesarise at around 590 cm − , which are also dominated bya displacement of oxygen atoms. The highest-frequencymode is an in-plane vibration, whereas the other mode isalong the bond between oxygen and antimony atoms per-pendicular to the direction of the Sb-O-Sb chain (Figs. 2and 3); see Fig. S10 for all displacement patterns.The additional oxygen atom in the type (II) Sb O structure leads to five atoms per unit cell for 1L, i.e.,15 phonon modes, which are illustrated in Fig. S11. Incomparison to the previously discussed structures, theadditional vibrational modes occur at roughly 300 cm − ,750 cm − , and in the range of 375 to 430 cm − for differ-ent layer numbers (Figs. 2 and 3).The phonon modes observed in the monolayer type(II) structures also appear in the respective 2L and 3Lstructures. The displacement patterns of these modesare qualitatively maintained and show overall similarfrequencies, though some are shifted by up to about50 cm − .The characteristic E g and A g Raman modes of pris-tine monolayer antimonene at 168 and 206 cm − areshown in Fig. 3(a). In Figs. 3(b)-3(e), the positions ofthe Raman active modes in the monolayer type (I) andtype (II) structures with different oxygen concentrationare presented. The characteristic frequencies shown inFig. 3 can be used for experimental identification of dif-ferent antimonene oxide structures. (a) antimonene(b) type (I)(c) type (II) Sb O(d) type (II) Sb O (e) type (II) Sb O − ) l og . R a m a n a c t i v i t y ( a r b . un i t s ) FIG. 3. Calculated frequencies of Raman-active vibrationalmodes in (a) antimonene, (b) type (I), (c)-(e) type (II) anti-monene oxide structures, as indicated in the insets and shownin Fig. 1. Except for (b), the height of the bars indicates thecalculated Raman activity (log. scale).
Our predictions of stable type (II) antimonene oxidelayers are in agreement with recent experiments on theoxidation behavior of liquid-phase exfoliated few-layerantimonene [33]. Reference [33] reports the formationof a passivation layer on the surface, which shows evi-dence for Sb O -like layers. Raman measurements revealcharacteristic modes in the range of 190-450 cm − . Thisexperimentally rules out the formation of type (I) struc-tures. Instead, predicted phonon modes of the Sb O layers fit reasonably well to the experimentally observedspectra [33].While bulk and few-layer antimonene are metallicdue to a partial covalent bonding between the layers,first-principles calculations on the GW level predict avalue of about 2.4 eV for nonoxidized monolayer anti-monene [21]. Fully oxidized monolayer antimonene withdouble-bonded oxygen atoms [type (I)] was previouslypredicted to be a topological insulator with a small“bulk” band gap if spin-orbit interaction is included [22].Figure 4 shows the calculated electronic band structuresof oxidized monolayer antimonene using the hybrid func-tional HSE12 [34], based on the atomic geometries ofFig. 1(a) [type (I)] and our proposed structures [type(II)] from Figs. 1(h) and 1(j). For type (I) Sb O , ourcalculations agree with the results of Ref. [22], showing aband gap of 168 meV; the system is metallic if spin-orbitinteraction is neglected [gray dashed lines in Fig. 4(a)].We refer to Sec. 5 of the Supplemental Material for adiscussion of the electronic band structure of the trilayer structures shown in Figs. 1(e), 1(g), 1(i). For type (II)Sb O , with the more stable chainlike configuration, ourcalculations predict the system to be a trivial insulatorwith a direct band gap of about 2.0 eV at the edge of theBrillouin zone [Fig. 4(b)]. In the latter case, there is nodiscernable effect of spin-orbit coupling on the electronicdispersion. Increasing the oxygen content in the unit cell[type (II) Sb O , Fig. 1(j)] causes a transition from directto indirect semiconductor and significantly increases theband gap to 4.9 eV. This suggests that both the size andthe nature of the fundamental band gap in oxidized anti-monene could be tuned from the visible to the ultravioletrange, if control over the oxidation can be achieved. Weexpect a natural formation of heterostructures in whichmultiple semimetallic antimonene layers are sandwichedbetween semiconducting oxidized antimonene layers, dueto the high reactivity between antimony and oxygen ob-served in Ref. [33].In summary, we present layered antimonene oxidestructures with the oxygen atoms incorporated into theantimonene sheet [type (II)]. They are more stable thanconfigurations with Sb=O double bonds perpendicularto the antimonene plane [type (I)]. Distinct differences inthe vibrational frequencies between type (I) and differenttype (II) antimonene oxides allow an experimental iden-tification of the structures via Raman spectroscopy. Thisis in good agreement with recent experimental findingson liquid-phase exfoliated few-layer antimonene [33]. Alltype (II) single-layer antimonene oxides presented hereare semiconductors with stochiometry-dependent bandgaps ranging from approximately 2.0 to 4.9 eV. Our re-sults thus pave the way for tailoring the electronic bandstructure of antimonene flakes via controlled oxidationand will guide future development of antimonene-based2D materials and heterostructures.Computational resources used for the calculations wereprovided by the HPC of the Regional Computer Cen-tre Erlangen (RRZE). This work has been supportedby the Deutsche Forschungsgemeinschaft (DFG) withinthe CRC 953 (B13), by the European Union (ERC-2018-StG 804110-2D-PnictoChem to G.A.), and by theSpanish MINECO (Structures of Excellence Mar´ıa deMaeztu MDM-2015-0538). G.A. acknowledges supportby the Generalitat Valenciana (CIDEGENT/2018/001),and the DFG (FLAG-ERA AB694/2-1).See Supplemental Material at [URL will be insertedby publisher] for computational details, structural pa-rameters, images of all structures, snapshots of molec-ular dynamic calculations, vibrational modes, displace-ment patterns, phonon dispersion relations of the mono-layer structures, electronic band structures of few layersystems, and the atomic positions used for the phonondispersion calculations. Refs. [35–38] are cited in theSupplemental Material. - 101 G M S b O t y p e I E-EFermi (eV) MK M G K M ( a ) - 2- 1012( b ) K ' M ' ' K ' '
E-EFermi (eV) M G M 'K
S b O t y p e I I GGG
K ''K K ' G M M ' M '' - 2024( c ) G MK ''K ' K M G E-EFermi (eV)
M K
S b O t y p e I I M ' K ' M ' ' K ' ' G M 'M '' GG FIG. 4. Electronic band structures and density of states cal-culated using the hybrid functional HSE12 [34] and inclusionof spin-orbit interactions (SOI) for (a) type (I) Sb O , (b)type (II) Sb O , and (c) type (II) Sb O monolayers. Graydashed lines are the results without SOI and are almost con-gruent with the bands with SOI for (b) and (c). The zero ofenergy is set to the Fermi energy. ∗ stefan.wolff@fau.de[1] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang,Y. Zhang, S. V. Dubonos, I. V. Grigorieva, and A. A.Firsov, Science , 666 (2004).[2] A. K. Geim and K. S. Novoselov, Nat. Mater. , 183(2007). [3] G. Slotman, A. Rudenko, E. van Veen, M. I. Katsnelson,R. Rold´an, and S. Yuan, Phys. Rev. B , 155411 (2018).[4] C. C. Mayorga-Martinez, R. Gusmo, Z. Sofer, andM. Pumera, Angew. Chem. Int. Ed. , 134 (2019).[5] G. Pizzi, M. Gibertini, E. Dib, N. Marzari, G. Iannac-cone, and G. Fiori, Nat. Commun. , 12585 (2016).[6] J. Chen, Z. Yang, W. Zhou, H. Zou, M. Li, andF. Ouyang, Phys. Status Solidi , 1800038 (2018).[7] H. Zhang, J. Xiong, M. Ye, J. Li, X. Zhang, R. Quhe,Z. Song, J. Yang, Q. Zhang, B. Shi, J. Yan, W. Guo,J. Robertson, Y. Wang, F. Pan, and J. Lu, Phys. Rev.Applied , 064001 (2019).[8] G. Wang, S. Higgins, K. Wang, D. Bennett, N. Milosavl-jevic, J. J. Magan, S. Zhang, X. Zhang, J. Wang, andW. J. Blau, Appl. Opt. , E147 (2018).[9] M. Fortin-Deschˆenes, O. Waller, T. O. Mentes, A. Lo-catelli, S. Mukherjee, F. Genuzio, P. L. Levesque,A. H´ebert, R. Martel, and O. Moutanabbir, Nano Lett. , 4970 (2017).[10] X. Wu, Y. Shao, H. Liu, Z. Feng, Y.-L. Wang, J.-T. Sun,C. Liu, J.-O. Wang, Z.-L. Liu, S.-Y. Zhu, Y.-Q. Wang,S.-X. Du, Y.-G. Shi, K. Ibrahim, and H.-J. Gao, Adv.Mater. , 1605407 (2017).[11] H.-A. Chen, H. Sun, C.-R. Wu, Y.-X. Wang, P.-H. Lee,C.-W. Pao, and S.-Y. Lin, ACS Appl. Mater. & Inter-faces , 15058 (2018).[12] P. Ares, F. Aguilar-Galindo, D. Rodr´ıguez-San-Miguel,D. A. Aldave, S. D´ıaz-Tendero, M. Alcam´ı, F. Mart´ın,J. G´omez-Herrero, and F. Zamora, Adv. Mater. , 6332(2016).[13] C. Gibaja, D. Rodriguez-San-Miguel, P. Ares, J. G´omez-Herrero, M. Varela, R. Gillen, J. Maultzsch, F. Hauke,A. Hirsch, G. Abell´an, and F. Zamora, Angew. Chem.Int. Ed. , 14345 (2016).[14] P. Ares, J. J. Palacios, G. Abell´an, J. G´omez-Herrero,and F. Zamora, Adv. Mater. , 1703771 (2018).[15] V. Lloret, M. ´A. Rivero-Crespo, J. A. Vidal-Moya,S. Wild, A. Dom´enech-Carb´o, B. S. J. Heller, S. Shin, H.-P. Steinr¨uck, F. Maier, F. Hauke, M. Varela, A. Hirsch,A. Leyva-P´erez, and G. Abell´an, Nat. Commun. , 509(2019).[16] O. U. Akt¨urk, V. O. ¨Oz¸celik, and S. Ciraci, Phys. Rev.B , 235446 (2015).[17] S. K. Gupta, Y. Sonvane, G. Wang, and R. Pandey,Chem. Phys. Lett. , 169 (2015).[18] G. Wang, R. Pandey, and S. P. Karna, ACS Appl. Mater.& Interfaces , 11490 (2015).[19] S. Zhang, Z. Yan, Y. Li, Z. Chen, and H. Zeng, Angew.Chem. Int. Ed. , 3112 (2015).[20] J. Ji, X. Song, J. Liu, Z. Yan, C. Huo, S. Zhang, M. Su,L. Liao, W. Wang, Z. Ni, Y. Hao, and H. Zeng, Nat.Commun. , 13352 (2016).[21] Y. Wang, P. Huang, M. Ye, R. Quhe, Y. Pan, H. Zhang,H. Zhong, J. Shi, and J. Lu, Chem. Mater. , 2191(2017).[22] S. Zhang, W. Zhou, Y. Ma, J. Ji, B. Cai, S. A. Yang,Z. Zhu, Z. Chen, and H. Zeng, Nano Lett. , 3434(2017).[23] H. Liu, A. T. Neal, Z. Zhu, Z. Luo, X. Xu, D. Tom´anek,and P. D. Ye, ACS Nano , 4033 (2014).[24] A. Favron, E. Gaufr`es, F. Fossard, A.-L. Phaneuf-L’Heureux, N. Y.-W. Tang, P. L. L´evesque, A. Loiseau,R. Leonelli, S. Francoeur, and R. Martel, Nat. Mater. , 826 (2015).[25] L. Li, J. Kim, C. Jin, G. J. Ye, D. Y. Qiu, F. H. da Jor-nada, Z. Shi, L. Chen, Z. Zhang, F. Yang, K. Watan-abe, T. Taniguchi, W. Ren, S. G. Louie, X. H. Chen,Y. Zhang, and F. Wang, Nat. Nanotechnol. , 21(2016).[26] G. Abell´an, S. Wild, V. Lloret, N. Scheuschner, R. Gillen,U. Mundloch, J. Maultzsch, M. Varela, F. Hauke, andA. Hirsch, J. Am. Chem. Soc. , 10432 (2017).[27] R. S. Meng, M. Cai, J. K. Jiang, Q. H. Liang, X. Sun,Q. Yang, C. J. Tan, and X. P. Chen, IEEE ElectronDevice Lett. , 134 (2017).[28] A. A. Kistanov, Y. Cai, D. R. Kripalani, K. Zhou, S. V.Dmitriev, and Y.-W. Zhang, J. Mater. Chem. C , 4308(2018).[29] T. Zhang, Y. Wan, H. Xie, Y. Mu, P. Du, D. Wang,X. Wu, H. Ji, and L. Wan, J. Am. Chem. Soc. ,7561 (2018).[30] G. Abell´an, P. Ares, S. Wild, E. Nuin, C. Neiss, D. R.-S.Miguel, P. Segovia, C. Gibaja, E. G. Michel, A. Grling,F. Hauke, J. G´omez-Herrero, A. Hirsch, and F. Zamora,Angew. Chem. Int. Ed. , 14389 (2017).[31] J. P. Allen, J. J. Carey, A. Walsh, D. O. Scanlon, andG. W. Watson, J. Phys. Chem. C , 14759 (2013). [32] C. A. Cody, L. DiCarlo, and R. K. Darlington, Inorg.Chem. , 1572 (1979).[33] M. Assebban, C. Gibaja, M. Fickert, I. Torres, E. Weinre-ich, S. Wolff, R. Gillen, J. Maultzsch, M. Varela, S. T. J.Rong, K. P. Loh, E. G. Michael, F. Zamora, andG. Abell´an, (to be published).[34] J. E. Moussa, P. A. Schultz, and J. R. Chelikowsky, J.Chem. Phys. , 204117 (2012).[35] P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car,C. Cavazzoni, D. Ceresoli, G. L. Chiarotti, M. Cococ-cioni, I. Dabo, A. Dal Corso, S. de Gironcoli, S. Fabris,G. Fratesi, R. Gebauer, U. Gerstmann, C. Gougoussis,A. Kokalj, M. Lazzeri, L. Martin-Samos, N. Marzari,F. Mauri, R. Mazzarello, S. Paolini, A. Pasquarello,L. Paulatto, C. Sbraccia, S. Scandolo, G. Sclauzero, A. P.Seitsonen, A. Smogunov, P. Umari, and R. M. Wentz-covitch, J. Phys. Condens. Matter , 395502 (2009).[36] K. F. Garrity, J. W. Bennett, K. M. Rabe, and D. Van-derbilt, Comput. Mater. Sic. , 446 (2014).[37] M. van Setten, M. Giantomassi, E. Bousquet, M. Ver-straete, D. Hamann, X. Gonze, and G.-M. Rignanese,Comput. Phys. Commun. , 39 (2018).[38] J. M. Soler, E. Artacho, J. D. Gale, A. Garc´ıa, J. Jun-quera, P. Ordej´on, and D. S´anchez-Portal, J. Phys. Con-dens. Matter14