# Ab Initio Prediction of Boron Compounds Arising from Borozene: Structural and Electronic Properties

- G Forte
^{1}Email author, - A La Magna
^{2}, - I Deretzis
^{2, 3}and - R Pucci
^{4}

**5**:158

**DOI: **10.1007/s11671-009-9458-8

© to the authors 2009

**Received: **15 September 2009

**Accepted: **2 October 2009

**Published: **21 October 2009

## Abstract

Structure and electronic properties of two unusual boron clusters obtained by fusion of borozene rings have been studied by means of first principles calculations based on the generalized-gradient approximation of the density functional theory. Moreover, a semiempirical tight-binding model has been appropriately calibrated for transport calculations on these clusters. Results show that the pure boron clusters are topologically planar and characterized by (3c–2e) bonds, which can explain, together with the aromaticity (estimated by means of NICS), the remarkable cohesive energy values obtained. Such feature makes these systems competitive with the most stable boron clusters to date. The energy gap values indicate that these clusters possess a semiconducting character, while when the larger system is considered, zero-values of the density of states are found exclusively within the HOMO–LUMO gap. Electron transport calculations within the Landauer formalism confirm these indications, showing semiconductor-like low bias differential conductance for these structures. Differences and similarities with carbon clusters are highlighted in the discussion.

### Keywords

Boron clusters Borozene DFT NICS Transport## Introduction

Boron is the first element in group IIIA of the periodic table, presents the external electronic configuration *s*
^{2}
*p*
^{1} and possesses a variety of compounds second only to carbon. In several boron compounds [1, 2] the existence of multicenter bonds has been discovered, which arise from the electron deficiency of this element. Moreover, boron has a special place among the elements of the periodic table because of the wide variety of crystalline structure forms, i.e., polymorphism, which include nanotubes [3], nanoribbons [4], and nanoclusters [5, 6].

The interest in boron-based nanostructures has recently increased due to new studies of both the synthesis of single walled boron nanotubes (SWBNTs) and the prediction of ballistic conduction in SWBNTs [3, 6]; these findings, together with the properties that all boron nanotubes (a) are predicted to be metallic [7] and (b) are superconducting at low temperatures [8, 9], promoted their prospective applications in fabrication of novel electronic devices. The most stable boron structure is the α-rhombohedral bulk where boron icosahedra are centered on the edges of a rhombohedral unit cell [10].

Unlike the bulk boron compounds, boron clusters B_{
n
} (*n* < 20) are quasiplanar, or even planar, with a symmetrical bond distribution, aromatic [11–13] and their existence is confirmed by the experiment [14]. From the Aufbau principle postulated by Boustani [5] it follows that these quasiplanar isomers are more stable than their icosahedral counterparts. Recently Szwacki et al. [15] have predicted the existence of a planar and aromatic boron compound, named borozene, which has strong similitudes with benzene.

_{60}H

_{12}and B

_{228}H

_{24}, in which the molecule of borozene can be considered as the building block, as the benzene ring represents the embryo of compounds such Coronene, Coronene 19 etc. In general, we will refer to these compounds as boron clusters whose external dangling bonds are saturated by hydrogen atoms; they are obtained by fusing together the outer boron pairs of borozene molecules bonded to a hydrogen atom, see Fig. 1. We have organized the rest of the paper as follows: the computational methods adopted are presented in “Computational Method” section. In “Results” section we present and discuss our results in comparison also with carbon compounds, and, finally, we give a summary in “Conclusions” section.

## Computational Method

The molecule B_{60}H_{12} here considered was built by fusing six borozene rings, for this reason it can be considered as the boron counterpart of coronene, whereas the structure of B_{228}H_{24} was obtained by surrounding B_{60}H_{12} with one series of borozene rings, therefore this cluster is constituted by a total of 24 borozene rings. A first optimization energy procedure was performed in the framework of the molecular mechanics approximation applying the CVFF Force Field [16, 17] which is enclosed in the Materials Studio package [18]. The geometries obtained were fully optimized at a B3LYP/STO-3G [19–23] and B3LYP/6-311G [24, 25] level by using the quadratically convergent Self Consistent Field procedure [26].

In detail, due to the large size, the optimizations of B_{228}H_{24} were carried out by means of the minimal basis set STO-3G whereas the more extended Pople basis set 6-311G was used in the optimizations of B_{60}H_{12}. In order to estimate the degree of aromaticity, the calculation of Nuclear Independent Chemical Shifts [27] on the plane of the aromatic system (NICS0) was computed using the Gaussian 03 package [28], applying the GIAO method [29, 30]. To obtain the contour plot of NICS, ghost atoms were placed on the plane of the molecule with a step size of about 1 Å.

Finally, electronic transport has been evaluated in the framework of the Nonequilibrium Green’s Function theory using a Landauer expression for the calculation of the current–voltage (I–V) characteristics [31]. Consistently to the electron structure findings we assume that only *p*
_{z} orbitals contribute to the low bias transport along the molecules and that the Fermi energy is at the center of HOMO–LUMO gap. The molecular device configuration considered consists of two vertical gold leads in contact with the horizontal molecules forming ideal bonds with the boron atoms indicated in Fig. 1.

## Results

### Structural Properties

The analysis of the smaller cluster, henceforth named **B6**, was performed by using both basis sets mentioned above in order to make a consistent comparison with B_{228}H_{24}, henceforth named **B24**, analyzed only with the minimal basis set. We point out that the results obtained in the two cases are qualitatively equivalent, for this reason, unless specified, the data shown below are referred to the minimal basis set. As far as the boron–boron bond length is concerned, a shorter value has been found in the STO-3G optimized structure, in particular, taking into account cluster **B6**, the average value of this parameter is calculated to be, respectively, 1.649 and 1.638 Å for 6-311G and STO-3G basis set, while a bond length average of 1.629 Å is obtained for the cluster **B24**.

**B6**and 6.449 eV for

**B24**. These values were calculated from

In the expressions above *n* is the number of boron atoms and the value of the B–H bond energy is calculated in the first approximation as 1/3 of the binding energy value of BH_{3}. Structural parameters evaluated are competitive, in terms of stability, with the more stable flat two-dimensional structures considered up to date [32, 33]. It is well known that Boron has a variety of compounds containing multicenter bonds, in particular the three-center, two-electron (3c, 2e) bond is present in molecules such diborane [34], boron clusters [32] and boron sheets [35].

Previous works have shown that (3c, 2e) bonds preclude the formation of boron rings in boron clusters [6, 36], whereas, on the other hand, more recently the three-center bonding has been proposed to explain the stability of boron fullerenes [34, 37]. This peculiar feature is also seen for **B6** and **B24,** while it is not present in Coronene and its larger clusters such Coronene 19, Coronene 37 and Coronene 61. Hence it is logical to assume that, as for boron fullerenes, it plays a pivotal role in maintaining a two-dimensional stable planar structure.

*P*is the total density matrix and

*S*is the overlap matrix. The bond order indices of three-center bonds are positive with a theoretical maximum of ≈0.296. In Fig. 2, we report the more relevant values of

*I*for both clusters, in general one can affirm that each boron is involved at least in two different three-center bonds, i.e., each boron is directly linked at least with four boron atoms.

### Electronic Properties

*p*

_{z}atomic orbitals [11, 12, 14]. As it will be discussed below,

**B6**and

**B24**show these features; before analyzing in detail we underline that a more extended electronic delocalization gives rise to a smaller GAP in carbon clusters [41], described as the HOMO–LUMO energy difference. In accordance with these calculations, the GAP values obtained for

**B6**and

**B24**are 1.33 and 1.17 eV, respectively; their density of states (DOS) as a function of energy (eV) is shown in Fig. 3. At energy

*E*the density of states is written as

*i*goes over all energy levels and δ is the Delta function.

**B24**has a larger density of states, while differently from cluster

**B6**, it shows a zero value of the DOS only within the HOMO–LUMO gap; (b), the composition of the molecular orbitals, calculated by means of Mulliken Population Analysis, reported in Table 1, clearly highlights how this contributes to the HOMO of both clusters, shown in the insets of Fig. 3 and to their nearer molecular orbitals, are mainly due to the

*p*

_{z}atomic orbitals, confirming the stabilizing effect of π-delocalization. This result is in agreement with the one evidenced in Coronene, whose HOMO and DOS are reported, in Fig. 3; (c), as already observed for the set of Carbon clusters previously studied [41], the peaks near the HOMO energy can be joined by an almost straight line, reproducing the linear dependence shown by the infinite system near the Fermi level.

Energies and percentual contributes of *p*
_{z} orbitals to the composition of HOMO, LUMO and their nearest MOs

B6 | B24 | |||
---|---|---|---|---|

| % |
| % | |

HOMO − 3 | −4.73 | 0.36 | −3.67 | 100.00 |

HOMO − 2 | −4.52 | 99.87 | −3.67 | 99.94 |

HOMO − 1 | −4.52 | 99.80 | −3.43 | 99.99 |

HOMO | −3.56 | 99.97 | −3.28 | 99.99 |

LUMO | −2.23 | 99.81 | −2.12 | 100.00 |

LUMO + 1 | −1.50 | 91.19 | −2.06 | 100.00 |

LUMO + 2 | −1.36 | 95.71 | −1.89 | 99.93 |

LUMO + 3 | −1.36 | 99.94 | −1.89 | 99.94 |

**B6**and

**B24**, we find it necessary to investigate this aspect. Figure 4 shows the plot of the nucleus independent chemical shift (NICS), which represents the magnetic criterion to evaluate the ring current for cluster

**B24**.

The symmetry of the plots reflects the assumed symmetry in the device configuration. A semiconductor-like behavior is evidenced in both structures. However, the zero bias differential conductance is one order of magnitude higher for the **B24** cluster, this is due not only to the lower HOMO–LUMO gap but also to the larger value of the DOS. Delocalized (along the cluster) *p*
_{z} molecular orbitals allow an efficient charge transport through the cluster for larger bias (of the order of the gap) and a diode-like characteristic can be observed.

## Conclusions

In this work first principles and semiempirical calculations were carried out to investigate both structural and electronic properties of two clusters obtained by condensation of 6 and 24 borozene molecules, considered as the analogs of Coronene and Coronene 19. Calculations predict a planar geometry for the pure clusters. Both (3c–2e) bonds and wide regions of aromaticity contribute to this stabilization, with cohesive energy values that are comparable with the most stable boron clusters considered to date.

Due to the high connectivity among boron atoms, we hypothesize that planar geometry could be compromised when impurities are introduced. Density of states spectra evidence a small gap, which decreases by increasing the cluster size, suggesting, at variance with carbon clusters, a semiconducting character in small sized clusters; furthermore the population analysis shows that the main contribution to the molecular orbitals near the GAP is due to π-bonds which derive from *p*
_{z} orbitals. Calculated low bias differential conductance for these structures confirms this semiconductor-like character.

## Declarations

### Acknowledgment

G. Forte wishes to thank the Consorzio Interuniversitario Cineca for the computational support.

## Authors’ Affiliations

## References

- Janotti A, Van de Walle CG:
*Nat. Phys.*. 2007,**6:**44. COI number [1:CAS:528:DC%2BD2sXot1On]Google Scholar - Li X,
*et al*.:*Science*. 2007,**315:**356. COI number [1:CAS:528:DC%2BD2sXmt1KgsQ%3D%3D]; Bibcode number [2007Sci...315..356L] 10.1126/science.1133767View ArticleGoogle Scholar - Ciuparu D, Klie FR, Zhu Y, Pfefferle LJ:
*Phys. Chem. B*. 2004,**108:**3967. COI number [1:CAS:528:DC%2BD2cXhvFWltbs%3D] 10.1021/jp049301bView ArticleGoogle Scholar - Xu TT, Zheng J, Wu N, Nicholls WA, Roth RJ, Dikin AD, Ruoff RS:
*Nano Lett.*. 2004,**4:**963. COI number [1:CAS:528:DC%2BD2cXis1Citrw%3D]; Bibcode number [2004NanoL...4..963X] 10.1021/nl0498785View ArticleGoogle Scholar - Boustani I:
*Phys. Rev. B*. 1997,**55:**16426. COI number [1:CAS:528:DyaK2sXkt1yls7Y%3D]; Bibcode number [1997PhRvB..5516426B] 10.1103/PhysRevB.55.16426View ArticleGoogle Scholar - Lau KC, Pandey R:
*Comput. Lett.*. 2005,**1**(Special Issue: Clusters):259. doi:10.1163/157404005776611312View ArticleGoogle Scholar - Boustani I, Quandt A:
*Europhys. Lett.*. 1997,**39:**527. COI number [1:CAS:528:DyaK2sXmtV2hsL8%3D]; Bibcode number [1997EL.....39..527B] 10.1209/epl/i1997-00388-9View ArticleGoogle Scholar - Kociak M,
*et al*.:*Phys. Rev. Lett.*. 2001,**86:**2416. COI number [1:CAS:528:DC%2BD3MXhvFSjt70%3D]; Bibcode number [2001PhRvL..86.2416K] 10.1103/PhysRevLett.86.2416View ArticleGoogle Scholar - Takesue I,
*et al*.:*Phys. Rev. Lett.*. 2006,**96:**057001. COI number [1:STN:280:DC%2BD287gtF2msA%3D%3D]; Bibcode number [2006PhRvL..96e7001T] 10.1103/PhysRevLett.96.057001View ArticleGoogle Scholar - Fujimori M, Nakata T, Nakayama T, Nishibori E, Kimura K, Takata M, Sahata M:
*Phys. Rev. Lett.*. 1999,**82:**4452. COI number [1:CAS:528:DyaK1MXjtlelsbw%3D]; Bibcode number [1999PhRvL..82.4452F] 10.1103/PhysRevLett.82.4452View ArticleGoogle Scholar - Boustani I:
*Int. J. Quantum Chem.*. 1994,**52:**1081. COI number [1:CAS:528:DyaK2cXmvVKqs7o%3D] 10.1002/qua.560520432View ArticleGoogle Scholar - Aihara J, Kanno H, Ishida T:
*J. Am. Chem. Soc.*. 2005,**127:**13324. COI number [1:CAS:528:DC%2BD2MXpsVGhtrg%3D] 10.1021/ja053171iView ArticleGoogle Scholar - Sergeeva AP, Zubarev DY, Zhai HJ, Boldyrev AI, Wang LS:
*J. Am. Chem. Soc.*. 2008,**130:**7244. COI number [1:CAS:528:DC%2BD1cXmtVGitbs%3D] 10.1021/ja802494zView ArticleGoogle Scholar - Zhai HJ, Kiran B, Li J, Wang LS:
*Nat. Mater.*. 2003,**2:**827. COI number [1:CAS:528:DC%2BD3sXptlemtrY%3D]; Bibcode number [2003NatMa...2..827Z] 10.1038/nmat1012View ArticleGoogle Scholar - Szwacki NG, Weber V, Tymczak CJ:
*Nanoscale Res. Lett.*. 2009,**4:**1085. Bibcode number [2009NRL.....4.1085S] Bibcode number [2009NRL.....4.1085S] 10.1007/s11671-009-9362-2View ArticleGoogle Scholar - Hagler AT, Dauber P, Lifson S:
*J. Am. Chem. Soc.*. 1979,**101:**5131. COI number [1:CAS:528:DyaL3cXlvV2r] 10.1021/ja00512a003View ArticleGoogle Scholar - Hagler AT, Dauber P, Lifson S:
*J. Am. Chem. Soc.*. 1979,**101:**5122. COI number [1:CAS:528:DyaE1MXmtFSks7Y%3D] 10.1021/ja00512a002View ArticleGoogle Scholar - Accelrys Inc.:
*MS Modeling Getting Started*. San Diego: Accelrys Inc.; 2003.Google Scholar - Becke AD:
*Phys. Rev. A*. 1988,**38:**3098. COI number [1:CAS:528:DyaL1cXmtlOhsLo%3D]; Bibcode number [1988PhRvA..38.3098B] 10.1103/PhysRevA.38.3098View ArticleGoogle Scholar - Lee C, Yang W, Parr RG:
*Phys. Rev. B*. 1988,**37:**785. COI number [1:CAS:528:DyaL1cXktFWrtbw%3D]; Bibcode number [1988PhRvB..37..785L] 10.1103/PhysRevB.37.785View ArticleGoogle Scholar - Becke AD:
*J. Chem. Phys.*. 1993,**98:**5648. COI number [1:CAS:528:DyaK3sXisVWgtrw%3D]; Bibcode number [1993JChPh..98.5648B] 10.1063/1.464913View ArticleGoogle Scholar - Hehre WJ, Stewart RF, Pople JA:
*J. Chem. Phys.*. 1969,**51:**2657. COI number [1:CAS:528:DyaF1MXltVGnsrY%3D]; Bibcode number [1969JChPh..51.2657H] 10.1063/1.1672392View ArticleGoogle Scholar - Collins JB, von Ragué Schleyer P, Binkley JS, Pople JA:
*J. Chem. Phys.*. 1976,**64:**5142. COI number [1:CAS:528:DyaE28XltlSrs78%3D]; Bibcode number [1976JChPh..64.5142C] 10.1063/1.432189View ArticleGoogle Scholar - McLean AD, Chandler GS:
*J. Chem. Phys.*. 1980,**72:**5639. COI number [1:CAS:528:DyaL3cXksFCnu7c%3D]; Bibcode number [1980JChPh..72.5639M] 10.1063/1.438980View ArticleGoogle Scholar - Krishnan R, Binkley JS, Seeger R, Pople JA:
*J. Chem. Phys.*. 1980,**72:**650. COI number [1:CAS:528:DyaL3cXpvFyitA%3D%3D]; Bibcode number [1980JChPh..72..650K] 10.1063/1.438955View ArticleGoogle Scholar - Bacskay GB:
*Chem. Phys.*. 1981,**61:**385. COI number [1:CAS:528:DyaL3MXmt1Gqtb8%3D]; Bibcode number [1981CP.....61..385B] 10.1016/0301-0104(81)85156-7View ArticleGoogle Scholar - von Rague Schleyer P, Maerker C, Dransfeld A, Jiao H, van Eikema Hommes NJR:
*J. Am. Chem. Soc.*. 1996,**16:**6317. 10.1021/ja960582dView ArticleGoogle Scholar - Gaussian 03, Revision D.02, Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Montgomery JA Jr, Vreven T, Kudin KN, Burant JC, Millam JM, Iyengar SS, Tomasi J, Barone V. Gaussian, Inc., Wallingford CT; 2004.Google Scholar
- McWeeny R:
*Phys. Rev.*. 1962,**126:**1028. Bibcode number [1962PhRv..126.1028M] Bibcode number [1962PhRv..126.1028M] 10.1103/PhysRev.126.1028View ArticleGoogle Scholar - Ditchfield R:
*Mol. Phys.*. 1974,**27:**789. COI number [1:CAS:528:DyaE2cXkvVektrk%3D]; Bibcode number [1974MolPh..27..789D] 10.1080/00268977400100711View ArticleGoogle Scholar - Deretzis I, La Magna A:
*Nanotechnology*. 2006,**17:**5063. COI number [1:CAS:528:DC%2BD28Xht1OjtLvP]; Bibcode number [2006Nanot..17.5063D] 10.1088/0957-4484/17/20/005View ArticleGoogle Scholar - Ismail-Beigi S, Tang H:
*Phys. Rev. Lett.*. 2007,**99:**115501. Bibcode number [2007PhRvL..99k5501T] Bibcode number [2007PhRvL..99k5501T] 10.1103/PhysRevLett.99.115501View ArticleGoogle Scholar - Yang X, Ding Y, Ni J:
*Phys. Rev. B*. 2008,**77:**041402. Bibcode number [2008PhRvB..77d1402Y] Bibcode number [2008PhRvB..77d1402Y] 10.1103/PhysRevB.77.041402View ArticleGoogle Scholar - Sannigrahi AB, Kar T:
*Chem. Phys. Lett.*. 1990,**173:**569. COI number [1:CAS:528:DyaK3cXmtlygtLk%3D]; Bibcode number [1990CPL...173..569S] 10.1016/0009-2614(90)87254-OView ArticleGoogle Scholar - Kunstmann J, Quandt A:
*Phys. Rev. B*. 2006,**74:**035413. Bibcode number [2006PhRvB..74c5413K] Bibcode number [2006PhRvB..74c5413K] 10.1103/PhysRevB.74.035413View ArticleGoogle Scholar - Chacko S, Kanhere GD, Boustani I:
*Phys. Rev. B*. 2003,**68:**035414. Bibcode number [2003PhRvB..68c5414C] Bibcode number [2003PhRvB..68c5414C] 10.1103/PhysRevB.68.035414View ArticleGoogle Scholar - Szwacki NG, Sadrzadeh A, Yakobson BI:
*Phys. Rev. Lett.*. 2007,**98:**166804. Bibcode number [2007PhRvL..98p6804G] Bibcode number [2007PhRvL..98p6804G] 10.1103/PhysRevLett.98.166804View ArticleGoogle Scholar - Kar T, Marcos ES:
*Chem. Phys. Lett.*. 1992,**192:**14. COI number [1:CAS:528:DyaK38XktlCjtrs%3D]; Bibcode number [1992CPL...192...14K] 10.1016/0009-2614(92)85420-FView ArticleGoogle Scholar - deGiambiagi MS, Giambiagi M, Fortes MD:
*J. Mol. Struct. Theochem*. 1997,**391:**141. COI number [1:CAS:528:DyaK2sXivFajt7g%3D] 10.1016/S0166-1280(96)04815-4View ArticleGoogle Scholar - Ponec R, Mayer I:
*J. Phys. Chem. A*. 1997,**101:**1738. COI number [1:CAS:528:DyaK2sXhsFWnsLw%3D] 10.1021/jp962510eView ArticleGoogle Scholar - Forte G, Grassi A, Lombardo GM, La Magna A, Angilella GGN, Pucci R, Vilardi R:
*Phys. Lett. A*. 2008,**372:**6168. COI number [1:CAS:528:DC%2BD1cXhtFWnsLzI]; Bibcode number [2008PhLA..372.6168F] 10.1016/j.physleta.2008.08.014View ArticleGoogle Scholar