Theoretical Study of Sarin Adsorption On

Chemical Physics Letters 738 (2020) 136816

Contents lists available at ScienceDirect

Chemical Physics Letters

journal homepage: www.elsevier.com/locate/cplett

Research paper Theoretical study of sarin adsorption on (12,0) boron nitride nanotube doped with silicon atoms T ⁎ ⁎ Jeziel Rodrigues dos Santosa, , Elson Longo da Silvab, Osmair Vital de Oliveirac, , José Divino dos Santosa a Universidade Estadual de Goiás, Campus Anápolis, CEP: 75.132-903 GO, Brazil b INCTMN, LIEC, Departamento de Química da Universidade Federal de São Carlos, CEP: 13.565-905 São Carlos, SP, Brazil c Instituto Federal de Educação, Ciência e Tecnologia de São Paulo, Campus Catanduva, CEP: 15.808-305 Catanduva, SP, Brazil

HIGHLIGHTS

• DFT method was used to study the adsorption of nerve agent sarin by BNNT. • Electronic properties of pristine BNNT are improved by Si impurity atoms. • The adsorption of sarin by Si-doped BNNT is highest favorable than the pure BNNT. • Si-doped BNNT can be a new gas sensor for sarin gas detection and its derivatives.

ARTICLE INFO ABSTRACT

Keywords: Sarin gas is one of the most lethal nerve agent used in chemical warfare, which its detection is import to prevent Nerve agent sarin a chemical attack and to identify a contamination area. Herein, density functional theory was used to investigate Gas sensor the (12,0) boron nitride nanotube (BNNT) and Si–doped BNNT as possible candidates to sarin detection. The Si- Boron nitride nanotube atoms doped improve the electronic properties of nanotubes by altering the electrostatic potential, HOMO and DFT LUMO energies. Based in the adsorption energies and the conductivity increased to ~33 and 350%, respectively, for Si- and 2Si-BNNT imply that they can be used for sarin detection.

1. Introduction War II, Iraq/Iran War (1981–1989), in the Tokyo subway attack (1995), in the Syrian Civil War (2012), and recently in the Syrian attack in Neurotoxic chemical agents are a class of substances which direct 2017. Therefore, the detection of sarin can be useful for military and and indirectly perturb the Human and animal nervous system by acting civil defense to prevent chemical attack using this neurotoxic agent and in the neural cell or in the metabolic process of this system [1]. These others. In this way, different methodologies have been used to detect substances are chemically classified into the organophosphorus group chemical agents like infrared spectroscopy [3], mobility spectroscopy and they are organic compounds degradable. Moreover, these com- [4], calorimetric [5], surface acoustic waver sensors [6], electro- pounds are very harmful and/or lethal for Human, consequently they chemical detectors [7], carbon nanotube (CNT) chemical sensor [8]. has been used as high-impact military artifice in called chemical war- Among them, chemical sensors based in CNT have an advantage to fare. These agents inhibit the acetylcholinesterase irreversibly leading produce portable sensing method and they can be coupled with elec- the loss of the control of central nervous system. Among the neurotoxic trical devices as conductometric, electrochemical, and others [8]. agents, the most used in the chemical warfare are belong the G-series as However, CNTs offer non-specific sensing responses for nerve agent and tabun, sarin and soman gases, which these two last were developed their mimics. Another disadvantage, the electronic properties of CNTs during the World War II [2]. In this series, they are absorbed through are high dependent of their chirality [9]. So, the boron nitride nanotube the lungs causing seizures, loss of body control, paralyses muscles in- (BNNT) appears as an excellent candidate to substitute the CNT because cluding heart and diaphragm. Sarin gas is the most neurotoxic agent their electronic properties are independent of their diameter and chir- known and largely used in chemical attack. It was used in the World ality [10–12]. For instance, the band gap of the BNNT is 5–6eV[13,14]

⁎ Corresponding authors. E-mail addresses: [email protected] (J.R. dos Santos), [email protected] (O.V. de Oliveira). https://doi.org/10.1016/j.cplett.2019.136816 Received 14 August 2019; Received in revised form 13 September 2019; Accepted 1 October 2019 Available online 09 October 2019 0009-2614/ © 2019 Elsevier B.V. All rights reserved. J.R. dos Santos, et al. Chemical Physics Letters 738 (2020) 136816 independently of these properties. Therefore, the BNNT is an excellent optimized structures with minimum energy obtained at DFT method. insulator, contrary to CNTs which are semimetallic and semiconductor Overall, as can see in the Fig. 1, it was not observed significant material [15]. Moreover, BNNTs are stable in oxidation resistance and structural change in the BNNT doped with Si atoms. The main differ- it has high thermal resistance above 900 °C. Experimentally, BNNT was ence noted is attributed to the Si atoms that are outside of 0.09 nm from synthesized by first time in 1995 thought arc discharge [16], and ac- the B95N96SiH24 and B94N96Si2H24 surface. This is due the Van der tually others methods like chemical vapor deposition [17], laser abla- Waals radius (0.210 nm) of the Si atom to be slight high than the boron tion [18] and thermal plasma jet [19] are used to their synthesis. atom (0.192 nm). Wang [43] obtained this same geometrical distortion BNNTs have been applied for polymer composite reinforcement [20], in Si-doped BNNT using periodic DFT calculations. Regarding the for piezo actuators [21], drug carrier [22],infield emission technology structural change of nanotubes, the same pattern was observed for the [23,24], sensing [25,26], etc. Recently, the possible use of BNNTs as gas complexes, implying that the sarin adsorbed preserve the BNNT, sensor has attracted the attention of many researchers. In the manner B95N96SiH24 and B94N96Si2H24 structures. These observations are in that, theoretical methods were used to study the adsorption of oxazole concordance with the lowest root mean square deviation (RMSD) va- and isoxazole [27], hydrogen halides [28], hydrogen cyanide [29], lues (< 0.01 nm) calculated from the superposition between Si–doped carbon monoxide [30], cyanogen chloride [31] and ammonia [32] in BNNT and pristine BNNT structure. For sarin–BNNT, it was observed a BNNTs. Regarding the chemical agents, the adsorption of soman and shortest distance between the boron and oxygen atom (sarin) with chlorosoman by (8,0) BNNT [33], and a sarin derivative by (6,0) BNNT distance of 0.294 nm. Whereas, for the sarin–B95N96SiH24 and sar- [34] were studied at theoretical level. In both studies, the authors find in–B94N96Si2H24, this distance was, respectively, 0.165 and 0.178 nm. that these nerve agents interact weakly with pristine BNNT. Therefore, From the NBO analysis, it was confirmed that the oxygen bound to improve the BNNT electronic sensitivity, herein a large zigzag (12,0) covalently with Si atom of the B95N96SiH24 and partially with BNNT and it doped with silicon atoms were studied at density func- B94N96Si2H24 compounds with bond order of 1.11 and 0.63, respec- tional theory (DFT) with intention to enhance the sarin-BNNT inter- tively. Contrary, there is not formation of covalent bond between BNNT actions. For instance, Si-doped BNNT was forecast by Guerini [35] from and sarin, keeping bond order of 0.04. Therefore, the OeSi bond ob- theoretical methods, and posterior it was synthesized in 2009 by Cho tained in our calculations is in good agreement with experimental data [36] using thermal chemical vapor deposition. In both studies, it was (0.161 nm). For all nanotubes, the boron-nitrogen bond is confirmed the improvement of BNNT reactivity up replacing B by the Si 0.144–0.145 nm in excellent agreement with the hexagonal crystal atom. structure of the boron nitride (0.145 nm). In the next section, the electronic structures for pure compounds and for complexes are pre- 2. Methodology sented and discussed to understand the energetic process involved.

The initial structure of the zigzag (12,0) BNNT formed by 96 boron 3.2. Electronic structure and 69 nitrogen atoms was built using a script written in-house. The end atoms were saturated by 24 hydrogen atoms to avoid the boundary The electronic properties were used to clarify the chemisorption – effects, forming the B96N96H24 compound. This nanotube model has 9.4 process of the sarin by BNNT and Si atoms doped BNNT. Initially, it is and 15.4 Å of diameter and length, respectively. For the complexes, the important evaluate the Si–doped BNNT stability, which this can be fi initial con gurations were built by positioning sarin (C4H10FO2P) mo- addressed by the formation energy (ΔEformation) using the equation, lecule in different regions around the BNNT and Si atoms–doped BNNT ΔE(EE)(EEformation=+−+ doped - BNNTnn B BNNT Si) (2) surfaces. The sarin, the pristine BNNT and it doped with Si atom

(B95N96SiH24 in doublet state and B94N96Si2H24 in singlet state), and the where Edoped-BNNT is the total energy of one or two silicon doped with – – – complexes (sarin BNNT, sarin B95N96SiH24 and sarin B94N96Si2H24) BNNT, n is the number of Si or B atoms substituted, nEB is the total were optimized initially with the semiempirical Hamiltonian PM7 [37] energy B atom, EBNNT is the total energy of the pristine BNNT and nESi is using MOPAC2016 program [38]. Posteriorly, the structures with the total energy of Si atom. The ΔEformation values for B95N96SiH24 and minimum energies were re-optimized with DFT calculations con- B94N96Si2H24 are 3.46 and 8.30 eV, respectively. For 1Si–doped BNNT, sidering the B3LYP hybrid functional [39] with 6-31G(d) basis set. This the ΔEformation has low value compared to the reported in the literature procedure was chosen to balance the computational cost and quality of (4.06 eV) using DFT//B3LYP/ECP method [44], but these values are in the results. Moreover, some works reported in the literature show that good agreement considering the specificity and limitation of each basis this theory level is sufficiently to describe pristine BNNT [28,34,40]. set used. Table 1 summarizes some important electronic properties The stationary points were characterized as a point of minimum energy obtained and discussed here. using the harmonic vibrational states, which it was not observed ne- The reactivity parameters were based in the highest occupied mo- gative frequencies. All DFT calculations were carried out in vacuum lecular orbital (HOMO, εHOMO) and lowest unoccupied molecular or- using the Gaussian 09 package [41]. The natural bond orbital (NBO) bital (LUMO, εLUMO) energies. For instance, according to the Koopmans’ method [42] was used to compute the atomic charges. The adsorption theorem [45] the ionization potential (IP) is described by the negative energy (Ead) was calculated using the equation, of the HOMO energy, and the electronic affinity (AE) can be approxi-

mated by the LUMO energy with opposite sign. Energy gap (Eg) is an- Ead=−E(EE) complex nanotube + sarin (1) other important property obtained from these orbital energies by the where Ecomplex is the total energy of complexes (sarin–BNNT, sar- difference between HOMO and LUMO energies in module. The large – – in B95N96SiH24 and sarin B94N96Si2H24), Enanotube is the total energy of and small Eg values imply a high stability and reactivity, respectively, of BNNT, B95N96SiH24 or B94N96Si2H24, and Esarin is the total energy of the a compound in chemical reactions. Therefore, the sarin is the most sarin. stable compound studied here with Eg value of 6.46 eV with highest IP and with lowest AE. The lowest reactivity of the sarin predicted here is 3. Results and discussion in agreement with a recent work reported by our research group using DFT//B3LYP/6-31(d,p) [46]. In the same way, the large energy gap 3.1. Structural analysis (6.05 eV) obtained here show that BNNT is an insulator in agreement with literature [13,14], which they reported a band gap of 5–6 eV. In the present work, DFT method was used to study the pristine Moreover, the energy gap obtained for others authors [28,32–33,40,44] BNNT and it doped with Si atoms with intention to obtain a new gas using different theoretical methods and different size of BNNT are in sensor to identify the presence of the nerve agent sarin. Fig. 1 shows the good accordance with our predicted Eg value. Therefore, these

2 J.R. dos Santos, et al. Chemical Physics Letters 738 (2020) 136816

Fig. 1. Structures of the compounds optimized with DFT//B3LYP/6-31G(d) method. Nitrogen in blue, oxygen in red, silicon in yellow, carbon in gray, hydrogen in white, phosphorous in or- ange, ﬂuorine in green and boron in salmon colors. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.)

Table 1 Fig. 2 shows lowest density of the HOMO energies for all nanotubes, Electronic properties obtained from DFT//B3LYP/6-31G(d) calculations. which the highest density is observed on Si atoms (Si–atoms doped

a BNNT) and, for pure BNNT, there small density energy on the N atoms. Compounds εHOMO (eV) εLUMO (eV) Eg Ead (kcal/mol) On the other hand, the BNNT and B95N96SiH24 present high density in sarin –7.93 1.47 6.46 – LUMO energy centered in the boron-nitrogen bonds, while for BNNT –6.43 –0.38 6.05 – B94N96Si2H24, the LUMO energy is condensed only in the Si atoms. For B N SiH –5.28 –0.37 4.91 – 95 96 24 nerve agent sarin, the LUMO is distributed along the molecule and the B94N96Si2H24 –4.06 –3.82 0.24 – sarin–BNNT –6.34 –0.28 6.06 –3.79 HOMO energy is centered in the oxygen atom. Therefore, the highest

sarin–B95N96SiH24 –4.77 –0.12 4.65 –13.44 reactivity of Si atoms-doped BNNT can be attributed the generation of – – – – sarin B94N96Si2H24 2.96 1.23 1.74 39.43 specific active sites, which they are suitable to react with others chemical species. a |ε – ε | in eV. HOMO LUMO From MEP analysis, it is interesting to point that the 2Si–doped BNNT does not change the electrostatic surface of the BNNT as can see experimental and theoretical studies validate our BNNT model, and in Fig. 2. Contrary, it was observed a significant change caused by one likewise the theory level used is accurate enough to describe the elec- Si impurity in the BNNT due the increasing of the atomic charges of B tronic properties of the present pristine BNNT. and N atoms. For instance, the average of NBO atomic charges obtained For Si–doped BNNT, it was observed an increase in the reactivity for B and N are, respectively, 1.18 electron (e) and −1.18e for pristine with E values of 4.91 and 0.24 eV for B N SiH and B N Si H , g 95 96 24 94 96 2 24 BNNT. Whereas, for B N SiH are 0.59 (B) and −0.59e (N), and for respectively. This improvement of the reactivity caused by the Si im- 95 96 24 B N SiH are 1.18 (B) and −1.18e (N). Except the edge regions, the purity on BNNT was also obtained in others theoretical [35,43,44] and 95 96 24 distribution of MEP is homogeneous for pristine BNNT implying that experimental [36] studies. Moreover, the Si doped with a small (6,0) there are not specific sites for nucleophilic (donates an electron pair) BNNT enhance it reactivity in one way that it can be used in catalysis and/or electrophilic (an electron pair acceptor) attack. Therefore, the for CO oxidation [47],N O reduction by SO [48] and CO oxidation by 2 2 lowest reactivity of BNNT (high E value of 6.05 eV, Table 1) is re- N O [49]. According to the Wang [43], the Si-doped BNNT reactivity is g 2 lationship with its electrostatic potential distribution in agreement with relationship with its structural change in the doping process. The lowest its low density of HOMO energy (Fig. 2). Otherwise, in both Si–doped E of B N Si H indicates that it is a conductor nanotube, whereas g 94 96 2 24 BNNT structures, there is the presence of electrophilic sites (blue region the B N SiH is a semiconductor. A Si atom doped in BNNT, does not 95 96 24 in the middle nanotube structure) in the Si atom and, it may explain alter significant the LUMO energy of the BNNT, which they have a si- their high reactivity as confirmed by the low E values (Table 1) and the milar value of ~−0.38 eV. Contrary, the LUMO energy (−3.82 eV) of g high density of HOMO energies in the Si (Fig. 2). Therefore, from the the 2Si–doped BNNT decrease drastically in ten times compared with HOMO, LUMO and MEP analysis, we can infer that the sarin acts as pristine BNNT (−0.38 eV), which become the B N Si H an ex- 94 96 2 24 nucleophile and Si–atoms doped BNNT acts as an electrophile. For in- cellent electron acceptor. The doped atoms increase the HOMO energy stance, the NBO charges of the silicon are 1.28 and 1.73e for 1Si and in 1.15 and 2.37 eV for B N SiH and B N Si H , respectively, 95 96 24 94 96 2 24 2Si–doped BNNT, respectively, and for oxygen atom (sarin) is −1.05e. compared to the pure BNNT. For best visualization of the electronic Consequently, as expected, it was observed strong interactions be- density, in the Fig. 2 is presented the HOMO and LUMO energies, and tween sarin and Si-doped BNNT, confirmed by the covalent bond for- the molecular electrostatic potential (MEP) obtained from NBO atomic mation, as presented in the Fig. 1. The E values (Table 1) follower the charges at DFT//B3LYP/6-31G(d) level. ad

3 J.R. dos Santos, et al. Chemical Physics Letters 738 (2020) 136816

Fig. 2. Representation of HOMO and LUMO energies, and molecular electrostatic potential (MEP, in eV) of the sarin, BNNT and Si–atoms doped BNNT. For MEP, the negative and positive charges range from red to blue colors, respectively. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.)

order: −39.43 < −13.44 < −3.79 kcal/mol for B94N96Si2H24, pristine BNNT. The Ead is very favorable for sarin-B95N96SiH24 inter- B95N96SiH24 and pristine BNNT, respectively. The highest Ead value for action, and the conductivity is increased in ~33% (using equation III) BNNT indicates a weakly physical adsorption between sarin and BNNT implying that this nanotube may be promissory as sensor gas for neu- due the Van der Waals interactions. This finding is in agreement with rotoxic sarin. Likewise, the B94N96Si2H24 appear to be an excellent others studies reporting the nerve agents adsorbed on pristine BNNT sensor to detect the presence of sarin due its lowest Ead and by the [33,44]. Indeed, the energy gap (Table 1) is very similar of BNNT and decreasing of ~350% (from equation III) of the conductivity along the sarin–BNNT complex indicating that the adsorption of sarin does not adsorption process of sarin. Although all nanotubes interact favorably alters significantly the electronic properties of the BNNT. Consequently, with sarin, its presence does not affect significantly the density of this result suggests that the BNNT is not suitable to be used as gas HOMO and LUMO energies of them as can see in the Figs. 2 and 3. sensor for the nerve agent sarin detection using electric conductivity. In details, the sarin adsorbed on BNNT and B95N96SiH24 decrease For instance, the conductivity can be estimated from Eg using the slight their density of LUMO energy; whereas for B94N96Si2H24 the equation below [50], LUMO is located only in the sarin molecule (see Figs. 2 and 3). For HOMO energy, the 1Si and 2Si-doped BNNT use their electronic density −E g centered in the Si atom to bind with oxygen from sarin molecule. An σ ∝ e2kbT (3) interesting find is that in sarin–B94N96Si2H24 complex, there is another where σ is the electric conductivity, Eg is the energy gap, kb is the density of HOMO energy and a partial positive charge in the Si atom Boltzmann’s constant and T is the temperature. Therefore, lowest Eg unbound with sarin, suggesting that a second sarin can bind in this site. value implies a higher electric conductivity. A good sensor chemical must be recovered after it worked. Thus, the As expected, the BNNT doped with 1 and 2Si atoms have the lowest sarin desorption from nanotubes surface can be obtained using physical

Ead due their high reactivity as conﬁrmed by their lowest Eg. Among (e.g. temperature) and/or chemical process, because the HOMO and − them, the lowest Ead value ( 39.43 kcal/mol) of B94N96Si2H24 imply LUMO energies are located only in the sarin molecule in B95N96SiH24 that the sarin molecule interact strongly with B94N96Si2H24 due elec- and B94N96Si2H24. The MEP analyses show that the sarin adsorbed on trostatic interactions, because in this nanotube the large charges of B nanotubes does not alters signiﬁcantly their electrostatic surfaces. The and N atoms of BNNT (1.18 for B and −1.18e for N atom) are pre- main changes observed are due the charge transfer from sarin to na- served. On the other hand, for B95N96SiH24 the B and N charges were notubes (−0.02, −0.21 and −0.24e to pristine BNNT, B95N96SiH24 and decreased and increased, respectively, by two times compared with

4 J.R. dos Santos, et al. Chemical Physics Letters 738 (2020) 136816

Fig. 3. Representation of HOMO and LUMO energies, and molecular electrostatic potential (MEP) for all complexes. For MEP, the negative and positive charges range from red to blue colors, respectively. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.)

B94N96Si2H24, respectively.) Finally, our results show the improvement Acknowledgments of the electronic properties of BNNT by substitution the B by Si atom, which become the Si-doped BNNT as a promissory candidate to be used This research was carried out with the support of the High for sarin gas detection. Performance Computing Center at the Universidade Estadual de Goiás. Also, Jeziel Rodrigues dos Santos acknowledge the Fundação de 4. Conclusions Amparo à Pesquisa do Estado de Goiás (FAPEG) for the award of a scholarship. In this work, DFT method was used to study the adsorption of nerve agent sarin in BNNT and Si–doped BNNT with intention to obtain a new References gas sensor for this lethal neurotoxic agent. The BNNT structure is pre- served along the doping process by Si atom. However, the electronic [1] P.S. Spencer, P.J. Lein, Neurotoxicity, Encycl. Toxicol. Third Ed. 3 (2014) 489–500. properties of the BNNT with Si impurity atoms were improved in- [2] A.M. Costero, S. Royo, R. Martínez-Máñez, S. Gil, M. Parra, F. Sancenón, Chem. Commun. (2007) 4839–4847. creasing their reactivity by altering the HOMO and LUMO energies, and [3] E.H. Braue, Mikrochim. Acta 1970 (1988) 11–16. the electrostatic potential. The adsorption energy values indicate that [4] B.M. Kolakowski, Analyst. 132 (2007) 842–864. the interaction between the nanotubes and sarin is a favorable process, [5] J.G. Weis, T.M. Swager, ACS Macro Lett. 4 (2015) 138–142. − [6] B. Joo, J. Huh, D. Lee, Sensors Actuators B Chem. 121 (2007) 47–53. which 2Si-doped BNNT has the lowest Ead with value of 39.43 kcal/ [7] G. Liu, Y. Lin, Anal. Chem. 77 (2005) 5894–5901. mol. The large Ead and energy gap imply that the pristine BNNT cannot [8] V. Schroeder, S. Savagatrup, M. He, S. Lin, T.M. Swager, Chem. Rev. 119 (2018) be used as gas sensor based in the conductivity. On the other hand, for 599–663. – B N SiH and B N Si H it was observed a variation of ~33 and [9] E. Artacho, D. Sa, A. Rubio, P. Ordejo, Phys. Rev. B. 59 (1999) 678 688. 95 96 24 94 96 2 24 [10] A. Soltani, A.A. Peyghan, Z. Bagheri, Phys. E Low Dimens. Syst. Nanostruct. 48 350% of the conductivity along the adsorption process. Therefore, (2013) 176–180. among than, the B94N96Si2H24 is more suitable to be a new gas sensor [11] M. Salazar Villanueva, E. Chigo Anota, M. del R. Melchor Martínez, L. Tepech – for sarin molecule. Overall, the theoretical results herein can be used to Carrillo, D. García Toral, Superlattices Microstruct. 89 (2015) 319 328. [12] J.C. Ordaz, E.C. Anota, M.S. Villanueva, M. Castro, New J. Chem. 41 (2017) support future experimental and theoretical studies concerning the use 8045–8052. of BNNT doped with atoms to be used as sensor gas to detect nerve [13] X. Blase, A. Rubio, S.G. Louie, M.L. Cohen, Europhys. Lett. 335 (1994) 335–340. – agents. [14] J.S. Lauret, R. Arenal, F. Ducastelle, A. Loiseau, Phys. Rev. Lett. 037405 (2005) 1 4. [15] J.W.G. Wilder, L.C. Venema, A.G. Rinzler, R.E. Smalley, C. Dekker, Nature 391 (1998) 59–62. Declaration of Competing Interest [16] D. Golberg, Y. Bando, C.C.C. Tang, C.Y.Y. Zhi, Adv. Mater. 19 (2007) 2413–2432. [17] M. Endo, J. Phys. Chem. Solids 54 (1994) 1841–1848. ﬁ [18] T. Guo, P. Nikolaev, A.G. Rinzler, D. Tombnek, D.T. Colbert, R.E. Smalley, J. Phys. The authors declare that they have no known competing nancial Chem. 99 (1995) 10694–10697. interests or personal relationships that could have appeared to inﬂu- [19] Y. Shimizu, Y. Moriyoshi, H. Tanaka, S. Komatsu, Y. Shimizu, Y. Moriyoshi, ence the work reported in this paper. H. Tanaka, Appl. Phys. Lett. 75 (1999) 929–931.

5 J.R. dos Santos, et al. Chemical Physics Letters 738 (2020) 136816

[20] C. Zhi, Y. Bando, C. Tang, S. Honda, H. Kuwahara, D. Golberg, J. Mater. Res. 21 [38] MOPAC2016, J.J.P. Stewart, Stewart Computational Chemistry, Colorado Springs, (2006) 2794–2800. CO, USA, 2016. http://OpenMOPAC.net. [21] J.H. Kang, G. Sauti, C. Park, V.I. Yamakov, K.E. Wise, S.E. Lowther, C.C. Fay, [39] K.D. Dobbs, W.J. Hehre, J. Comput. Chem. 8 (1987) 880–893. S.A. Thibeault, R.G. Bryant, ACS Nano 30 (2015) (1950) 11942–11951. [40] M. Noei, H. Asadi, A.A. Salari, S. Mohammad, R. Hosseini, Indian J. Fundam. Appl. [22] D. Abu, J. Niskanen, Y. Xue, D. Golberg, F.M. Winnik, A. Sosnik, Mater. Today Life Sci. 4 (2014) 679–685. Chem. 6 (2017) 45–50. [41] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, X. Cheeseman, [23] K.N. Yun, Y. Sun, J.S. Han, Y.-H. Song, C.J. Lee, A.C.S. Appl, Mater. Interfaces 9 G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, (2017) 1562–1568. K. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, [24] K.N. Yun, S.H. Lee, J.S. Han, Y. Song, Nanotechnology 29 (2018) 085203. M. Ehara, T.V. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, [25] Q. Huang, Y. Bando, L. Zhao, C.Y. Zhi, D. Golberg, Nanotechnology 20 (2009) O. Kitao, H. Nakai, K.N.K.J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro, 415501. M. Bearpark, J.J. Heyd, E. Brothers, S.S.V.N. Staroverov, R. Kobayashi, J. Normand, [26] Y. Yu, H. Chen, Y. Liu, L. Hua, Y. Chen, Electrochem. Commun. 30 (2013) 29–33. K. Raghavachari, A. Rendell, J.C. Burant, V.B. Iyengar, J. Tomasi, M. Cossi, N. Rega, [27] J. Kaur, P. Singla, N. Goel, Appl. Surf. Sci. 328 (2014) 632–640. J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, C.C. Adamo, J. Jaramillo, R. Gomperts, [28] M. Yoosefian, N. Etminan, M.Z. Moghani, S. Mirzaei, S. Abbasi, Superlattices R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, J. Pomelli, J.W. Ochterski, Microstruct. 98 (2016) 325–331. R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, [29] A. Habibi-Yangjeh, H. Basharnavaz, Mol. Phys. 116 (2018) 1320–1327. J.C.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman, J.V. Ortiz, [30] H. Basharnavaz, A. Habibi-yangjeh, Mol. Phys. 116 (2017) 204–211. D.J. Fox, Revision C.01, Gaussian, Inc., Gaussian 09, 2009. [31] T. Movlarooy, M.A. Fadradi, Chem. Phys. Lett. 700 (2018) 7–14. [42] A.E. Reed, R.B. Weinstock, F. Weinhold, J. Chem. Phys. 83 (1985) 735–746. [32] P. Sripadung, N. Nunthaboot, B. Wanno, Monatshefte Für Chemie – Chem. Mon. [43] R. Wang, D. Zhang, Y. Liu, C. Liu, Nanotechnology 20 (2009) 505704. 150 (2019) 1011–1018. [44] A. Bahari, M. Bagheri, M. Amiri, Solid State Commun. 267 (2017) 1–5. [33] M. Bezi, A. Soltani, A.S. Ghasemi, E.T. Lemeski, N. Gholami, H. Balakheyli, Appl. [45] T. Koopmans, Physic. 1 (1934) 104–113. Surf. Sci. 411 (2017) 1–10. [46] O.V. de Oliveira, T. Cuya, E.C. Ferreira, S. Gonçalves, Chem. Phys. Lett. 706 (2018) [34] R. Amirkhani, M.H. Omidi, R. Abdollahi, H. Soleymanabadi, J. Clust. Sci. 29 (2018) 82–86. 757–765. [47] M.D. Esrafili, N. Saeidi, Struct. Chem. 27 (2016) 595–604. [35] S. Guerini, T. Kar, P. Piquini, Eur. Phys. J. B 38 (2004) 515–518. [48] M.D. Esrafili, N. Saeidi, RSC Adv. 403 (2017) 43–50. [36] Y.J. Cho, C.H. Kim, H.S. Kim, J. Park, H.C. Choi, H. Shin, G. Gao, H.S. Kang, Chem. [49] M.D. Esrafili, N. Saeidi, P. Nematollahi 5 (2015) 100290–100298. Mater. 21 (2009) 136–143. [50] S.S. Li, Semiconductor Physical Electronics, Springer Science & Business Media, [37] J.J.P. Stewart, J. Mol. Model. (2013) 1–32. 2012.