1: %\documentclass[aps,prb,preprint]{revtex4}

2: \documentclass[aps, prb, twocolumn, showpacs, floatfix]{revtex4}

3: \usepackage{graphicx}

4: 

5: \begin{document}

6: 

7: \title{Electronic structures of organic molecule encapsulated BN nanotubes under transverse electric field}

8: 

9: \author{Wei He}

10: \author{Zhenyu Li}

11: \author{Jinlong Yang}

12: \thanks{Corresponding author. E-mail: jlyang@ustc.edu.cn}

13: \author{J. G. Hou}

14: 

15: \affiliation{Hefei National Laboratory for Physical Sciences at

16:      Microscale,  University of Science and Technology of

17:      China, Hefei,  Anhui 230026, China}

18: 

19: \date{\today}

20: 

21: \begin{abstract}

22: The electronic structures of boron nitride nanotubes (BNNTs) doped

23: by different organic molecules under a transverse electric field

24: were investigated via first-principles calculations. The external

25: field reduces the energy gap of BNNT, thus makes the molecular bands

26: closer to the BNNT band edges and enhances the charge transfers

27: between BNNT and molecules. The effects of the electric field

28: direction on the band structure are negligible. The electric field

29: shielding effect of BNNT to the inside organic molecules is

30: discussed. Organic molecule doping strongly modifies the optical

31: property of BNNT, and the absorption edge is red-shifted under

32: static transverse electric field.

33: \end{abstract}

34: 

35: \pacs{75.75.+a, 73.22.-f, 72.80.Rj}

36: 

37: \maketitle

38: 

39: \section{introduction}

40: 

41: Nanotube is a very important actor in nanodevice applications due to

42: its novel properties. Compared to the more widely studied carbon

43: nanotube (CNT), boron nitride nanotube (BNNT) has its own

44: distinctive properties. Besides the excellent mechanical stiffness

45: and thermal conductivity comparable to CNT, \cite{review} BNNT has a

46: unified electronic structure regardless of its diameter, chirality,

47: and the number of walls of the tube. \cite{bn1, bn2, bn3} BNNT also

48: shows pronounced resistance to oxidation, and it is stable up to 700

49: $^{\circ}$C in air, while CNT survives only below 400 $^{\circ}$C.

50: \cite{bn5} All these properties make BNNT an attractive candidate

51: for nano-electronics.

52: 

53: Pristine BNNT is a wide-gap semiconductor. For electronics

54: application, it is desirable to make $p$- or $n$-type doping to

55: BNNT. Our previous study shows that the electronic structure of BNNT

56: can be modified by organic molecule encapsulation. \cite{bnhe}

57: Electrophilic molecule introduces acceptor states in the wide gap of

58: BNNT close to the valence band edge (VBE), which makes the doped

59: system a $p$-type semiconductor. However, with typical nucleophilic

60: organic molecules, instead of shallow donor states, only deep

61: occupied molecular states are observed. There is a significant

62: electron transfer from BNNT to an electrophilic molecule, while the

63: charge transfer between a nucleophilic molecule and BNNT is

64: negligible.

65: 

66: On the other hand, previous theoretical studies \cite{field0,

67: field1, field2} showed that the band gap of CNT or BNNT can be

68: reduced and even closed by applying a transverse electric field, due

69: to the giant Stark effect (GSE). \cite{field1} The electric field

70: can mix the nearby subband states in the valence band (VB) complex

71: and conduction band (CB) complex separately, leading to an electric

72: field dependence for the band gap. This effect is more remarkable in

73: BNNTs than in CNTs, due to the reduced screening of the electric

74: field in BNNTs. The GSE was lately confirmed experimentally using

75: the bias dependent scanning tunneling microscopy and scanning

76: tunneling spectroscopy. \cite{field3}

77: 

78: Since the electronic structure of BNNT can be tuned by both organic

79: molecule encapsulation and transverse electric field, it is thus

80: interesting to see what will happen if we apply both. In this paper,

81: by performing density functional theory (DFT) calculations, we study

82: the electronic structures of organic molecule encapsulated BNNTs

83: under a transverse electric field. The electrostatic shielding of

84: BNNT to the inside molecules and the effects of organic molecule

85: doping and electric field on the optical properties of BNNT are also

86: studied.

87: 

88: Following our previous study, \cite{bnhe} several typical

89: electrophilic and nucleophilic molecules are considered for BNNT

90: encapsulation. The two electrophilic molecules studied are

91: tetracyano-p-quinodimethane (TCNQ) and

92: tetrafluorocyano-p-quinodimethane (F4TQ). Three nucleophilic

93: molecules are selected: tetrakis(dimethylamino)ethylene (TDAE),

94: anthracene (ANTR), and tetrathia-fulvalene (TTF). The (16,0) BNNT is

95: chosen as a prototype in this study. We name the organic molecules

96: capsulated BNNT as M@BN, where M is the name of the doped organic

97: molecules.

98: 

99: \begin{figure}[!hbp]

100: \includegraphics[width=8cm]{total.eps}

101: \caption{(Color online) Optimized geometrical structures of the M@BN

102: systems under a 0.5 V/\AA\ electric field. B in dark-salmon, N in

103: blue, C in grey, H in yellow, F in aqua, and S in goldenrod. }

104: \label{fig:geo}

105: \end{figure}

106: 

107: The rest of the paper is organized as follows: In Sec. II, the

108: theoretical approach and computational details are briefly

109: described. In Sec. III, the calculated band structures, charge

110: transfers between BNNT and molecules, shielding effect under

111: electric field, and optical properties are discussed. Finally, we

112: conclude in Sec. IV.

113: 

114: \section{Computational Details}

115: 

116: To investigate the geometry structure and electronic states of the

117: organic molecule encapsulated BNNT, we performed first-principles

118: spin-polarized DFT calculations. The computational details of

119: geometry optimization and band structure calculation can be found in

120: our previous study, \cite{bnhe} and we only give a brief summary

121: here. We used the projector augmented-wave (PAW) method \cite{20,21}

122: implemented in the Vienna Ab Initio Simulation Package (VASP).

123: \cite{17,18} The Perdew, Burke and Ernzerhof (PBE)

124: exchange-correlation functional was chosen. \cite{19} We note that

125: more rigorous band structure calculations are required to yield

126: correct band gaps. \cite{Baumeier07} However, PBE functional has

127: been widely used in similar systems, and it should give correct

128: trends which are interested in this study. The energy cutoff for

129: plane-waves was 400 eV. In our calculation, each BNNT was separated

130: by 10 \AA\ of vacuum, and the minimum distance between periodically

131: repeated organic molecules along the BNNT tube axis was

132: $\sim$8.7\AA. A 4$\times$1$\times$1 $\Gamma$-centered Monhkorst-Pack

133: $k$-point grid was used for Brillouin zone sampling.

134: 

135: 

136: The way to handle electric field in VASP is adding an artificial

137: dipole sheet in the middle of the vacuum part in the periodic

138: supercell. \cite{vaspdipole1, vaspdipole2} In our study, the

139: direction of the electric field was chosen to be perpendicular to

140: the nanotube's axis (along the $y$ direction), with a magnitude of

141: 0.20 or 0.50 V/\AA. As discussed later, the selected fields are

142: strong enough to make significant change of the electronic

143: structure, and weak enough to avoid artificial field emission.

144: 

145: 

146: \begin{figure}[!hbp]

147: \includegraphics[width=8cm]{E0.5vacuum.eps}

148: \caption{Electrostatic potential (in volt) across the BNNT in the

149: M@BN systems under 0.5 V/\AA\ external electric field. The top of

150: the highest occupied bands are marked by dash lines. The numbers on

151: the top of the curves are the effective electric field (in V/\AA),

152: calculated by the slope of the line connecting the left

153: high-potential point and the right low-potential point in vacuum.}

154: \label{fig:potential}

155: \end{figure}

156: 

157: All linear optical properties, such as absorption coefficient and

158: refractive index, can be obtained as functions of dielectric

159: function. The real part of the dielectric function can be obtained

160: from its imaginary part by Kramer-Kronig transformation, and the

161: imaginary part $\epsilon_2$ can be calculated based on the

162: independent-particle approximation, \cite{Adolph0108} i.e.,

163: \begin{equation}

164:  \epsilon_2(\omega)=\frac{4\pi}{\Omega\omega^2}\sum_{i,j}\sum_\mathbf{k}

165:  W_\mathbf{k}|P_{ij}|^2\delta(E_{\mathbf{k}j}-E_{\mathbf{k}i}-\omega)

166: \end{equation}

167: where $\Omega$ is the unit cell volume, and $\omega$ is the photon

168: energy. Summations for $i$ and $j$ are performed for valence and

169: conduction bands, respectively. $P_{ij}$ denotes the dipole

170: transition matrix elements obtained from the self-consistent band

171: structure calculations and $W_\mathbf{k}$ is the $\mathbf{k}$ point

172: weighting. The same method has been applied to h-BN, and the results

173: are in reasonably good agreement with experiments. \cite{Guo0502}

174: The number of bands used for dielectric function calculations are

175: the sum of the electron number and the ion number. The number of

176: occupied bands is half of the number of electron, and the number of

177: unoccupied bands are more than 380 for all the systems studied in

178: this paper. The optical polarization in dielectric function

179: calculations is perpendicular to the nanotube axis. In this way, for

180: some parts of the tube wall, the optical polarization is nearly

181: perpendicular to the BN layer, while for the other parts of the tube

182: wall, it is roughly parallel to the BN layer.

183: 

184: \section{RESULTS AND DISCUSSION}

185: 

186: \subsection{Geometric and band structures}

187: 

188: As reported in our previous study, \cite{bnhe} after the organic

189: molecules are encapsulated into BNNT, their geometries changes

190: little. This suggests the weak interaction between BNNT and the

191: encapsulated molecules, due to the large distances between them. The

192: separations between neighboring molecules and between molecule and

193: BNNT are larger than additions of the corresponding atomic van der

194: Waals radius. Optimizations under the transverse electric field gave

195: similar structure as in the zero-field case. The optimized

196: geometrical structures under a 0.5 V/\AA\ electric field are shown

197: in Fig. \ref{fig:geo}. The relaxed geometries of the M@BN systems

198: before and after applying electric field change little. The largest

199: change of bond length, comparing to the geometry under zero field,

200: is smaller than 0.01 \AA. Consistently, the direction of the

201: transverse field does not affect the geometric and electronic

202: structures. Our test calculation gave almost identical band

203: structures for F4TQ@BN under 0.5 V/\AA\ transverse electric field

204: along two different directions. For simplicity, in the rest part of

205: the paper, all electronic structure calculations were done based on

206: the geometrical structures obtained without electric field.

207: 

208: 

209: 

210: Before calculating the electronic structure of M@BN with transverse

211: electric field applied, two important issues are needed to be

212: discussed. One is that the vacuum region should be broad enough to

213: avoid the overlap between the charge density and the artificial

214: dipole sheet used to generate electric field. The width of the

215: vacuum layer between two neighboring BNNTs in our calculation is 10

216: \AA, and the work function of (16,0) BNNT is 6.36 eV. The charge

217: density of BNNT roughly decay as $\exp(-\sqrt{\phi}r)$, where $\phi$

218: is the work function in eV and $r$ is the distance from the BNNT in

219: \AA. \cite{vaspdipole3} The charge density thus decays one magnitude

220: every 0.91 \AA. The dipole sheet is 5 \AA\ far from the BNNT, and

221: there is thus only negligible charge density there. The other issue

222: is that the vacuum region should not be too broad so as to attract

223: electrons to vacuum (field emission). In Fig. \ref{fig:potential},

224: we plot the electrostatic potential of the M@BN system under a 0.5

225: V/\AA\ electric field, the strongest electric field used in our

226: calculations. We can see that the lowest potential in vacuum region

227: is still higher than the edge of the highest occupied band (HOBand),

228: which confirms that the artificial field emission has not been

229: introduced in this study.

230: 

231: \begin{figure}[!hbp]

232: \includegraphics[width=8cm]{BNNTBand.eps}

233: \includegraphics[width=8cm]{charge.eps}

234: \caption{Band structures (from $\Gamma$ to X) of the pristine BNNT

235: under (a) 0.0, (b) 0.2, and (c) 0.5 V/\AA\ electric fields. Fermi

236: levels are marked by dashed lines. Density distribution for HOBand,

237: LUBand, and LUBand+1 in pristine BNNT under (d-f) 0.0 and (g-i) 0.5

238: V/\AA\ electric fields. The electric field direction is from up to

239: down.} \label{fig:bandbn}

240: \end{figure}

241: 

242: We first calculate the band structure of the pristine BNNT under the

243: 0.2 and 0.5 V/\AA\ electric fields, and the results are shown in

244: Fig. \ref{fig:bandbn}. The band gap decreases from 4.45 to 3.73 and

245: 1.98 eV, when the field increases from 0 to 0.2 and 0.5 V/\AA,

246: respectively. Our result is consistent with the previous theoretical

247: study on similar systems. \cite{field1} Such band gap decrease is a

248: result of the extensive mixing among the subband states within the

249: valence and conduction band complex under electric filed. The band

250: mixing is also reflected by the band-resolved charge density

251: distributions, as shown in Fig. \ref{fig:bandbn}, where the density

252: of the HOBand, the lowest unoccupied band (LUBand), and LUBand+1

253: under 0.0 and 0.5 V/\AA\ field are plotted. The orbital density move

254: along the electric field for occupied states, and against the field

255: direction for unoccupied states. From the density distributions, the

256: LUBand state can be identified as a nearly free electron (NFE)

257: state. \cite{field1}

258: 

259: \begin{figure}[!hbp]

260: \includegraphics[width=8cm]{E0.5bandcompare.eps}

261: \caption{Band structures (from $\Gamma$ to X) of the pristine BNNT

262: and the M@BN systems under 0.5 V/\AA\ electric field. Fermi levels

263: are marked by dashed lines. } \label{fig:band0.5}

264: \end{figure}

265: 

266: The band structures of the M@BN systems under 0.2 and 0.5 V/\AA\

267: electric field are then calculated. Similar to results under zero

268: electric field, \cite{bnhe} the bands are either mainly from

269: molecules or mainly from BNNT. The molecular bands are much flatter

270: than the BNNT bands. Since there is no qualitative difference

271: between the band structure of the M@BN systems under 0.2 and 0.5

272: V/\AA\ electric fields, we will focus on the 0.5 V/\AA\ case (Fig.

273: \ref{fig:band0.5}). For the two electrophilic molecules (TCNQ and

274: F4TQ), the lowest unoccupied molecular bands become very close to

275: the BNNT VBE, which means good $p$-type doping. It also suggests a

276: considerable charge transfer between the molecules and the BNNT. We

277: define the gap between the lowest unoccupied molecular band and the

278: BNNT VBE as $E_p$, and the gap between the highest occupied

279: molecular band and the BNNT conduction band edge (CBE) as $E_n$.

280: Their values are listed in Table \ref{tbl:eg}. For the nucleophilic

281: organic molecules TDAE and TTF, $E_n$ drops from 2.30 and 2.85 eV at

282: zero field \cite{bnhe} to 1.08 and 1.54 eV, respectively. However,

283: it is still too large to form good $n$-type doping. For ANTR@BN, the

284: highest occupied molecular band is pushed into the BNNT valence band

285: manifest, and there is only an unoccupied molecular band close to

286: the BNNT CBE. Therefore, neither $n$-type or $p$-type doping is

287: introduced to BNNT by ANTR encapsulation.

288: 

289: \begin{table}[b]

290: \caption{Energy gaps $E_p$, $E_n$, and $E_g^{BN}$ of the pristine

291: BNNT and the M@BN systems under electric field $\mathcal{E}$. The

292: energy gaps are in eV, and the electric field $\mathcal{E}$ is in

293: V/\AA. } \label{tbl:eg}

294: \begin{tabular}{cccccccc}

295:  \hline\hline

296:  Gap & $\mathcal{E}$ & BN & TCNQ & F4TQ & TDAE & TTF & ANTR   \\

297:  \hline

298: $E_p$      & 0.0 &  --- & 0.15 & 0.09 & 4.37 & 3.55 & 2.96 \\

299:            & 0.2 &  --- & 0.17 & 0.11 &  --- & 3.25 & 2.65 \\

300:            & 0.5 &  --- & 0.09 & 0.05 &  --- &  --- & 1.93 \\

301: $E_n$      & 0.0 &  --- &  --- &  --- & 2.30 & 2.85 & 3.71 \\

302:            & 0.2 &  --- &  --- &  --- & 1.86 & 2.49 & 3.40 \\

303:            & 0.5 &  --- &  --- &  --- & 1.08 & 1.54 &  --- \\

304: $E_g^{BN}$ & 0.0 & 4.45 & 4.45 & 4.48 & 4.45 & 4.46 & 4.43 \\

305:            & 0.2 & 3.73 & 4.00 & 4.18 & 3.63 & 3.79 & 3.81 \\

306:            & 0.5 & 1.98 & 2.79 & 2.91 & 2.31 & 2.14 & 2.18 \\

307:  \hline\hline

308: \end{tabular}

309: \end{table}

310: 

311: To consider the molecular effect on the intrinsic electronic

312: structure of BNNT under electric field, we define the energy gap

313: between the CBE and VBE of BNNT as $E_g^{BN}$. As shown in Table

314: \ref{tbl:eg}, without external field, it is not very sensitive to

315: the type of molecules. However, $E_g^{BN}$ for TCNQ@BN and F4TQ@BN

316: under electric field is much larger than other systems under the

317: same field. This can be easily understood by their electronic

318: structures. Under strong external field, the energy gaps of TCNQ@BN

319: and F4TQ@BN become very small, which means enhanced metallicity and

320: stronger electrostatic screening. With stronger screening, the

321: reduction of electrostatic potential along the direction of electric

322: field is slower (see Fig. \ref{fig:potential}), therefore,

323: $E_g^{BN}$ reduction compared to pristine BNNT and other M@BN

324: systems is smaller. Similar behavior has been observed in our study

325: on defective BNNTs under an electric field. \cite{bnhu}

326: 

327: 

328: 

329: \subsection{Charge transfer}

330: 

331: An important property of the molecule encapsulated system is the

332: charge transfer between the molecules and BNNT. We calculate the

333: charge transfer using the method described in our previous work,

334: \cite{bnhe} originally proposed by Lu et al..\cite{charge1} Briefly

335: speaking, the boundary between the organic molecule and the BNNT is

336: determined by the maximum/minimum position of the

337: cylinder-integrated differential electron density curve, and the

338: corresponding maximum/minimum value is the charge transfer value.

339: The differential electron density is defined as the difference

340: between electron densities of the M@BN system under electric field

341: and the sum of charge densities of BNNT and molecules without

342: electric field. The range of the radial position $R_b$ of boundary

343: between molecule and BNNT can be estimated by the density

344: distribution of the BNNT bands. For electrophilic molecules, $R_b$

345: should be inside the region where the BNNT HOBand has a high density

346: ($R_b<6$ \AA), while for nucleophilic molecules, $R_b$ should be

347: inside region where the BNNT LUBand has a high density ($R_b<3.5$

348: \AA). \cite{bnhe}

349: 

350: \begin{figure}[!hbp]

351: \includegraphics[width=8cm]{E0.5charge.eps}

352: \caption{(Color online) The cylinder-integrated differential

353: electron density curves for the the pristine BNNT and the M@BN

354: systems under a 0.5 V/A electric field. The blue vertical lines

355: indicate the radial boundary between molecules and the BNNT, in the

356: viewpoint of charge transfer. The black vertical line in the first

357: panel marks the radius of the BNNT.} \label{fig:E0.5charge}

358: \end{figure}

359: 

360: For pristine BNNT, the differential electron density is just the

361: density change before and after applying electric field. As shown in

362: Fig. \ref{fig:E0.5charge}, the electron density increases inside the

363: nanotube wall, while it decrease outside the wall. Of course, the

364: absolute value of this change is very small ($\sim$0.01 electron).

365: For the electrophilic organic molecules TCNQ and F4TQ, under 0.5

366: V/\AA\ electric field, there are 0.28 and 0.38 electrons per

367: molecule transfer from BNNT, about twice as the value without field

368: (0.10 and 0.24 electron). \cite{bnhe} This is due to the closer

369: molecular state in energy to VBE. The boundary radius is 4.3 \AA,

370: which is basically the midpoint between the molecule and the BNNT as

371: in the no field case. For the nucleophilic organic molecules TDAE,

372: TTF and ANTR, the charge transfers form molecules to BN also

373: increase, which is 0.006, 0.011, and 0.008 electron per molecule,

374: respectively (0.004, 0.008, and 0.005 electrons without external

375: field). Such small electron transfers also indicate that the doped

376: system is not a good $n$-type semiconductor.

377: %, which is mainly contributed by the NFE states ({\it empty states})

378: 

379: 

380: \subsection{Shielding effect}

381: 

382: It is interesting to study the effects of BNNT to the geometry and

383: electronic structure of the inside molecules under an electric

384: field. In Fig. \ref{fig:dos}, we compare the molecular partial

385: densities of states (PDOS) with and without external field for two

386: representative molecules: an electrophilic one (F4TQ) and a

387: nucleophilic one (TTF). The calculated PDOS shows different

388: behaviors for these two types of molecules. For the electrophilic

389: organic molecules F4TQ (Fig. \ref{fig:dos}a), there is an nearly

390: rigid upshift of the molecular PDOS after applying the electric

391: field. This is due to the electron transfer from BNNT to the

392: molecules. For the nucleophilic organic molecules TTF (Fig.

393: \ref{fig:dos}c), the molecular PDOS almost does not change after

394: applying the electric field, which is consistent with the little

395: charge transfers between BNNT and the molecules.

396: 

397: \begin{figure}[!hbp]

398: \includegraphics[width=8cm]{dos.eps}

399: \caption{(Color online) Molecular PDOS of (a) F4TQ@BN and (c) TTF@BN

400: under 0.0 (solid) and 0.5 (dashed) V/\AA\ electric field. DOS of

401: individual (b) F4TQ and (d) TTF molecules under 0.0 (solid) and 0.5

402: (dashed) V/\AA\ electric field.} \label{fig:dos}

403: \end{figure}

404: 

405: We also check the shielding effect by applying electric field to

406: individual molecules without a BNNT encapsulation. First, we relax

407: the geometry of the individual molecules under the 0.5 V/\AA\

408: electric field. Contrasting to the encapsulated case, we found a

409: significant geometry relaxation after applying the electric field.

410: Then, we fix the molecular geometry to the one as inside BNNT, and

411: calculate the electronic structure under zero and finite electric

412: field. The effect of electric field to the molecular electronic

413: structures for individual molecules is different from that for

414: molecules encapsulated by BNNT (see Fig. \ref{fig:dos}). For F4TQ

415: molecule, the DOS energy shift caused by electric field is smaller

416: than that for PDOS of F4TQ@BN. However, we can see a notable change

417: of the overall DOS shape after applying the electric field. For TTF,

418: electric field shifts the DOS of individual molecules, but it does

419: not change the molecular PDOS. In summary, our results indicates,

420: although not good as CNT, the BNNT still provides a significant

421: electrostatic shielding for the inside molecules: the geometry of

422: the inside molecules and the electronic structures of nucleophilic

423: molecules are almost not affected by an external field under the

424: protection of BNNT.

425: 

426: \subsection{Optical properties}

427: 

428: The calculated imaginary part of the dielectric function of the

429: pristine BNNT without external electric field is shown in the inset

430: of Fig. \ref{fig:optics}a. A much denser $k$-point sampling is

431: required to obtain smooth curves. \cite{Guo0502} However, the main

432: features of $\epsilon_2$ are already presented in our results. And

433: our calculated $\epsilon_2$ of prinstine BNNT agrees well with the

434: previous theoretical result on the zigzag BNNTs. \cite{Guo0502}

435: Mainly, there are three peaks at about 5.6, 10.6, and 13.5 eV,

436: respectively. The strong peak around 5.6 eV corresponds to the

437: $\pi\rightarrow\pi^*$ interband transitions. The two other peaks

438: higher in energy and weaker in intensity are associated with

439: interband transitions also involving the $\sigma$ bands. We note

440: that the peak around 10.6 eV does not appear in our test calculation

441: with the optical polarization along the tube axis. By optical

442: selection rule analysis, \cite{Guo0502} the 10.6 eV peak should

443: correspond to $\pi\rightarrow\sigma^*$ and $\sigma\rightarrow\pi^*$

444: transitions, and the 13.5 peak mainly contributed by

445: $\sigma\rightarrow\sigma^*$ transitions. In Fig. \ref{fig:optics}a,

446: we also plot $\epsilon_2$ of the pristine BNNT under transverse

447: electric field. Similar to a previous study on pristine BNNT,

448: \cite{field5} the whole imaginary dielectric function is not

449: strongly affected by the static field, but the absorption edge is

450: red-shifted.

451: 

452: \begin{figure}[!hbp]

453: \includegraphics[width=8cm]{opticsper.eps}

454: \caption{(Color online) The imaginary part ($\epsilon_2$) of the

455: theoretical dielectric function of the BNNT and the M@BN systems,

456: with (red) and without (blue) the external static electric field.

457: Inset: $\epsilon_2$ plotted in a larger energy scale. }

458: \label{fig:optics}

459: \end{figure}

460: 

461: 

462: With molecule intercalation, all the three peaks are still clearly

463: shown in the M@BN system. However, new features appear in the

464: low-energy region from 0 to 5 eV. These peaks below the BNNT

465: absorption edge ($\sim$4.5 eV) are clearly from transitions between

466: molecular bands and the BNNT bands or between molecular bands

467: themselves. For the two electrophilic molecules, there are peaks

468: close to the zero energy. This is consistent with the very narrow

469: band gap calculated for these two systems. There is no significant

470: peaks below 4.5 eV for TDAE@BN, and the new peaks for TTF@BN and

471: ANTR@BN are close to the BNNT absorption edge. Our results indicate

472: that the optical property of BNNT can be strongly modified by

473: different organic molecule encapsulation.

474: 

475: Next, we study the case where both molecule intercalation and

476: electric field exist. The theoretical imaginary part of the

477: dielectric functions of the molecule doped BNNT under a 0.5 V/\AA\

478: transverse static electric field are also presented in Fig.

479: \ref{fig:optics}. The three peaks from BNNT are still not strongly

480: affected by the static field. However, if we look into the details

481: below 5 eV, we can see three different behaviors. For transitions

482: between valence band and conductance bands of BNNT, the electric

483: field GSE broadens the band manifests, thus broadens the absorption

484: peaks too. This broadening effect red-shifts the absorption edge of

485: BNNT as shown in Fig. \ref{fig:optics}d. For transitions between

486: molecular bands, due to electrostatic shielding of BNNT, the

487: electric field does not strongly affect the peaks, as shown in Fig.

488: \ref{fig:optics}e and \ref{fig:optics}f. For transitions between

489: BNNT bands and molecular bands, the electric field affects both the

490: amplitude of the peaks and the peak positions, as shown in Fig.

491: \ref{fig:optics}b and \ref{fig:optics}c.

492: 

493: 

494: \section{conclusion}

495: 

496: In summary, we calculate the electronic structure and optical

497: properties of BNNT under transverse electric field with different

498: molecule intercalations. When applying transverse electric field,

499: the band gap of BNNT decrease, while the band structures of organic

500: molecules change little, or only with a rigid shift. After applying

501: electric field, there is more charge transfer between molecules and

502: nanotube in the $p$-type semiconductors TCNQ@BN and F4TQ@BN. BNNT

503: provides relatively good electrostatic shielding for the geometry

504: relaxation of the inside organic molecular chain, and for the

505: electronic structure of nucleophilic molecules. Organic molecule

506: doping strongly modifies the optical properties of the composite

507: M@BN systems. The absorption edge is slightly red-shifted under

508: static electric field.

509: 

510: 

511: \begin{acknowledgements}

512: The postprocessing routine downloaded from the homepage of Prof.

513: Furthmuller has been used in the optical property calculations. This

514: work is partially supported by the National Natural Science

515: Foundation of China (50721091, 20533030, 50731160010), by National

516: Key Basic Research Program under Grant No.2006CB922004, by the

517: USTC-HP HPC project, and by the SCCAS and Shanghai Supercomputer

518: Center.

519: \end{acknowledgements}

520: 

521: \begin{thebibliography}{99}

522: 

523: \bibitem{review} D. Golberg, Y. Bando, C. C. Tang, and C. Y. Zhi, Adv. Mater. {\bf 19}, 2413(2007).

524: 

525: \bibitem{bn1} A. Rubio, J. L. Corkill, and M. L. Cohen, Phys. Rev. B  {\bf 49}, 5081 (1994).

526: 

527: \bibitem{bn2} N. G. Chopra, R. J. Luyken, K. Cherrey, V. H. Crespi, M. L. Cohen, S. G. Louie, and A. Zettl, Science {\bf 269}, 966 (1995).

528: 

529: \bibitem{bn3} X. Blase, A. Rubio, S. G. Louie, and M. L. Cohen, Europhy. Lett. {\bf 28}, 335 (1994).

530: 

531: %\bibitem{bn4} M.Terauchi, M.Tanaka, K.Suzuki, A.Ogino, and K.Kimura, Chem. Phys. Lett. {\bf 324}, 359 (2000).

532: 

533: \bibitem{bn5} Y. Chen, J. Zou, S. J. Campbell, and G. L. Caer, Appl. Phys. Lett. {\bf 84}, 2430 (2004).

534: 

535: \bibitem{bnhe} W. He, Z. Li, J. Yang, and J. G. Hou, J. Chem. Phys., 128, 164701 (2008).

536: 

537: \bibitem{field0} J. O'Keeffe, C. Wei, and K. Cho, Appl. Phys. Lett. {\bf 80}, 676 (2002).

538: 

539: \bibitem{field1} K. H. Khoo, M. S. C. Mazzoni, and S. G. Louie, Phys, Rev. B {\bf 69}, 201401 (2004)

540: 

541: \bibitem{field2} C. W. Chen, M. H. Lee, adn S. J. Clark, Nanotechnology {\bf 15}, 1837 (2004)

542: 

543: \bibitem{field3} M. Ishigami, J. D. Sau, S. Aloni, M. L. Cohen, and A. Zettl, Phys, Rev. Lett. {\bf 94}, 056804 (2005)

544: 

545: \bibitem{20} G. Kresse, D. Joubert, Phys. Rev. B {\bf 59}, 1758 (1999).

546: 

547: \bibitem{21} P. E. Blochl, Phys. Rev. {\bf 50}, 17953 (1994).

548: 

549: \bibitem{17} G. Kresse, J. Furthm¨¹ller, Phys. Rev. B {\bf 64}, 11169 (1996).

550: 

551: \bibitem{18} G. Kresse, J. Furthmuller, Comput. Mater. Sci. {\bf 6}, 15 (1996).

552: 

553: \bibitem{19} J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. {\bf 77}, 3865 (1996).

554: 

555: \bibitem{Baumeier07} B. Baumeier, P. Kruger, and J. Pollmann, Phys.

556: Rev. B {\bf 76}, 085407 (2007).

557: 

558: \bibitem{vaspdipole1} J. Neugebauer and M. Scheffler, Phys. Rev. B 46, 16067 (1992).

559: 

560: \bibitem{vaspdipole2} G. Kresse and J. Furthmuller, VASP Guide.

561: 

562: \bibitem{Adolph0108} B. Adolph, J. Furthmuller, and F. Bechsted,

563: Phys. Rev. B {\bf 63}, 125108 (2001)

564: 

565: \bibitem{Guo0502} G. Y. Guo and J. C. Lin, Phys. Rev. B {\bf 71},

566: 165402 (2005)

567: 

568: \bibitem{vaspdipole3}P. J. Feibelman, Phys. Rev. B 64, 125403(2001).

569: 

570: \bibitem{bnhu} S. Hu, Z. Li, X. C. Zeng, and J. Yang, J. Phys. Chem.

571: C, ASAP article (2008)

572: 

573: \bibitem{charge1} X. Lu, M. Grobis, K. H. Khoo, S. G. Louie, and M. F. Crommie, Phys. Rev. B 70, 115418 (2004).

574: 

575: %\bibitem{shield1} E. N. Brothers, G. E. Scuseria, and K. N. Kudin, J Chem Phys {\bf 124}, 041101

576: %(2006)

577: 

578: %\bibitem{shield3} L. Wang, J. Lu, L. Lai, W. Song, M. Ni, Z. Gao, and W. N. Mei, J Phys Chem C {\bf 111}, 3285

579: %(2007)

580: 

581: %\bibitem{shield2} B. Kozinsky and N. Marzari, Phys Rev Lett {\bf 96}, 166801

582: %(2006)

583: 

584: \bibitem{field5} C. W. Chen, M. H. Lee, and Y. T. Lin, Appl. Phys. Lett. {\bf 89}, 223105 (2006).

585: 

586: %\bibitem{organic0} T. Takenobu, T. Takano, M. Shiraishi, Y. Murakami, M. Ata, H. Kataura, Y. Achiba, and Y. Iwasa, Nat. Mater. {\bf 2}, 683 (2003).

587: 

588: %\bibitem{organic1} J. Lu, S. Nagase, D. Yu, H. Ye, R. Han, Z. Gao, S. Zhang, and L. Peng, Phys. Rev. Lett.{\bf 93}, 116804 (2004).

589: 

590: %\bibitem{organic2} W. Liang, J. Yang, and J. Sun, Appl. Phys. Lett. {\bf 86}, 223113 (2005).

591: 

592: %\bibitem{Yamashiro0310} A. Yamashiro, Y. Shimoi, K. Harigaya, and K.

593: %Wakabayashi, Phys. Rev. B \textbf{68}, 193410 (2003)

594: 

595: \end{thebibliography}

596: 

597: \end{document}

598: