1: \documentclass[aps,prl,twocolumn,superscriptaddress,showpacs,floatfix]{revtex4}
2: \usepackage{graphicx}
3: \bibliographystyle{apsrev.bst}
4: \begin{document}
5: \title{Symmetry restrictions in chirality dependence of physical properties
6: of single wall nanotubes}
7: \author{Fei Ye}
8: \affiliation{Center for Advanced Study, Tsinghua University,
9: Beijing 100084, China}
10: \author{Bing-Shen Wang}
11: \affiliation{National Laboratory of Semiconductor Superlattice and
12: Microstructure
13: \\ and
14: Institute of Semiconductor, Academia Sinica, Beijing 100083,
15: China\\}
16: \author{Zhao-Bin Su}
17: \affiliation{Institute of Theoretical Physics, Academia Sinica, Beijing
18: 100080, China }
19: \affiliation{Center for Advanced Study, Tsinghua University, Beijing 100084,
20: China}
21: \begin{abstract}
22: We investigate the chirality dependence of physical properties of nanotubes
23: which are wrapped by the planar hexagonal lattice including graphite and
24: boron nitride sheet, and reveal its symmetry origin. The observables under
25: consideration are of scalar, vector and tensor types. These exact chirality
26: dependence obtained are useful to verify the experimental and numerical
27: results and propose accurate empirical formulas. Some important features of
28: physical quantities can also be extracted by only considering the symmetry
29: restrictions without complicated calculations.
30: \end{abstract}
31:
32:
33: \pacs{78.67.Ch, 61.46.+w, 73.22.-f} \maketitle
34: Since the discovery of carbon nanotube(CNT)\cite{Iijima}, there have been
35: extensive investigations on the unusual physical properties of this novel
36: nano-material. The simplest CNT is the single wall carbon nanotube (C-SWNT)
37: consisting of only one rolled up graphite sheet, which was first synthesized
38: in 1993\cite{Iijima2,Bethune} and now can be produced in large
39: scale\cite{Bronikowski, Connell}. As an analogue to the graphite, the III-V
40: layered compound, namely, boron nitride(BN) sheet has also the hexagonal
41: lattice structure and can be wrapped into various nanotubes too. The single
42: wall BN nanotube(BN-SWNT) was synthesized in 1996\cite{Loiseau}, which also
43: attracts much attentions very recently\cite{Lauret, Ishigami}. Unlike the
44: nonpolar C-SWNT which could be metallic or semiconducting with a moderate
45: gap\cite{Saito}, this heteropolar nanotube is expected to be always a wide
46: gap semiconductor\cite{Rubio,Ng}. Both of these two kinds of nanotubes have
47: the similar descriptions of their chiral structures and constitute promising
48: materials for wide applications\cite{Baughman}.
49:
50: The properties of nanotubes are determined by their chiral structures.
51: Therefore, to investigate the chirality dependence of various physical
52: quantities of nanotubes is an interesting topic all the time. Particularly,
53: to analyze the chiral composition of bulk samples will be helpful to the
54: production of nanotubes with different species. Numerous experimental and
55: theoretical investigations are devoted to this subject
56: \cite{Bachilo,Telg,Li,Wu,Ando,Kane,Reich,Yang, Ivchenko,Ye}. Most recently,
57: it has been reported that detailed chirality distributions in the bulk
58: samples of the separated C-SWNT can be obtained by the fluorescence
59: measurement and by the resonant Raman spectroscopy\cite{Bachilo,Telg}. Their
60: assignments of the chiral number $(n_1,n_2)$ to the observed spectra are
61: based upon the comparison between the tight-binding (TB) calculations and
62: the experimental data.
63:
64: In the fluorescence experiment\cite{Bachilo}, the authors also gave the
65: empirical formulas of the measured von-Hove singularities with respect to
66: the structures of C-SWNTs, which are expressed in terms of the chiral angle
67: $\theta$ and chiral index $\nu$ defined as
68: $\theta=\mbox{atan}\sqrt{3}n_{2}/(2n_{1}+n_{2})$, and
69: $\nu\equiv\mbox{mod}(n_{1}-n_{2},3)$.
70: Actually, for the purpose to give the dependence of physical observables on
71: the chiral structure of nanotubes, it is more suitable and efficient to
72: use $\nu$ and $\theta$ instead of the chiral numbers $(n_{1},n_{2})$. Since
73: there are many types of nanotubes not being observed in the experiments or
74: numerical calculations, an appropriate empirical formula are quite useful
75: and convenient to predict the properties of those unobserved ones. Generally
76: speaking, there are two possible methods to obtain these empirical formulas
77: for different physical quantities. One is through fitting the numerical and
78: experimental data, the other is from the analytical expansion around the two
79: Dirac points of the hexagonal Brillouin zone based on various TB models and
80: effective mass approximations. However it has technical difficulties in the
81: second way to get higher order terms which are sometimes important.
82:
83: In this paper, we show the symmetry restrictions on the general chirality
84: dependence of physical quantities of various types such as scalar, vector
85: and tensor. This leads to compact forms of the chirality dependence for
86: these observables on $\theta$ and $\nu$. Since the results are \emph{model
87: independent} and exact, they not only can be used to propose accurate
88: empirical formulas from the numerical or experimental data, but also can
89: indicate some important features of physical quantities without complicated
90: calculations. This idea was originated in our previous study of the natural
91: optical activity of C-SWNT\cite{Ye}. Our symmetry analysis is essentially
92: based on the hexagonal structure, and we will consider its application to
93: the properties of both the heteropolar BN-SWNT with large ionicity and
94: nonpolar C-SWNT. We examine the following physical quantities as concrete
95: examples to illustrate our method, such as the excitation gap, electric
96: polarization, dielectric tensor, and piezoelectricity.
97:
98: Considering a hexagonal lattice with base vectors $\vec{a}_{1}$ and
99: $\vec{a}_{2}$(see Fig.~(\ref{fig:sixvectors})), it can be rolled up into a
100: nanotube along the chiral vector $\vec{R}=n_{1}\vec{a}_{1} +
101: n_{2}\vec{a}_{2}$\cite{Saito}, so that each nanotube can be simply
102: represented by a pair of chiral numbers $(n_{1},n_{2})$ and the chiral angle
103: $\theta$ is just the angle between $\vec{R}$
104: and $\vec{a}_{1}$. It then can be established a mapping $f$ from the space of
105: chiral vectors on the planar sheet to that of the nanotube structure in a
106: fixed way of wrapping.
107: However this mapping is not one-to-one. For a given chiral vector
108: $\vec{R}_{0}$ with chiral angle $\theta$ on the BN sheet with $C_{3v}$
109: symmetry, there are another 2 equivalent vectors, obtained by rotating
110: $\vec{R}_{0}$ by $2\pi/3$ successively, corresponding to the same nanotube.
111: These vectors, as shown in Fig.~(\ref{fig:sixvectors}), have the explicit
112: forms
113: \begin{eqnarray}
114: \begin{array}{l}
115: \vec{R}_{0}=n_{1}\vec{a}_{1}+n_{2}\vec{a}_{2},\\
116: \vec{R}_{2}=n_{2}\vec{a}_{1}-(n_{1}+n_{2})\vec{a}_{2},\\
117: \vec{R}_{4}=-(n_{1}+n_2)\vec{a}_{1}+n_{1}\vec{a}_{2}
118: \end{array}\;.
119: \end{eqnarray}
120: All of them constitute an invariant subspace of the three fold symmetry of
121: BN sheet with the same chiral index $\nu$. Another three chiral vectors
122: $\vec{R}_{1}$, $\vec{R}_{3}$ and $\vec{R}_{5}$ in
123: Fig.~(\ref{fig:sixvectors}) have chiral angle $\theta+\pi/3$, $\theta+\pi$
124: and $\theta+5\pi/3$, respectively, and can be written as
125: \begin{eqnarray}
126: \begin{array}{l}
127: \vec{R}_{1}=(n_{1}+n_{2)}\vec{a}_{1}-n_{1}\vec{a}_{2},\\
128: \vec{R}_{3}=-n_{1}\vec{a}_{1}-n_{2}\vec{a}_{2},\\
129: \vec{R}_{5}=-n_{2}\vec{a}_{1}+(n_{1}+n_2)\vec{a}_{2}
130: \end{array}\;.
131: \end{eqnarray}
132: \begin{figure}[htpb]
133: \begin{center}
134: \scalebox{0.65}[0.65]{\includegraphics{sixvectors.eps}}
135: \end{center}
136: \caption{Illustration of hexagonal lattices. The solid and open
137: circles represent boron and nitride atoms for the BN-SWNT and both
138: are the carbon atoms for the C-SWNT.}
139: \label{fig:sixvectors}
140: \end{figure}
141: They also form an invariant subspace, but with an opposite chiral index
142: $-\nu$. The nanotube mapping from $\vec{R}_{1}$ is related to that from
143: $\vec{R}_{0}$ by rotating the latter tube upside down. In addition, there is
144: another special chiral vector $\vec{R}'_{0}=(n_{1}+n_{2}) \vec{a}_{1}-n_{2}
145: \vec{a}_{2}$, which is the reflection of $\vec{R}_{0}$ about the base vector
146: $\vec{a}_1$ on the BN sheet and has the same chiral index $\nu$ of
147: $\vec{R}_{0}$. When mapping onto a nanotube, it corresponds to the mirror
148: image of that from $\vec{R}_{0}$ with respect to the section along the tube
149: axis. For the graphite sheet, since the two atoms in one unit cell are the
150: very same, it has a higher symmetry $C_{6v}$. The six vectors $\vec{R}_{i}
151: (i=0,\cdots,5)$ all correspond to the same nanotube which means that a
152: C-SWNT may be represented by opposite chiral indices, $\nu$ and $-\nu$.
153:
154: Briefly, through the mapping $f$ the manipulations on $\vec{R}$ in the
155: chiral vector set lead to the change of the structure in the nanotube set,
156: and we have the following three observations
157: \renewcommand{\labelenumi}{\roman{enumi}.}
158: \begin{enumerate}
159: \item when $\theta\rightarrow\theta+{2\pi}/{3}$, the
160: nanotube keeps unchanged, and $\nu$ is also unchanged.
161: \item when $\theta\rightarrow\theta+{\pi}/{3}$, the
162: nanotube is rotated upside down, and
163: $\nu\rightarrow -\nu$.
164: \item when $\theta\rightarrow-\theta$, the nanotube is
165: reflected with respect to the section along the tube axis,
166: and $\nu$ is unchanged.
167: \end{enumerate}
168: According to the first observation, we know that the physical quantities of
169: nanotube can always be expanded in terms of the triangle function of
170: $\theta$ with respect to each class of $\nu$
171: \begin{eqnarray}
172: Q^{(\nu)}(\theta)=\sum_{n=0}^{\infty}a_{n}^{(\nu)}\cos(3n\theta)+b_{n}^{(\nu)}\sin(3n\theta)\;.
173: \label{Fourierexpandsion}
174: \end{eqnarray}
175: $Q$ is some physical quantity. The coefficients $a_{n}^{(\nu)}$'s and
176: $b_{n}^{(\nu)}$'s are functions of those chirality independent variables,
177: such as the tube radius $r$ and some external parameters. Our analysis will
178: reveal \emph{ the characteristic role of the chiral index $\nu$ in
179: classifying the chirality dependence}. In the following we will consider
180: some examples to show that the second and third observations together with
181: features of the physical quantity under consideration will reduce the above
182: chiral expressions Eq.(\ref{Fourierexpandsion}). For the convenience of
183: discussion, the Cartesian coordinates for nanotube is introduced as: the
184: tube axis is set as $z$ direction and the cross section is the $xy$ plane
185: with $x$ axis passing through one atom on the tube surface.
186:
187: 1. \emph{Excitation Gap $\Delta$}. We can treat the band gap for the C-SWNT
188: and BN-SWNT as a scalar since its value does not change under rotation or
189: mirror reflection of the nanotube. Hence, its $\theta$-dependence must
190: satisfy
191: \begin{eqnarray}
192: &&\Delta^{(\nu)}(\theta+{\pi}/{3})=\Delta^{(-\nu)}
193: (\theta)\nonumber\\
194: &&\Delta^{(\nu)}(-\theta)=\Delta^{(\nu)}(\theta)\;.
195: \label{gapsymmetry}
196: \end{eqnarray}
197: Notice that in the first expression in Eq.(\ref{gapsymmetry}), $\nu$ becomes
198: $-\nu$ when $\theta \rightarrow \theta+\pi/3$. Then, from
199: Eq.(\ref{Fourierexpandsion}) the chirality dependence of $\Delta$ reads
200: \begin{eqnarray}
201: \Delta^{(\pm)}(\theta)&=& a_{0}\pm
202: a_{1}\cos(3\theta)+a_{2}\cos(6\theta)+\cdots\nonumber\\
203: \Delta^{(0)}(\theta)&=& a_{0}+a_{1}\cos(6\theta)+a_{2}\cos(12\theta)+\cdots\;.
204: \label{gapfourier}
205: \end{eqnarray}
206: Clearly, the coefficients of $\cos(3\theta)$ for $\nu=\pm1$ should have the
207: same value but opposite signs.
208:
209: For the semiconducting C-SWNT with $\nu=\pm 1$, the longitudinal optical
210: excitation can be measured in the fluorescence experiment\cite{Bachilo} and
211: is quite useful in analyzing the chiral composition of bulk samples. In
212: Ref.[\onlinecite{Bachilo}], the authors fitted the experimental data of the
213: von-Hove singularities for $\nu=1$ and $-1$ separately. According to their
214: fitting function, the absolute values of $a_{1}$ are quite different for
215: $\nu=1$ and $\nu=-1$, which does not agree with the symmetry analysis
216: Eq.(\ref{gapfourier}) and suggests that it may need to consider higher order
217: terms like $\cos(6\theta)$ for more accurate fitting functions. In fact we
218: have tried to fit all their data for both $\nu=\pm1$ by just one
219: four-parameters formula with a $\cos(6\theta)$ term
220: \begin{eqnarray}
221: \frac{p_{1}}{r}+\frac{p_{2}}{r^{2}}+\nu\frac{p_{3}\cos3\theta}{r^{2}}
222: +\frac{p_{4}\cos6\theta}{r^{3}}\nonumber
223: \end{eqnarray}
224: with parameters $p_i's$ to be determined, which satisfies the symmetry
225: restriction Eq.(\ref{gapfourier}). It turns out that the result has a
226: smaller root-mean-square deviation than that in Ref. [\onlinecite
227: {Bachilo}].
228:
229: Recently, an interesting temperature dependence of bandgap $\Delta(T)$ was
230: reported for the semiconducting C-SWNT in Ref.[\onlinecite{Capaz}], that is
231: when $\theta$ is small, the temperature dependence of gap is monotonic for
232: $\nu=1$ and nonmonotonic for $\nu=-1$. The different behaviors of
233: $\Delta(T)$ for $\nu=\pm 1$ exist even for very close $\theta$. This could
234: be understood by assuming that both $a_{0}$ and $a_{1}$ are decreasing
235: functions of temperature. Then from Eq.(\ref{gapfourier}), it is clear that
236: the bandgap for $\nu=1$ is monotonically decreasing with temperature. For
237: $\nu=-1$ and small $\theta$, the sign of $a_{1}$ is negative, so that
238: $\Delta$ can be nonmonotonic as temperature varying as a result of the
239: interplay between $a_{0}(T)$ and $a_{1}(T)$. However when $\theta$ is close
240: to $\pi/6$, the $\cos(3\theta)$ is nearly vanishing and only $a_{0}$ takes
241: the dominant role, so that the temperature dependence of bandgap should
242: behave similarly for both $\nu=1$ and $-1$.
243:
244: 2. \emph{Electric Polarization(EP) $\vec{P}$.} The macroscopic electric
245: polarization along the nanotube axis is the consequence of the broken
246: sublattice symmetry of BN-SWNT, which was studied as a geometric phase in
247: Ref.[\onlinecite{Mele}]. They found the sign and size of the longitudinal
248: polarization are dramatically dependent on the chiral structure of nanotube.
249: This in fact has its symmetry origin, and we will show below that some
250: remarkable properties of EP can be extracted by the symmetry analysis.
251:
252: Due to the helical symmetry of the nanotube\cite{ctwhite,Damnjanovic}, this
253: EP vector only exists in the tube axis direction($z$-axis) and vanishes in
254: the cross section($xy$ plane), i.e., $P_{z}\ne 0$ and $P_{x,y}=0$. According
255: to observations ii and iii and the vector nature of EP, we have
256: \begin{eqnarray}
257: P^{(\nu)}_{z}(\theta+{\pi}/{3})&=&-P^{(-\nu)}_{z}(\theta)\nonumber\\
258: P^{(\nu)}_{z}(-\theta)&=&P^{(\nu)}_{z}(\theta)\;,
259: \label{epsymmetry}
260: \end{eqnarray}
261: which leads to the chirality dependence of $P_{z}$
262: \begin{eqnarray}
263: P_{z}^{(0)}(\theta)&=& a_{1}\cos(3\theta)+a_{3}\cos(9\theta)+\cdots\nonumber\\
264: P_{z}^{(\pm)}(\theta)&=& \pm a_{0}+ a_{1}\cos(3\theta)\pm a_{2}\cos(6\theta)+\cdots
265: \label{ep}
266: \end{eqnarray}
267: Eq.(\ref{ep}) shows evidently that the armchair tubes($n_{1}=n_{2}=N$) with
268: $\theta=\pi/6$ and $\nu=0$ have no electric polarization. Another type of
269: achiral tube, zigzag tubes($n_{1}=N,n_{2}=0$), show different pictures. If
270: we assume reasonably the coefficients of higher order terms are small enough
271: comparing to the zero order term $a_{0}$, we then obtain $P^{(0)}_{z}
272: \approx 0$ for $\mbox{mod}(N,3)=0$, and $P^{(\pm)}_{z}\approx \pm a_{0}$ for
273: $\mbox{mod}(N,3)=\pm 1$, i.e., when $N$ is increasing, the EP is oscillating
274: among $1$, $0$ and $-1$, which is just the striking result of BN-SWNT found
275: in Ref.[\onlinecite{Mele}].
276:
277: For the C-SWNT, the six fold $C_{6v}$ symmetry in the chiral vector set
278: guarantees $P^{(0)}_{z}(\theta+\pi/3) =P^{(-\nu)}_{z}(\theta)$, which
279: combined with Eq.~(\ref{epsymmetry}) leads to $P^{(\nu)}_{z}=0$, i.e.,
280: \emph{no EP in C-SWNT}.
281:
282: 3. \emph{Dielectric tensor $\epsilon$}. This second rank tensor is
283: restricted to have the following form by the helical symmetry of
284: nanotube\cite{Damnjanovic}:
285: \begin{eqnarray}
286: \epsilon=\left(
287: \begin{array}{ccc}
288: \epsilon_{xx}&\epsilon_{xy}&0\\
289: \epsilon_{yx}&\epsilon_{yy}&0\\
290: 0&0&\epsilon_{zz}
291: \end{array}
292: \right)
293: \label{dtsymmtry}
294: \end{eqnarray}
295: with $\epsilon_{xx}=\epsilon_{yy}$ and $\epsilon_{xy}=-\epsilon_{yx}$. The
296: diagonal matrix elements of $\epsilon$, denoted by $\epsilon_{ii}$ with
297: $i=x,y,z$, have the same chirality dependence. Similar to the analysis of EP
298: vector, we obtain
299: \begin{eqnarray}
300: &&\epsilon^{(\nu)}_{ii}(\theta+{\pi}/{3})=\epsilon^{(-\nu)}_{ii}(\theta)\nonumber\\
301: &&\epsilon^{(\nu)}_{ii}(-\theta)=\epsilon^{(\nu)}_{ii}(\theta)\nonumber\\
302: &&\epsilon^{(\nu)}_{xy}(\theta+{\pi}/{3})=-\epsilon^{(-\nu)}_{xy}(\theta)\nonumber\\
303: &&\epsilon^{(\nu)}_{xy}(-\theta)=-\epsilon^{(\nu)}_{xy}(\theta)
304: \label{dt}
305: \end{eqnarray}
306: by noticing that the diagonal term is unchanged when the tube is reversed or
307: reflected, and the off-diagonal term get its sign changed. Then
308: \begin{eqnarray}
309: \epsilon_{ii}^{(\pm)}(\theta)&=&a_0\pm a_1\cos(3\theta)+a_2\cos(6\theta)+\cdots\nonumber\\
310: \epsilon_{ii}^{(0)}(\theta)&=&a_0+ a_2\cos(6\theta)+a_4\cos(12\theta)+\cdots\nonumber\\
311: \epsilon_{xy}^{(\pm)}(\theta)&=&b_1\sin (3\theta)\pm b_2\sin (6\theta)+\cdots\nonumber\\
312: \epsilon_{xy}^{(0)}(\theta)&=&b_1\sin (3\theta)+b_3\sin(9\theta)+\cdots
313: \label{dtfourier}
314: \end{eqnarray}
315: The coefficients $a_{n}'s$ and $b_{n}'s$ in the above four expressions have
316: no direct relationships. Obviously, the diagonal and off-diagonal terms have
317: quite different chirality dependence and the off-diagonal terms vanish for
318: the zigzag tube whose chiral angle is $0$.
319:
320: The discussion above is for the heteropolar BN-SWNT. For the C-SWNT, the
321: higher symmetry requires $\epsilon^{(\nu)}_{xy}(\theta+{\pi}/
322: {3})=\epsilon^{(-\nu)}_{xy}(\theta)$, which together with Eq.~(\ref{dt})
323: leads to \emph{ $\epsilon^{(\nu)}_{xy}=0$ for any kind of C-SWNT}.
324:
325: 4. \emph{Piezoelectricity $e$.} Piezoelectricity is the response of the EP
326: of the material to the mechanical strain, which is a third rank tensor
327: defined by the derivative of EP vector with respect to the elastic strain
328: tensor $u$ as $e_{i,jk}={\partial P_{i}}/{\partial u_{jk}}$.
329: For the quasi-one-dimensional nanotube, we are concerned with the response
330: of EP along the $z$ direction to uniaxial($s$) and torsional ($t$) strains,
331: \begin{eqnarray}
332: e_{s}={\partial P_{z}}/{\partial u_{s}}, \hspace{0.5cm}
333: e_{t}={\partial P_{z}}/{\partial u_{t}}\;.
334: \label{piezotube}
335: \end{eqnarray}
336: $u_{s}$ is the stretch strain along the tube, and $u_{t}$ is the torsional
337: strain around the tube circumference. They can be related to the second rank
338: tensor $u_{ij}$ in the Cartesian coordinates system through
339: equations $u_{s}=u_{zz}$ and $u_{t}=(xu_{zy}-yu_{zx})/{r^{2}}$
340: with $r^{2}=x^{2}+y^{2}$. Then, it is clear that how
341: $u_{s}$ and $u_{t}$ transform under the rotation of the tube upside down or
342: the mirror reflection with respect to the section along the tube axis.
343: Consequently, we have
344: \begin{eqnarray}
345: &&e^{(\nu)}_{s}(\theta+{\pi}/{3})=-e^{(-\nu)}_{s}(\theta) \nonumber \\
346: &&e^{(\nu)}_{s}(-\theta)=e^{(\nu)}_{s}(\theta) \nonumber \\
347: &&e^{(\nu)}_{t}(\theta+{\pi}/{3})=-e^{(-\nu)}_{t}(\theta) \nonumber \\
348: &&e^{(\nu)}_{t}(-\theta)=-e^{(\nu)}_{t}(\theta)\;,
349: \label{piezosymmetry}
350: \end{eqnarray}
351: and the chirality dependence of piezoelectricity then reads
352: \begin{eqnarray}
353: &&e^{(\pm)}_{s}(\theta)= \pm a_{0}+a_{1}\cos(3\theta)\pm a_{2}\cos(6\theta)+\cdots\nonumber\\
354: &&e^{(0)}_{s}(\theta)= a_{1}\cos(3\theta)+a_{3}\cos(9\theta)+\cdots\nonumber\\
355: &&e^{(\pm)}_{t}(\theta)= b_{1}\sin(3\theta)\pm b_{2}\sin(6\theta)+b_{3}\sin(9\theta)+\cdots\nonumber\\
356: &&e^{(0)}_{t}(\theta)= b_{1}\sin(3\theta)+ b_{3}\sin(9\theta)+\cdots\;.
357: \label{piezoexpression}
358: \end{eqnarray}
359: Eq.~(\ref{piezoexpression}) implies the zigzag($\theta=0$) tube can only
360: have piezoelectric response to the longitudinal stretch and no response to
361: the torsion strain around the circumference. On the contrary, the armchair
362: ($\theta=\pi/6$ and $\nu=0$) tube has response to torsion strain but not to
363: stretch strain. These conclusions are in consistent with the numerical
364: results of the \textit{ab initio} and TB calculations in
365: Ref.[\onlinecite{Sai}], where the chiral angle by definition has a
366: difference $\pi/2$ from ours. Apart from this difference in the $\theta$
367: definition, in Ref.[\onlinecite{Sai}] only $\cos(3\theta)$ and
368: $\sin(3\theta)$ terms appear in the chirality dependence of $e_{s}$ and
369: $e_{t}$, respectively, while there are extra terms according to our results
370: Eq.~(\ref{piezoexpression}). In fact, their $\theta$-dependence of $e_s$ and
371: $e_t$ do not agree with their numerical data very well. The reason for this
372: difference is that in Ref.[\onlinecite{Sai}] the chirality dependence of
373: $e_{t}$ and $e_{s}$ is inherited from the BN planar sheet\cite{3theta}, so
374: that only the $3\theta$ terms are permitted by the $C_{3v}$ symmetry of BN
375: sheet. However when the sheet is rolled up, this planar $C_{3v}$ symmetry is
376: broken, therefore those $C_{3v}$-forbidden terms are not forbidden any more,
377: although they should be small by the continuous argument from sheet to
378: nanotube, i.e., they vanish when the tube radius tends to infinity.
379:
380: For the C-SWNT, the symmetry leads to additional restriction $e^{(\nu)}
381: _{\alpha} (\theta+{\pi}/{3})= e^{(-\nu)} _{\alpha} (\theta)$, which combined
382: with Eq.~(\ref{piezosymmetry}) requires $e^{(\nu)}_{\alpha} (\theta)=0$ for
383: both $\alpha=s$ and $t$, namely, \emph{for C-SWNT there should be no
384: piezoelectricity due to its nonpolar feature.}
385:
386: In the above discussion, the expansion coefficients $a^{(\nu)}_{n}$'s and
387: $b^{(\nu)}_{n}$'s can not be determined by the symmetry argument which in
388: fact depend on the chirality independent parameters, such as the tube
389: diameter, magnetic flux, temperature, and so on. By tuning the external
390: parameters, one can adjust the magnitude of $a^{(\nu)}_{n}$'s and
391: $b^{(\nu)}_{n}$'s, which will be helpful to identify the chirality of the
392: tubes. As examples, one could refer to Ref.[\onlinecite{Ye,Capaz}] to find the
393: different dependence of these coefficients on magnetic flux or temperature.
394:
395: As a conclusion, we give the explicit $\theta$-dependence of physical
396: quantities for different values of $\nu$ by a symmetry analysis. It shows
397: clearly that the chiral index $\nu$ plays a characteristic role in
398: describing the chirality dependence. This model independent method may be
399: used to verify the numerical and experimental data and also can give rise to
400: some important properties qualitatively without complicated calculations.
401: In addition, this method is not restricted to the examples illustrated in
402: this paper and could be extended to other situations.
403:
404: The author F. Ye would like to thank Dr. J. Wu at CASTU for helpful
405: discussion.
406: \begin{thebibliography}{}
407: \bibitem{Iijima} S. Iijima, Nature \textbf{354}, 56(1991)
408: \bibitem{Iijima2} S. Iijima and T. Ichihashi, Nature \textbf{363},
409: 603(1993)
410: \bibitem{Bethune} D. S. Bethune, \textit{et al.}, Nature \textbf{363},
411: 605(1993)
412: \bibitem{Bronikowski} M. J. Bronikowski, \textit{et al.}, J. Vac. Sci. Technol. A \textbf{19}, 1800(2001)
413: \bibitem{Connell} M. O' Connell \textit{et al.}, Science \textbf{297}, 593(2002)
414: \bibitem{Loiseau} A. Loiseau, F. Willaime, N. Demoncy, G. Hug, and H.
415: Pascard, Phys. Rev. Lett.
416: \textbf{76}, 4737(1996)
417: \bibitem{Lauret} J. S. Lauret, \textit{et. al.}, Phys. Rev. Lett.
418: \textbf{94}, 037405(2005)
419: \bibitem{Ishigami} M. Ishigami, J. D. Sau, S. Aloni, M. L. Cohen, and A.
420: Zettl, Phys. Rev. Lett.
421: \textbf{94}, 056804(2005)
422: \bibitem{Saito} S. Saito, G.Dresselhaus and M. S. Dresselhaus, {\it Physical Properties of Carbon Nanotube}(Imperial College Press, London, 1998)
423: \bibitem{Rubio} A. Rubio, J. L. Corkill, and M. L. Cohen, Phys. Rev. B
424: \textbf{49}, R5081(1994)
425: \bibitem{Ng} M. F. Ng and R. Q. Zhang, Phys. Rev. B \textbf{69},
426: 115417(2004)
427: \bibitem{Baughman} For the potential applications of CNT, one can refer
428: to the review article, R. H. Baughman, A. A. Zakhidov, and W. A. de Heer,
429: Science \textbf{297}, 787(2002)
430: \bibitem{Bachilo} Sergei M. Bachilo, \textit{et al.} Science \textbf{298}, 2361(2002)
431: \bibitem{Telg}H. Telg, J. Maultzsch, S. Reich, F. Hennrich, and C. Thomsen,
432: Phys. Rev. Lett. \textbf{93}, 177401 (2004)
433: \bibitem{Li} Lain-Jong Li, R. J. Nicholas, R. S Deacon, and P. A. Shields, Phys. Rev. Lett. \textbf{93}, 156104(2004)
434: \bibitem{Wu} J. Wu, \textit{et al}, Phys. Rev. Lett. \textbf{93}, 017404(2004)
435: \bibitem{Ando} H. Ajiki and T. Ando, J. Phys. Soc. Jap. \textbf{65}, 505(1996)
436: \bibitem{Kane} C. L. Kane and E. J. Mele, Phys. Rev. Lett. \textbf{78}, 1932(1997)
437: \bibitem{Reich} S. Reich and C. Thomsen, Phys. Rev. B \textbf{62}, 4273(2000)
438: \bibitem{Yang} Liu Yang and Jie Han, Phys. Rev. Lett. \textbf{85}, 154(2000)
439: \bibitem{Ivchenko} E. L. Ivchenko and B. Spivak, Phys. Rev. B \textbf{66}, 155404(2002)
440: \bibitem{Ye} Fei Ye, Bing-Shen Wang and Zhao-Bin Su, Phys. Rev. B \textbf{70}, 153406(2004)
441: \bibitem{Capaz} Rodrigo B. Capaz, C. D. Spataru, P. Tangney, M. L.
442: Cohen, and S. G. Louie, Phys. Rev. Lett.
443: \textbf{94}, 036801(2005)
444: \bibitem{Sai} Na Sai and E.J. Mele, Phys. Rev. B \textbf{68}, 241405(R)(2003)
445: \bibitem{Mele} E. J. Mele and Petr Kr\'al, Phys. Rev. Lett. \textbf{88}, 056803(2002)
446: \bibitem{ctwhite} C. T. White, D. H. Robertson and J. W. Mintmire, Phys. Rev. B \textbf{47}, R5485(1993)
447: \bibitem{Damnjanovic}M. Damnjanovi\'{c}, I. Milo\v{s}evi\'c, T. Vukovi\'c and R. Sredanovi\'c, Phys. Rev. B \textbf{60}, 2728(1999)
448: \bibitem{3theta} It can be obtained through a rotation by $\theta$ of the piezoelectric tensor of the BN sheet.
449: \end{thebibliography}
450: \end{document}
451: