1: \documentclass[twocolumn,showpacs,showkeys]{revtex4}

2: \usepackage{graphicx}

3: \input{psfig.sty}

4: \newcommand {\gs}{\

5:   \raisebox{-0.2ex}{$\stackrel{\scriptstyle>}{\scriptstyle\sim}$}\ }

6: \newcommand{\ls}{\

7:   \raisebox{-0.2ex}{$\stackrel{\scriptstyle<}{\scriptstyle\sim}$}\ }

8: \parindent=0.5cm

9: \parskip=0.2 cm

10: \newcommand{\la}[1]{\label{#1}}

11: \newcommand{\bastar}{\begin{eqnarray*}}

12: \newcommand{\eastar}{\end{eqnarray*}}

13: \newskip\humongous \humongous=0pt plus 1000pt minus 1000pt

14: \def\caja{\mathsurround=0pt}

15: \def\eqalign#1{\,\vcenter{\openup1\jot \caja

16:         \ialign{\strut \hfil$\displaystyle{##}$&$

17:         \displaystyle{{}##}$\hfil\crcr#1\crcr}}\,}

18: \newif\ifdtup

19: \def\panorama{\global\dtuptrue \openup1\jot \caja

20:         \everycr{\noalign{\ifdtup \global\dtupfalse

21:         \vskip-\lineskiplimit \vskip\normallineskiplimit

22:         \else \penalty\interdisplaylinepenalty \fi}}}

23: \def\eqalignno#1{\panorama \tabskip=\humongous

24:         \halign to\displaywidth{\hfil$\displaystyle{##}$

25:         \tabskip=0pt&$\displaystyle{{}##}$\hfil

26:         \tabskip=\humongous&\llap{$##$}\tabskip=0pt

27:         \crcr#1\crcr}}

28: \relax

29: %%%%%%%%%%%%%%%%%%%%%%+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

30: \newcommand{\be}{\begin{equation}}

31: \newcommand{\ee}{\end{equation}}

32: \newcommand{\bea}{\begin{eqnarray}}

33: \newcommand{\eea}{\end{eqnarray}}

34: \newcommand{\X}{{\vec X}}

35: \newcommand{\pro}{\partial}

36: \newcommand{\n}{\hat n}

37: \newcommand{\oneg}{\displaystyle\frac{1}{g}}

38: \newcommand{\alfa}{\hat{\alpha}}

39: \newcommand{\D}{{\hat D}}

40: \newcommand{\hfi}{{\hat \phi}}

41: \newcommand{\B}{{\vec B}}

42: \newcommand{\C}{\stackrel{\ast}{C}}

43: \newcommand{\ksi}{{\hat \xi}}

44: \newcommand{\A}{{\vec A}}

45: \newcommand{\valpha}{{\vec \alpha}}

46: \newcommand{\vbeta}{{\vec \beta}}

47: \newcommand{\dfrac}{\displaystyle\frac}

48: \newcommand{\ba}{\begin{array}}

49: \newcommand{\ea}{\end{array}}

50: \newcommand{\dpst}{\displaystyle}

51: \newcommand{\vareps}{\varepsilon}

52: \newcommand{\uone}{\mbox{$U(1)$\,\,}}

53: \newcommand{\suu}{\mbox{$SU(2)$\,\,}}

54: \newcommand{\nn}{\nonumber}

55: \newcommand{\hn}{\hat n}

56: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

57: \begin{document}

58: \title{Knots in Condensed Matters}

59: \author{Y. M. Cho}

60: \email{ymcho@yongmin.snu.ac.kr}

61: \affiliation{Center for Theoretical Physics and School of Physics \\

62: College of Natural Sciences, 

63: Seoul National University,

64: Seoul 151-742, Korea  \\

65: }

66: \begin{abstract}

67: ~~~~~We propose two types of topologically stable knot solitons 

68: in condensed matters, one in two-component

69: Bose-Einstein condensates and one in two-gap superconductors. 

70: We identify the knot in Bose-Einstein condensates as

71: a twisted vorticity flux ring and the knot in two-gap superconductors

72: as a twisted magnetic flux ring.

73: In both cases we show that there is a remarkable interplay between

74: topology and dynamics which transforms the topologcal stability

75: to the dynamical stability, and vise versa.  

76: We discuss how these knots can be constructed in the spin-1/2 

77: condensate of $^{87}{\rm Rb}$ atoms and in two-gap superconductor

78: of $MgB_2$.

79: \end{abstract}

80: \pacs{03.75.Fi, 05.30.Jp, 67.40.Vs, 74.72,-h}

81: \keywords{Topological knot in BEC, Knot in two-gap superconductor}

82: \maketitle

83: 

84: 

85: \section{Introduction}

86: 

87: Topological objects (monopoles, vortices, skyrmions, etc.) 

88: have played an increasingly important role

89: in physics \cite{dirac,abri}. In particular finite energy topological

90: objects have been widely studied in theoretical physics

91: \cite{skyrme,thooft}. A recent addition to this family of finite

92: energy solitons has been the knots \cite{fadd1,prl01}.

93: The interest on these topological objects,

94: however, has been confined mainly to theoretical physics, 

95: because most of them (exept the vortices) exist in

96: ``hypothetical'' worlds which are very difficult to

97: create in laboratories. The only

98: topological objects which one can realistically expect to exist

99: in the ``standard'' models are the electroweak 

100: monopoles and dyons in Weinberg-Salam model

101: which have a non-trivial $W_\mu$

102: and $Z_\mu$ dressing \cite{plb97,yang}. Unfortunately these objects

103: could carry an infinite energy, which makes it impossible to

104: create them experimentally.

105: 

106: Fortunately the recent experimental advances in condensed matter

107: physics, in particular the construction of multi-component Bose-Einstein 

108: condensates (BEC) made of $^{87}{\rm Rb}$ \cite{bec,ruo}

109: and two-gap superconductors made of $MgB_2$ \cite{exp1,exp2}, 

110: have widely opened new opportunities for

111: us to create new topological objects experimentally which so far

112: have been only of theoretical interest. 

113: The purpose of this paper is to argue that these new 

114: multi-component condensates could allow us to have real

115: knots, topologically stable finite energy 3-dimensional solitons.

116: This is because, due to the multi-component structure, the new condensates

117: have a non-Abelian symmetry which provides the needed topology

118: for the stable knots.

119: 

120: To understand how the realistic knots can appear in these

121: condensed matters, it is necessary to understand 

122: the prototype Faddeev-Niemi knot in Skyrme theory. 

123: The Skyrme theory is well known to have a magnetic vortex 

124: known as the baby skyrmion \cite{piet}, and the

125: Faddeev-Niemi knot can be identified as a twisted vortex ring

126: made of a helical baby skyrmion \cite{prl01,plb04}. In the following

127: we show how one can construct similar knots from helical vortices 

128: in two-component Bose-Einstein condensates and two-gap superconductors. 

129: 

130: The paper is organized as follows. In Section II we discuss

131: how the helical vortex can give rise to the prototype knot 

132: in Skyrme theory for later purpose.

133: In Section III we discuss gauge theory of two-component BEC

134: which has the vorticity interaction, and show that the theory 

135: is closely related to Skyrme theory. With this we show that

136: the theory allows a topological knot similar to the one in 

137: Skyrme theory. We also identify that this knot is a vorticity knot 

138: very similar to the one in Gross-Pitaevskii theory of two-component BEC.

139: In Section IV we discuss the Landau-Ginzburg theory of two-gap

140: superconductor and show that the theory is closely related to 

141: the gauge theory of two-component BEC. With this we argue 

142: that the theory can also admit a knot, a twisted magnetic vortex 

143: ring, similar to the vorticity knot in two-component BEC.

144: In doing so we also establish the non-Abelian flux quantization

145: in two-gap superconductor. We demonstrate the existence of magnetic 

146: vortex whose flux is quantized in the unit $4\pi/g$, not $2\pi/g$, 

147: in two-gap superconductor.

148: Finally in Section V we discuss physical implications of 

149: our results.

150: 

151: \section{Knot in Skyrme theory}

152: 

153: The Skyrme theory has a rich topological structure.

154: It has skyrmion and baby skyrmion \cite{skyrme,piet}.

155: But recently it has been shown that the theory allows

156: a knotted soliton identical to

157: the Faddeev-Niemi knot in Skyrme-Faddeev

158: non-linear sigma model, which can be identified as a twisted 

159: vortex ring made of helical baby skyrmion \cite{prl01,plb04}. 

160: To see this, we review the knot in Skyrme theory first. 

161: 

162: Let $\omega$ and $\hn$ be the scalar field and 

163: the non-linear sigma field in Skyrme theory. With

164: \bea

165: &U = \exp (\omega \dfrac{\vec \sigma}{2i} \cdot \hat n)

166: = \cos \dfrac{\omega}{2} - i (\vec \sigma \cdot \hat n) 

167: \sin \dfrac{\omega}{2}, \nn\\

168: &L_\mu = U\partial_\mu U^{\dagger}, ~~~~~({\hat n}^2 = 1)

169: \eea

170: the Skyrme Lagrangian is expressed as

171: \bea

172: &{\cal L} = \dfrac{\mu^2}{4} {\rm tr} ~L_\mu^2 + \dfrac{\alpha}{32}

173: {\rm tr} \left( \left[ L_\mu, L_\nu \right] \right)^2  \nn \\

174: &= - \dfrac{\mu^2}{4} \Big[ \dfrac{1}{2} (\partial_\mu \omega)^2

175: +(1-\cos\omega)(\partial_\mu \hat n)^2 \Big]  \nn \\

176: & -\dfrac{\alpha}{16} \Big[ \dfrac{1-\cos\omega}{2} (\partial_\mu

177: \omega \partial_\nu \hat n - \partial_\nu \omega \partial_\mu \hat n)^2  \nn\\

178: & + (1-\cos\omega)^2 (\partial_\mu \hat n \times \partial_\nu \hat

179: n)^2 \Big].

180: \label{slag}

181: \eea

182: The equation of motion is given by

183: \bea

184: &\dfrac{\mu^2}{4} \left[ \partial^2 \omega -\sin\omega 

185: (\partial_\mu \hat n)^2 \right] \nn\\

186: &+\dfrac{\alpha}{32} \sin\omega (\partial_\mu

187: \omega \partial_\nu \hat n  -\partial_\nu \omega \partial_\mu \hat n)^2 \nn\\

188: &+\dfrac{\alpha}{8} (1-\cos\omega)

189: \partial_\mu \big[ (\partial_\mu

190: \omega \partial_\nu \hat n  -\partial_\nu \omega \partial_\mu \hat

191: n) \cdot \partial_\nu \hat n \big] \nn\\

192: & - \dfrac{\alpha}{8} (1-\cos\omega) \sin\omega (\partial_\mu \hat n \times

193: \partial_\nu \hat n)^2 =0, \nn \\

194: &\partial_\mu \Big\{ \dfrac{\mu^2}{4} (1-\cos\omega) \hat n \times

195: \partial_\mu \hat n \nn\\

196: & + \dfrac{\alpha}{16} (1-\cos\omega)

197: \big[ (\partial_\nu \omega)^2

198: \hat n \times \partial_\mu \hat n

199: -(\partial_\mu \omega \partial_\nu \omega) \hat n \times

200: \partial_\nu \hat n \big] \nn\\

201: &+\dfrac{\alpha}{8} (1-\cos\omega)^2

202: (\hat n \cdot \partial_\mu \hat n \times

203: \partial_\nu \hat n) \partial_\nu \hat n \Big\}=0.

204: \label{seq}

205: \eea

206: With the spherically symmetric ansatz

207: \bea

208: \omega = \omega (r),~~~~~\hat n = \hat r,

209: \eea

210: (3) is reduced to

211: \bea

212: &\dfrac{d^2 \omega}{dr^2} +\dfrac{2}{r} \dfrac{d\omega}{dr} -\dfrac{2

213: \sin\omega}{r^2}

214: -\dfrac{\alpha}{\mu^2} \Big[ \dfrac{\sin^2 (\omega /2)}{r^2}

215: \dfrac{d^2 \omega}{dr^2} \nn\\

216: &+\dfrac{\sin\omega}{2 r^2} (\dfrac{d\omega}{dr})^2

217: -\dfrac{2 \sin\omega \sin^2 (\omega /2)}{r^4} \Big] =0.

218: \eea

219: Imposing the boundary condition $\omega(0)= 2\pi$ and $\omega(\infty)= 0$,

220: one has the well-known skyrmion \cite{skyrme}.

221: 

222: A remarkable point of Skyrme theory is that $\omega=\pi$,

223: independent of $\hn$, becomes a classical solution \cite{prl01}.

224: So restricting $\omega$ to $\pi$, one can reduce the Lagrangian 

225: (\ref{slag}) to the Skyrme-Faddeev Lagrangian

226: \bea

227: {\cal L} \rightarrow -\dfrac{\mu^2}{2} (\partial_\mu

228: \hat n)^2-\dfrac{\alpha}{4}(\partial_\mu \hat n \times

229: \partial_\nu \hat n)^2,

230: \label{sflag}

231: \eea

232: in which case the equation of motion (\ref{seq}) is reduced to

233: \bea

234: &\hn \times

235: \partial^2 \hn + \dfrac{\alpha}{\mu^2} ( \partial_\mu N_{\mu\nu} )

236: \partial_\nu \hn = 0, \nn\\

237: &N_{\mu\nu} = \hn \cdot (\partial_\mu \hn \times \partial_\nu \hn)

238: =\pro_\mu C_\nu - \pro_\nu C_\mu.

239: \label{sfeq}

240: \eea

241: Notice that since $N_{\mu\nu}$ forms a closed two-form, 

242: it always admits a $U(1)$ potential $C_\mu$.

243: 

244: The equation (\ref{sfeq}) allows non-Abelian monopole,

245: baby skyrmion, helical baby skyrmion, and Faddeev-Niemi

246: knot as its solutions \cite{prl01,plb04}. But for our purpose it is 

247: important to understand the helical baby skyrmion, because this 

248: plays a crucial role for us to construct the knot. 

249: So we review the helical baby skyrmion.

250: 

251: To construct the desired helical baby skyrmion

252: let $(\varrho,\varphi,z)$ the cylindrical coodinates,

253: and choose the ansatz

254: \bea

255: &\n= \Bigg(\matrix{\sin{f(\varrho)}\cos{(mkz+n\varphi)} \cr

256: \sin{f(\varrho)}\sin{(mkz+n\varphi)} \cr \cos{f(\varrho)}}\Bigg).

257: \label{hvans}

258: \eea

259: With this the equation (\ref{sfeq}) is reduced to

260: \bea

261: &\Big(1+(m^2 k^2+\dfrac{n^2}{\varrho^2})

262: \dfrac{\sin^2{f}}{g^2 \rho^2}\Big) \ddot{f}

263: + \Big( \dfrac{1}{\varrho}+ 2\dfrac{\dot{\rho}}{\rho} \nn\\

264: &+(m^2 k^2+\dfrac{n^2}{\varrho^2})

265: \dfrac{\sin{f}\cos{f}}{g^2 \rho^2} \dot{f}

266: + \dfrac{1}{\varrho} (m^2 k^2-\dfrac{n^2}{\varrho^2})

267: \dfrac{\sin^2{f}}{g^2 \rho^2} \Big) \dot{f} \nn\\

268: &- (m^2 k^2+\dfrac{n^2}{\varrho^2}) \sin{f}\cos{f}=0.

269: \label{hveq}

270: \eea

271: So with the boundary condition

272: \bea

273: f(0)=\pi,~~f(\infty)=0,

274: \label{bc}

275: \eea

276: we obtain the non-Abelian vortex solutions shown in Fig.\ref{hbs}.

277: Notice that, when $m=0$, the solution describes

278: the well-known baby skyrmion \cite{piet}. But when $m$ is not zero,

279: it describes a helical baby skyrmion, a twisted magnetic vortex 

280: which is periodic in $z$-coodinate \cite{plb04}.

281: In this case, the vortex has a quantized magnetic flux

282: not only along the $z$-axis but also around the $z$-axis.

283: 

284: \begin{figure}[t]

285: \includegraphics[scale=0.5]{hbs.eps}

286: \caption{The baby skyrmion (dashed line) with $m=0,n=1$

287: and the helical baby skyrmion (solid line) with $m=n=1$ in Skyrme

288: theory. Here $\varrho$ is in the unit ${\sqrt \alpha}/\mu$

289: and $k=0.8~\mu/{\sqrt \alpha}$.}

290: \label{hbs}

291: \end{figure}

292: 

293: The helical baby skyrmion will become

294: unstable unless the periodicity condition is enforced by

295: hand. But it plays a very important role

296: because one can make it a vortex ring by smoothly connecting

297: two periodic ends. In this case the vortex ring 

298: acquires the topology of a knot, and thus becomes a knot 

299: itself \cite{prl01,plb04}. In fact it becomes a knot made of two

300: magnetic fluxes linked together, whose knot topology is described by

301: the Chern-Simon index of the potential $C_\mu$,

302: \bea

303: &Q = \dfrac{1}{32\pi^2} \int \epsilon_{ijk} C_i N_{jk} d^3x = mn,

304: \label{kqn}

305: \eea

306: which describes the Hopf mapping $\pi_3(S^2)$ defined by $\hn$.

307: 

308: Clearly the knot has a topological stability, because two

309: flux rings linked together can not be disjointed by a smooth

310: deformation of the field configuration.

311: Moreover the topological stability is 

312: backed up by a dynamical stability. This is because

313: the knot can be viewed as two magnetic fluxes linked together,

314: and the magnetic flux trapped in the knot disk

315: can not be squeezed out. This provides

316: a stablizing repulsive force which prevent the collapse 

317: of the knot \cite{plb04}.

318: 

319: From our discussion it becomes clear that the existence of 

320: the helical baby skyrmion is crucial for the existence of a topologically

321: stable knot. In the following we show that an identical mechanism 

322: works in two-component BEC and two-gap superconductor 

323: which guarantees the existence of a stable knot.

324: 

325: \section{Knot in Two-component Bose-Einstein Condensates}

326: 

327: The recent advent of multi-component BEC (in particular

328: the spin-1/2 condensate of $^{87}{\rm Rb}$ atoms)

329: has widely opened a new opportunity for

330: us to study novel topological objects which can not be

331: realized in ordinary (one-component) BEC \cite{bec,ruo}.

332: This is because the multi-component BEC naturally allows

333: a non-Abelian structure which accommodates a non-trivial

334: topological objects, in particular a topolgical knot which is

335: very similar to the knot in Skyrme theory \cite{bec1,bec5}.

336:  

337: There are two competing theories of two-component BEC,

338: the popular Gross-Pitaevskii theory \cite{ruo} and the gauge theory

339: of two-component BEC proposed recently \cite{bec1}. 

340: Both theories predict topological knots. 

341: Many authors have already claimed 

342: the existence of a knot in Gross-Pitaevskii theory \cite{ruo}.

343: Moreover it has been shown that this knot can be identified 

344: as a vorticity knot which is made of

345: two vorticity fluxes linked together, whose topology $\pi_3(S^2)$

346: is fixed by the Chern-Simon index of the velocity potential

347: of the condensate \cite{bec5}.

348: 

349: So in this section we discuss the gauge theory of two-component BEC,

350: and show that this theory also has a vorticity knot very similar 

351: to the knot in Gross-Pitaevskii theory. 

352: Consider a ``charged'' two-component BEC

353: described by a complex doublet $\phi$ interacting

354: ``electromagnetically'', which can be described by the following

355: non-relativistic gauged Gross-Pitaevskii type Lagrangian (we

356: will discuss the relativistic generalization later)

357: \bea

358: &{\cal L} = i \dfrac {\hbar}{2} \Big[\phi^\dag (\tilde D_t \phi)

359: -(\tilde D_t \phi)^\dag \phi \Big] 

360: -  \dfrac {\hbar^2}{2m} |\tilde D_i \phi|^2 \nn\\

361: & + \mu^2 \phi^\dag \phi - \dfrac {\lambda}{2}

362: (\phi^\dag \phi)^2 - \dfrac {1}{4} \tilde F_{\mu \nu} ^2,

363: \label{beclag}

364: \eea

365: where $\tilde D_\mu \phi = ( \pro_\mu - i g \tilde A_\mu ) \phi$,

366: and $\mu^2$ and $\lambda$ are the chemical potential 

367: and the quartic coupling constant.

368: Normalizing $\phi$ to $(\sqrt{2m}/\hbar)\phi$ and putting

369: \bea

370: \phi = \dfrac {1}{\sqrt 2} \rho \xi , ~~~~~(\xi^\dag \xi = 1)

371: \eea

372: we have the following Hamiltonian in the static limit,

373: \bea

374: &{\cal H} =  \dfrac {1}{2} (\pro_i \rho)^2

375: + \dfrac {\rho^2}{2} |\tilde D_i \xi |^2 \nn \\

376: &- \dfrac {\mu^2}{2} \rho^2 + \dfrac{\lambda}{8} \rho^4

377: + \dfrac {1}{4} \tilde F_{ij} ^2.

378: \label{becham}

379: \eea

380: Notice that here we have also

381: normalized $\mu^2$ and $\lambda$ to

382: $(\hbar^2/2m) \mu^2$ and $(\hbar^2/2m)^2 \lambda$.

383: From now on we will use the normalized Hamiltonian (\ref{becham}).

384: 

385: At this point it is important to realize that the ``electromagnetic''

386: interaction here should be self-induced, since we are dealing

387: with neutral condensates. So we identify

388: the ``electromagnetic'' potential by the velocity field of

389: the doublet $\xi$,

390: \bea

391: \tilde A_\mu = -\dfrac {i}{g} \xi^\dag \pro_\mu \xi.

392: \label{am}

393: \eea

394: A justification for this is that the $U(1)$ gauge invariance almost

395: dictates us to identify the velocity field as the gauge potential.

396: Indeed in the absence of the Maxwell term, (\ref{am}) becomes 

397: an equation of motion. With this identification 

398: the field strength becomes the vorticity field

399: \bea

400: &\tilde F_{\mu\nu} = \pro_\mu \tilde A_\nu - \pro_\nu \tilde A_\mu \nn\\

401: &= -\dfrac {i}{g} (\pro_\mu \xi^\dag \pro_\nu \xi

402: - \pro_\nu \xi^\dag \pro_\mu \xi).

403: \label{fmn}

404: \eea

405: One might wonder why we have included 

406: the vorticity interaction in the Lagrangian (\ref{beclag}), 

407: when this is absent in the Gross-Pitaevskii Lagrangian. 

408: The reason is because it costs energy to create

409: the vorticity in two-component BEC. In ordinary BEC one does

410: not have to worry about the vorticity because the vorticity is identically

411: zero (since the velocity is given by the gradient of the phase

412: of the one-component complex condensate). But a non-Abelian

413: (multi-component) BEC has a non-vanishing vorticity,

414: in which case it is natural to include

415: the vorticity interaction in the Lagrangian \cite{bec1,bec5}. 

416: 

417: With (\ref{am}) the Hamiltonian (\ref{becham}) becomes

418: \bea

419: &{\cal H} = \dfrac {1}{2} (\pro_i \rho)^2 + \dfrac {\rho^2}{2}\Big(|\pro_i

420: \xi |^2 - |\xi^\dag \pro_i \xi|^2 \Big) 

421: - \dfrac {\lambda}{8} (\rho^2 - \rho^2_0)^2 \nn\\

422: &- \dfrac {1}{4 g^2} (\pro_i \xi^\dag \pro_j \xi 

423: - \pro_j \xi^\dag \pro_i \xi)^2, \nn\\

424: &\rho^2_0= \dfrac{2\mu^2}{\lambda}.

425: \label{becham1}

426: \eea

427: Notice that now the coupling constant $g$ represents

428: the coupling strength of the vorticity.

429: Minimizing the Hamiltonian we have

430: the following equation of motion

431: \bea

432: & \pro^2 \rho - \Big(|\pro_i \xi |^2 - |\xi^\dag \pro_i \xi|^2 \Big)

433: \rho = \dfrac {\lambda}{2} (\rho^2 - \rho_0^2) \rho,\nn \\

434: &\Big\{(\pro^2 - \xi^\dag \pro^2 \xi) + 2 (\dfrac {\pro_i

435: \rho}{\rho} - \xi^\dag \pro_i\xi)(\pro_i - \xi^\dag \pro_i \xi) \nn\\

436: &- \dfrac {1}{g^2 \rho^2} \Big[\pro_i (\pro_i

437: \xi^\dag \pro_j \xi - \pro_j \xi^\dag \pro_i \xi) \Big] 

438: (\pro_j - \xi^\dag \pro_j \xi) \Big\} \xi \nn\\

439: &= 0.

440: \label{beckeq1}

441: \eea

442: To understand the meaning of this notice that with

443: \bea

444: \n = \xi^\dag \vec \sigma \xi,

445: \label{hn}

446: \eea

447: we have

448: \bea

449: & |\pro_\mu \xi|^2 - |\xi^\dag \pro_\mu \xi|^2

450: = \dfrac{1}{4} (\pro_\mu \hn)^2, \nn\\

451: &i (\pro_\mu \xi^\dag \pro_\nu \xi 

452: - \pro_\nu \xi^\dag \pro_\mu \xi ) 

453: = \dfrac{1}{2} \hn \cdot (\pro_\mu \hn \times \pro_\nu \hn) \nn\\

454: &= \dfrac{1}{2} N_{\mu\nu},

455: \label{id}

456: \eea

457: where $N_{\mu\nu}$ is mathematically identical to what we have 

458: in Skyrme theory in (\ref{sfeq}).

459: This tells that the potential $C_\mu$ for the two-form $N_{\mu\nu}$

460: in (\ref{sfeq}) is given by (up to a gauge transformation)

461: \bea

462: C_\mu = 2g \tilde A_\mu = -2i \xi^\dag \pro_\mu \xi. 

463: \label{cm}

464: \eea

465: More importantly, with (\ref{id}) the Hamiltonian (\ref{becham1}) 

466: can be expressed as

467: \bea

468: &{\cal H} = \dfrac{1}{2} (\pro_i \rho)^2

469: + \dfrac{\rho^2}{8} (\pro_i \hn)^2 

470: + \dfrac{\lambda}{8}(\rho^2-\rho^2_0)^2 \nn\\

471: &+ \dfrac{1}{16 g^2} (\pro_i \hn \times \pro_j \hn)^2.

472: \eea

473: So the theory becomes very similar to Skyrme-Faddeev

474: theory, which strongly indicates that 

475: the gauge theory of two-component BEC can allow a knot.

476: As importantly this strongly implies that the Skyrme-Faddeev theory 

477: could play an important role in condensed matter

478: physics.

479: 

480: To simplify the equation (\ref{beckeq1}) notice that 

481: from (\ref{am}) and (\ref{hn}) we have

482: \bea

483: \pro_\mu \xi = (ig \tilde A_\mu + \dfrac{1}{2} \vec \sigma

484: \cdot \pro_\mu \n) \xi.

485: \eea

486: With the identity the second equation of (\ref{beckeq1}) is reduced to 

487: \bea

488: &(A + \B \cdot \vec \sigma) \xi = 0, \nn\\

489: &A = (\pro_i \n)^2, \nn \\

490: &\B = \pro^2 \n + 2 \dfrac{\pro_i \rho}{\rho} \pro_i \n

491: - \dfrac{i}{2 g^2 \rho^2} \pro_i N_{ij} \pro_j \n, 

492: \eea

493: which can be written as

494: \bea

495: &A+ \vec B \cdot \hn=0, \nn\\

496: & \hn \times \vec B - i \hn \times (\hn \times \vec B) =0,

497: \eea

498: or 

499: \bea

500: \n \times (\pro^2 \n + 2 \dfrac{\pro_i \rho}{\rho} \pro_i \n) 

501: + \dfrac{1}{2 g^2 \rho^2} \pro_i N_{ij} \pro_j \n = 0.

502: \eea

503: So we can put (\ref{beckeq1}) into the form

504: \bea

505: &\pro^2 \rho - \dfrac{1}{4} (\pro_i \n)^2 \rho 

506: = \dfrac{\lambda}{2}(\rho^2 - \rho_0^2)

507: \rho, \nn \\

508: &\n \times \pro^2 \n + 2 \dfrac{\pro_i \rho}{\rho} \n \times \pro_i

509: \n + \dfrac{1}{g^2\rho^2} \pro_i N_{ij} \pro_j \n = 0.

510: \label{beckeq2}

511: \eea

512: This is the equation of two-component BEC that we are 

513: looking for. The similarity between this and the equation 

514: (\ref{sfeq}) of Skyrme-Faddeev theory is unmistakable.

515: 

516: Notice that the target space of $\xi$ and $\n$ is $S^3$ and

517: $S^2$, but here we have transformed the equation for $\xi$

518: in (\ref{beckeq1}) completely into the equation

519: for $\n$ in (\ref{beckeq2}).

520: This is made possible because, with the Abelian gauge invariance of (6),

521: the physical target space of $\xi$ becomes the gauge orbit space

522: $S^2 = S^3/S^1$ which forms a $CP^1$ space.

523: This means that we can view the theory as a self interacting

524: $CP^1$ theory (coupled to a scalar field $\rho$), and replace

525: $\xi$ completely in terms of $\n$.

526: 

527: To show that the theory has a knot solution we construct 

528: a helical vortex solution first. To do this 

529: we choose the ansatz

530: \bea

531: &\rho= \rho(\varrho), 

532: ~~~~~\xi = \Bigg( \matrix{\cos \dfrac{f(\varrho)}{2}

533: \exp (-in\varphi) \cr

534: \sin \dfrac{f(\varrho)}{2} \exp (imkz)} \Bigg), \nn\\

535: &\n= \xi^\dag \vec \sigma \xi

536: = \Bigg(\matrix{\sin{f(\varrho)}\cos{(mkz+n\varphi)} \cr

537: \sin{f(\varrho)}\sin{(mkz+n\varphi)} \cr \cos{f(\varrho)}}\Bigg). \nn\\

538: \label{bhvans}

539: \eea

540: With this (\ref{beckeq2}) is reduced to

541: \bea

542: &\ddot{\rho}+\dfrac{1}{\varrho}\dot\rho

543: - \dfrac{1}{4}\Big(\dot{f}^2+(m^2 k^2+\dfrac{n^2}{\varrho^2})

544: \sin^2{f}\Big)\rho \nn\\

545: &= \dfrac{\lambda}{2}(\rho^2-\rho_0^2)\rho, \nn\\

546: &\Big(1+(m^2 k^2+\dfrac{n^2}{\varrho^2})

547: \dfrac{\sin^2{f}}{g^2 \rho^2}\Big) \ddot{f}

548: + \Big( \dfrac{1}{\varrho}+ 2\dfrac{\dot{\rho}}{\rho} \nn\\

549: &+(m^2 k^2+\dfrac{n^2}{\varrho^2})

550: \dfrac{\sin{f}\cos{f}}{g^2 \rho^2} \dot{f}

551: + \dfrac{1}{\varrho} (m^2 k^2-\dfrac{n^2}{\varrho^2})

552: \dfrac{\sin^2{f}}{g^2 \rho^2} \Big) \dot{f} \nn\\

553: &- (m^2 k^2+\dfrac{n^2}{\varrho^2}) \sin{f}\cos{f}=0.

554: \label{bveq}

555: \eea

556: So with the boundary condition

557: \bea

558: &\dot \rho(0)=0,~~~~~\rho(\infty)=\rho_0, \nn\\

559: &f(0)=\pi,~~~~~f(\infty)=0,

560: \label{becbc}

561: \eea

562: we obtain the non-Abelian vortex solutions shown in Fig.\ref{becheli}.

563: 

564: \begin{figure}[t]

565: \includegraphics[scale=0.5]{becknot.eps}

566: \caption{The non-Abelian vortex (dashed line) with $m=0,n=1$ and

567: the helical vortex (solid line) with $m=n=1$ in the gauge theory

568: of two-component BEC. Here we have put $\lambda/g^2=1$,

569: $k=0.64~{\sqrt \lambda}\rho_0$, and $\varrho$ is in the unit of

570: $1/{\sqrt \lambda}\rho_0$.}

571: \label{becheli}

572: \end{figure}

573: 

574: There are three points that have to be

575: emphasized here. First, when $m=0$, the solution describes

576: the untwisted non-Abelian vortex \cite{bec1}. But when $m$ is not zero,

577: it describes a helical vortex which is periodic in $z$-coodinate.

578: In this case, the vortex has a non-vanishing velocity current

579: (not only around the vortex but also) along the $z$-axis.

580: Secondly, the doublet $\phi$ starts from

581: the second component at the core,

582: but the first component takes over completely at the infinity.

583: This is due to the boundary condition $f(0)=\pi$ and $f(\infty)=0$,

584: which assures that our solution describes a genuine

585: non-Abelian vortex. Thirdly, when $f=0$ (or $f=\pi$)

586: the doublet effectively becomes

587: a singlet, and (\ref{bveq}) describes the well-known

588: vortex in single-component BEC. Only when $f$ has a non-trivial

589: profile, we have a non-Abelian vortex.

590: 

591: In Skyrme theory the helical vortex is interpreted as

592: a twisted magnetic vortex whose flux is quantized \cite{prl01,plb04}

593: Now we show that the above vortex is a twisted vorticity flux

594: which is also quantized.

595: To see this notice that the non-Abelian structure of

596: the vortex is represented by the doublet $\xi$.

597: Moreover, the velocity field 

598: of the doublet is given by 

599: \bea

600: &\tilde A_\mu =-\dfrac{n}{2g}(\cos{f}+1) \pro_\mu \varphi \nn\\

601: &-\dfrac{mk}{2g}(\cos{f}-1) \pro_\mu z,

602: \label{gpvel}

603: \eea

604: which generates the vorticity

605: \bea

606: &H_{\mu\nu}= -\dfrac{i}{g}(\pro_\mu \xi^{\dagger} \pro_\nu \xi

607: -\pro_\nu \xi^{\dagger} \pro_\mu \xi) \nn\\

608: &=\dfrac{\dot{f}}{2g} \sin{f}\Big(n(\pro_\mu \varrho \pro_\nu \varphi

609: -\pro_\nu \varrho \pro_\mu \varphi) \nn\\

610: &+mk(\pro_\mu \varrho \pro_\nu z

611: - \pro_\nu \varrho \pro_\mu z) \Big).

612: \label{gpvor}

613: \eea

614: This has two vorticity fluxes, $\phi_{\hat z}$

615: along the $z$-axis

616: \bea

617: &\phi_{\hat z}=\dfrac{}{} \int H_{{\hat \varrho}{\hat \varphi}}

618: \varrho d \varrho d \varphi

619: = -\dfrac{2\pi n}{g},

620: \label{gpfluxz}

621: \eea

622: and $\phi_{\hat \varphi}$ around the the $z$-axis (in one

623: period section from $z=0$ to $z=2\pi/k$)

624: \bea

625: &\phi_{\hat \varphi}=\dfrac{}{} \int_{z=0}^{z=2\pi/k}

626: H_{{\hat z}{\hat \varrho}} d \varrho dz

627: = \dfrac{2\pi m}{g}.

628: \label{gpfluxphi}

629: \eea

630: This shows that the helical vortex is made of two quantized vorticity

631: fluxes, the $\phi_{\hat z}$-flux which is concentrated at the core

632: and the $\phi_{\hat \varphi}$-flux which surrounds it. 

633: This confirms that the helical vortex is a twisted vorticity

634: flux which is very similar to the helical vortex in Gross-Pitaevskii 

635: theory \cite{bec5}.

636: 

637: Now, with the helical vortex, one can easily make a twisted vortex ring

638: smoothly connecting the periodic ends together. 

639: And just as in Skyrme theory the twisted vortex ring

640: becomes a topological knot. But

641: here it is $\pi_3(S^3)$ of the doublet $\xi$ which provides the

642: non-trivial quantum number, 

643: \bea

644: q = - \dfrac {1}{4\pi^2} \int \epsilon_{ijk} \xi^{\dagger}

645: \partial_i \xi ( \partial_j \xi^{\dagger}

646: \partial_k \xi ) d^3 x.

647: \label{beckqn}

648: \eea

649: Of course this is identical to the expression (\ref{kqn}), due to the

650: Hopf fibering of $S^3$ to $S^2 \times S^1$ \cite{bec5}. This

651: tells that we can express the knot quantum number either

652: by $\pi_3(S^3)$ or by $\pi_3(S^2)$. 

653: 

654: Obviously this knot has a topological stability, because 

655: the knot topology (\ref{kqn}) can not be changed by a smooth 

656: deformation of the field configuration. Moreover it has 

657: a dynamical stability. To understand this notice that

658: the knot has a twisted velocity field so that it has a

659: non-vanishing velocity around the $z$-axis.

660: This means that it carries a non-vanishing angular momentum along 

661: the $z$-axis. And this angular momentum provides 

662: the dynamical stability, because it creates a centrifugal force that 

663: prevents the collapse of the knot. Notice that this dynamical 

664: stability originates from the knot topology, because 

665: the angular momentum comes from the twisted velocity field. 

666: In this sense the topological stability and the dynamical stability

667: have one and the same origin. It is this remarkable 

668: interplay between topology and dynamics which assures 

669: the stability of the knot. The nontrivial twisted topology of 

670: the knot expresses itself in the form of the angular momentum, 

671: which in turn provides the dynamical stability of the knot. 

672: This presence of the angular momnetum is 

673: what differentiates our

674: knot from the untwisted Abrikosov-type vortex ring 

675: which has no dynamical stability.  

676: 

677: There have been assertions that two-component BEC admits

678: a knot \cite{ruo}. But notice that this knot is based on 

679: Gross-Pitaevskii theory of two-component BEC, which has no

680: vorticity interaction. In contrast our knot is based on 

681: the gauge theory in which the vorticity 

682: interaction plays a crucial role. Nevertheless physically two knots 

683: are very similar \cite{bec5}. Both can be identified as 

684: a vorticity knot. This implies that both theories should be 

685: taken seriously as a theory of two-component BEC.

686: 

687: \section{Knot in Two-gap Superconductors}

688: 

689: In the above gauge theory of spin-1/2   

690: condensates the gauge interaction was a  

691: self-induced interaction. But when the doublet is charged,

692: the gauge interaction can be treated as independent.

693: In this case the theory can describe a two-gap superconductor.

694: But even in this case the knot topology and thus the knot itself

695: should survive. This implies that two-gap 

696: superconductor should also have

697: a topological knot.

698: 

699: The knot in two-gap superconductor could be either relativistic

700: or non-relativistic, and appear in both Abelian and non-Abelian

701: setting \cite{cm2}. In this paper we will discuss the relativistic 

702: knot in the Abelian setting (a non-relativistic

703: Gross-Pitaevskii type theory gives an identical result).

704: Consider a charged doublet scalar field $\phi$ coupled to

705: the real electromagnetic field,

706: \bea

707: &{\cal L} = - |D_\mu \phi|^2 + \mu^2 \phi^{\dagger}\phi 

708: - \dfrac{\lambda}{2} (\phi^{\dagger} \phi)^2

709: - \dfrac{1}{4} F_{\mu \nu}^2 , \nn\\

710: &D_\mu \phi = (\partial_\mu - ig A_\mu) \phi.

711: \label{sclag}

712: \eea

713: The Lagrangian has the equation of motion

714: \bea

715: &D^2\phi =\lambda(\phi^{\dagger} \phi

716: -\dfrac{\mu^2}{\lambda})\phi, \nn\\

717: &\partial_\mu F_{\mu \nu} = j_\nu =  i g \Big[(D_\nu

718: \phi)^{\dagger}\phi - \phi ^{\dagger}(D_\nu \phi) \Big].

719: \label{sceq1}

720: \eea

721: Now, with

722: \bea

723: \phi =\dfrac{1}{\sqrt 2} \rho \xi,~~~~~{\xi}^{\dagger}\xi = 1,

724: ~~~~~\hat n = \xi^{\dagger} \vec \sigma \xi,

725: \eea

726: we can reduce (\ref{sceq1}) to \cite{cm2}

727: \bea

728: &\partial ^2 \rho - \Big(

729: \dfrac{1}{4} (\partial _\mu \hat n)^2+ g^2

730: (A_\mu + \tilde A_\mu)^2 \Big) \rho

731: = \dfrac{\lambda}{2} (\rho^2-\rho_0^2)\rho, \nn\\

732: &\hat n \times \partial ^2 \hat n + 2 \dfrac{\partial_\mu \rho}{

733: \rho} \hat n \times \partial_\mu \hat n

734: + \dfrac{2}{g\rho^2} \partial_\mu F_{\mu\nu} \partial_\nu \hat n =0, \nn\\

735: &\partial_\mu F_{\mu\nu} = j_\mu

736: =g^2 \rho^2 (A_\mu + \tilde A_\mu), \nn\\

737: &\tilde A_\mu=-\dfrac{i}{g} \xi^{\dagger}\partial_\mu \xi,

738: ~~~~~\rho_0=\dfrac{2\mu^2}{\lambda}.

739: \label{sceq2}

740: \eea

741: This is the equation for two-gap superconductor.

742: Notice that with $A_\mu=-\tilde A_\mu$ the first two equations

743: reduce to (\ref{beckeq2}). This tells that the gauge theory of 

744: two-component BEC and the above theory of two-gap superconductor

745: are closely related.

746: 

747: To obtain the desired knot we first construct a superconducting

748: helical magnetic vortex. Let 

749: \bea

750: &\rho=\rho(\varrho), 

751: ~~~~~\xi = \Bigg( \matrix{\cos \dfrac{f(\varrho)}{2} \exp (-in\varphi) \cr

752: \sin \dfrac{f(\varrho)}{2} \exp (imkz)} \Bigg), \nn\\

753: &A_\mu= \dfrac{1}{g} \big(n A_1(\varrho) \partial_\mu\varphi

754: + mk A_2(\varrho) \partial_\mu z \big), \nn\\

755: &\n= \xi^\dag \vec \sigma \xi

756: = \Bigg(\matrix{\sin{f(\varrho)}\cos{(n\varphi+mkz)} \cr

757: \sin{f(\varrho)}\sin{(n\varphi+mkz)} \cr \cos{f(\varrho)}}\Bigg), \nn\\

758: &\tilde A_\mu=-\dfrac{n}{2g} \big(\cos{f(\varrho)}+1\big)

759: \partial_\mu \varphi \nn\\

760: &- \dfrac{mk}{2g} \big(\cos{f(\varrho)}-1\big) \partial_\mu z.

761: \label{scans}

762: \eea

763: With this we have

764: \bea

765: &j_\mu = g\rho^2 \Big(n \big(A_1-\dfrac{\cos{f}+1}{2}\big)

766: \partial_\mu \varphi \nn\\

767: &+ mk\big(A_2-\dfrac{\cos{f}-1}{2}\big)

768: \partial_\mu z \Big),

769: \label{sc}

770: \eea

771: and (\ref{sceq2}) becomes

772: \bea

773: &\ddot{\rho}+\dfrac{1}{\varrho}\dot\rho

774: - \Big[\dfrac{1}{4} \Big(\dot{f}^2

775: +\big(\dfrac{n^2}{\varrho^2} + m^2 k^2 \big) \sin^2{f}\Big) \nn\\

776: &+\dfrac{n^2}{\varrho^2} \Big(A_1-\dfrac{\cos{f}+1}{2}\Big)^2

777: + m^2 k^2 \Big(A_2-\dfrac{\cos{f}-1}{2}\Big)^2 \Big]\rho \nn\\

778: &= \dfrac{\lambda}{2}(\rho^2-\rho_0^2)\rho, \nn\\

779: &\ddot{f} + \big(\dfrac{1}{\varrho}

780: +2\dfrac{\dot{\rho}}{\rho} \big)\dot{f}

781: -2 \Big(\dfrac{n^2}{\varrho^2}\big(A_1-\dfrac{1}{2}\big) \nn\\

782: &+ m^2 k^2 \big(A_2+\dfrac{1}{2}\big) \Big)\sin{f} = 0, \nn\\

783: &\ddot{A}_1-\dfrac{1}{\varrho}\dot{A}_1 -g^2 \rho^2

784: \Big(A_1-\dfrac{\cos{f}+1}{2}\Big) = 0, \nn\\

785: &\ddot{A}_2+\dfrac{1}{\varrho}\dot{A}_2 -g^2 \rho^2

786: \Big(A_2-\dfrac{\cos{f}-1}{2}\Big) = 0.

787: \label{sceq3}

788: \eea

789: Now, we impose the following boundary condition for the non-Abelian

790: vortices \cite{cm2},

791: \bea

792: &\rho (0) = 0,~~~\rho(\infty) = \rho_0, 

793: ~~~f (0) = \pi,~~~f (\infty) = 0, \nn\\

794: & A_1 (0) = -1,~~~A_1 (\infty) =1.

795: \label{scbc}

796: \eea

797: This need some explanation, because the boundary

798: value $A_1(0)$ is chosen to be $-1$, not $0$. This is to

799: assure the smoothness of $\rho(\varrho)$ and $f(\varrho)$

800: at the origin. Only with this boundary value they

801: become analytic at the origin.

802: At this point one might object the boundary condition, because it creates

803: an apparent singularity in the gauge potential at the origin.

804: But notice that this singularity is an unphysical

805: (coordinate) singularity which can easily be removed by a gauge transformation.

806: In fact the singularity disappears with the gauge transformation

807: \bea

808: \phi \rightarrow \phi \exp(in\varphi),

809: ~~~~A_\mu \rightarrow A_\mu + \dfrac{n}{g} \partial_\mu \varphi,

810: \eea

811: which changes the boundary condition $A_1(0)=-1$ and 

812: $A_1(\infty)=1$ to $A_1(0)=0$ and $A_1(\infty) =2$.

813: Mathematically this boundary condition

814: has a deep origin, which has to do with the fact

815: that the Abelian $U(1)$ runs from $0$ to $2\pi$, but the $S^1$

816: fiber of $SU(2)$ runs from $0$ to $4\pi$ \cite{cm2}.

817: As for $A_2(\varrho)$, we choose $A_2(\infty)=0$ 

818: to make the supercurrent vanishing at infinity

819: and require the vortex superconducting. As we will see,

820: this requires a logarithmic divergence for $A_2(0)$.

821: The boundary condition will

822: have an important consequence in the following.

823: 

824: \begin{figure}

825: \includegraphics[scale=0.7]{nascknot.eps}

826: \caption{The non-Abelian vortex (dashed line) with

827: $m=0,n=1$ and the helical vortex (solid line) with $m=n=1$ in

828: two-gap superconductor. Here we have put $g=1,~\lambda=2$,

829: $k=\rho_0/10$, and $\varrho$ is in the unit of $1/\rho_0$.

830: Notice that $A_2$ has a logarithmic singularity at the origin.}

831: \label{heli-sc}

832: \end{figure}

833: 

834: With the boundary condition we can integrate (\ref{sceq3})

835: and obtain the non-Abelian vortex solution

836: of the two-gap superconductor, which is shown in Fig.\ref{heli-sc}.

837: The solution is very similar to the one we have in 

838: two-component BEC. When $m=0$, the solution (with $A_2=0$) 

839: describes an untwisted non-Abelian vortex \cite{cm2}. 

840: But when $m$ is not zero, it describes a helical magnetic vortex 

841: which is periodic in $z$-coordinate.

842: Moreover, the vortex starts from

843: the second component at the core,

844: but the first component takes over completely at the infinity.

845: This is due to the boundary condition $f(0)=\pi$ and $f(\infty)=0$,

846: which assures that our solution describes a genuine

847: non-Abelian vortex. This is true even when $m=0$. Only when 

848: $f=0$ (or $f=\pi$) the doublet effectively becomes

849: a singlet, and (\ref{sceq3}) describes the Abelian

850: Abrikosov vortex of one-gap superconductor.

851: 

852: There are important differences between the non-Abelian vortex 

853: and the Abrikosov vortex. First the non-Abelian vortex

854: has a non-Abelian magnetic flux quantization \cite{cm2}. 

855: Indeed the quantized magnetic flux $\hat \phi_z$ 

856: of the non-Abelian vortex along the $z$-axis is given by 

857: \bea

858: &H_z= \dfrac{n}{g} \dfrac{\dot A_1}{\varrho}, \nn\\

859: &\hat \phi_z = \dfrac{}{} \int H_z d^2x

860: = \dfrac{2\pi n}{g} \big[A_1(\infty) - A_1(0)\big] \nn\\

861: &= \dfrac{4\pi n}{g}.

862: \label{zflux}

863: \eea

864: Notice that the unit of the non-Abelian flux is $4\pi/g$,

865: not $2\pi/g$. This is a direct

866: consequence of the boundary condition (\ref{scbc}).

867: This non-Abelian quantization of magnetic flux comes

868: from the non-Abelian topology $\pi_2(S^2)$ of the doublet $\xi$,

869: or equivalently the triplet $\hn$, whose topological

870: quantum number is given by \cite{cm2}

871: \bea

872: &q = \dfrac {g}{4\pi } \int H_z d^2 x

873: = - \dfrac {1}{4\pi} \int \epsilon_{ij} \partial_i \xi^{\dagger}

874: \partial_j \xi d^2 x \nn\\

875: &= \dfrac {1}{8\pi} \int \epsilon_{ij} \hn \cdot (\partial_i \hn

876: \times \partial_j \hn) d^2 x = n.

877: \label{scqn}

878: \eea

879: This distinguishes our non-Abelian vortex from

880: the Abelian vortex whose topology is fixed by $\pi_1(S^1)$.

881: 

882: Another important feature of the non-Abelian vortex 

883: is that it carries a non-vanishing supercurrent

884: along the $z$-axis,

885: \bea

886: &i_z = mkg \dfrac{}{} \int \rho^2 \big(A_2

887: -\dfrac{\cos f +1}{2}\big)\varrho d\varrho d\varphi \nn\\

888: &= \dfrac{2\pi mk}{g} \int \big(\ddot A_2

889: +\dfrac{1}{\varrho} \dot A_2 \big) \varrho d\varrho \nn\\

890: &= \dfrac{2\pi mk}{g} (\varrho \dot A_2) \Big|_{\varrho=0}^{\varrho=\infty}

891: =-\dfrac{2\pi mk}{g} (\varrho \dot A_2) \Big|_{\varrho=0}.

892: \label{scz}

893: \eea

894: This is due to the logarithmic divergence of $A_2$ at the origin.

895: 

896: Notice that the superconducting helical vortex

897: has only a heuristic value, because one needs

898: an infinite energy to create it (since the magnetic flux

899: around the vortex becomes divergent because of the singularity

900: of $A_2$ at the origin).

901: With the helical vortex, however, one can make

902: a vortex ring by smoothly bending and connecting

903: two periodic ends. In the vortex ring the infinite magnetic

904: flux of $A_2$ can be made finite making the finite supercurrent

905: (\ref{scz}) of the vortex ring produce a finite flux,

906: and we can fix the flux to have the value $4\pi m/g$

907: by adjusting the current with

908: $k$. With this the vortex ring now

909: becomes a topologically stable knot.

910: 

911: To see this notice that the doublet $\xi$, after forming a knot,

912: acquires a non-trivial topology $\pi_3(S^2)$ which provides

913: the knot quantum number,

914: \bea

915: &Q = - \dfrac {1}{4\pi^2} \int \epsilon_{ijk} \xi^{\dagger}

916: \partial_i \xi ( \partial_j \xi^{\dagger}

917: \partial_k \xi ) d^3 x \nn\\

918: &= \dfrac{g^2}{32\pi^2} \int \epsilon_{ijk} C_i

919: (\partial_j C_k - \partial_k C_j) d^3x

920: =mn.

921: \label{bkqn}

922: \eea

923: This is nothing but the Chern-Simon index of the potential

924: $C_\mu$, which is mathematically identical to

925: the quantum number of the knots we discussed before. 

926: This tells that our

927: knot is also made of two quantized magnetic flux rings

928: linked together whose knot quantum number is fixed by

929: the linking number $mn$. Obviously two flux rings linked together

930: can not be separated by any continuous deformation of

931: the field configuration. This provides the topological stability

932: of the knot.

933: 

934: Again this topological stability is backed up

935: by a dynamical stability. To see this notice that the supercurrent

936: of the knot has two components, the one around

937: the knot tube which confines the magnetic flux along the knot,

938: but more importantly the other along the knot

939: which creates a magnetic flux

940: passing through the knot disk. This component of supercurrent

941: along the knot now generates a net

942: angular momentum which provides

943: the centrifugal repulsive force preventing the knot to collapse.

944: This makes the knot dynamically stable.

945: 

946: To compare our knot with the Abrikosov vortex ring (made of the

947: Abrikosov vortex in conventional superconductor),

948: notice that the Abrikosov knot is empty (i.e.,

949: does not carry a net supercurrent).

950: As importantly it is unstable, and collapses immediately.

951: In contrast our knot has a helical supercurrent, and is stable. 

952: Furthermore these two features are deeply related. 

953: The helical supercurrent plays a crucial role to stablize the vortex

954: ring by providing the net angular momentum, which prevents the

955: collapse of the vortex ring. And this helical supercurrent 

956: originates from the knot topology. This remarkable

957: interplay between topology and dynamics is what provides

958: the stability of the knot. The nontrivial topology expresses 

959: itself in the form of the helical supercurrent, which in turn

960: provides the dynamical stability of the knot. We emphasize that

961: this supercurrent is what distinguishes our knot from

962: the Abrikosov vortex ring, which has neither topological

963: nor dynamical stability.

964: 

965: \section{Discussion}

966: 

967: \begin{figure}[t]

968: \includegraphics[scale=0.7]{2schv.eps}

969: \caption{The regular helical magnetic vortex with $m=n=1$ in

970: two-gap superconductor. Here we have put $g=1,~\lambda=2$,

971: $k=0.12 \rho_0$, $A_2(0)=0.5$, and $\varrho$ is in the unit of $1/\rho_0$.

972: Notice that the solution is completely regular.}

973: \label{2schv}

974: \end{figure}

975: 

976: In this paper we have presented a compelling argument

977: for the existence of topological

978: knots in two-component BEC and two-gap superconductor. 

979: Similar knots have popped out almost everywhere,

980: in particular in high energy physics in QCD \cite{plb05} and 

981: Weinberg and Salam model \cite{ewknot}. 

982: But we emphasize that at the center of these

983: topological objects lies the baby skyrmion and the Faddeev-Niemi

984: knot \cite{prl01,plb04}. In fact, our helical vortices 

985: and knots in this paper are a straightforward

986: generalization of the baby skyrmion and the Faddeev-Niemi

987: knot. 

988: 

989: It has been assumed that the topological

990: objects in Skyrme theory can only be realized at high energy,

991: at the hadronic scale. But our analysis shows that

992: similar objects could exist in a completely different environment,

993: at a much lower scale, in low energy condensed matters.

994: If so, the challenge now is

995: to verify the existence of the topological

996: knot experimentally in condensed matters. Constructing the knot

997: might not be so easy at present moment. 

998: Nevertheless, with some experimental ingenuity,

999: one should be able to construct the knots

1000: in condensed matters.

1001: 

1002: Note Added: One might doubt the existence of a superconducting

1003: knot because the superconducting helical vortex we discussed 

1004: in Section IV was singular (and thus unphysical). In this note 

1005: we report a regular superconducting helical vortex which has a finite 

1006: magnetic flux around the axis and thus a finite energy. 

1007: The regular solution is obtained linking $A_2(0)$ with $k$. 

1008: For example for $k=0.12$ we obtain the regular solution shown 

1009: in Fig.\ref{2schv}, with $A_2(0)=0.5$. This type of regular

1010: helical vortex has vanishing supercurrent $i_z$, 

1011: but could still be called superconducting 

1012: because it has a non-trivial supercurrent density $j_z$

1013: which generates a net magnetic flux $H_\varphi$ around the vortex.

1014: The regular helical vortex strongly support the existence of

1015: a regular knot. The details will be published elsewhere.

1016:  

1017: {\bf ACKNOWLEDGEMENT}

1018: 

1019: ~~~The work is supported in part by the ABRL Program of

1020: Korea Science and Engineering Foundation (R14-2003-012-01002-0)

1021: and by the BK21 Project of the Ministry of Education.

1022: 

1023: \begin{thebibliography}{99}

1024: \bibitem{dirac}P. A. M. Dirac, Proc. Roy. Soc. {\bf A113}, 60 (1931);

1025: Phys. Rev. {\bf 74}, 817 (1948).

1026: \bibitem{abri} A. Abrikosov, Sov. Phys. JETP {\bf 5}, 1174 (1957).

1027: \bibitem{skyrme}T. H. R. Skyrme, Proc. Roy. Soc. (London) {\bf 260}, 127

1028: (1961); {\bf 262}, 237 (1961); Nucl. Phys. {\bf 31}, 556 (1962).

1029: See also, for example, I. Zahed and G. Brown,

1030: Phys. Rep. {\bf 142}, 1 (1986), and the references therein.

1031: \bibitem{thooft}G. 'tHooft, Nucl. Phys. {\bf 79}, 276 (1974);

1032: A. M. Polyakov, JETP Lett. {\bf 20}, 194 (1974).  

1033: \bibitem{fadd1} L. Faddeev and A. Niemi, Nature {\bf 387}, 58 (1997);

1034: J. Gladikowski and M. Hellmund, Phys. Rev. {\bf D56}, 5194 (1997);

1035: R. Battye and P. Sutcliffe, Phys. Rev. Lett. {\bf 81}, 4798

1036: (1998).

1037: \bibitem{prl01} Y. M. Cho, Phys. Rev. Lett. {\bf 87}, 252001 (2001).

1038: \bibitem{plb97} Y. M. Cho and D. Maison, Phys. Lett. {\bf B391},

1039: 360 (1997); W. S. Bae and Y. M. Cho, JKPS {\bf 46}, 791 (2005).

1040: \bibitem{yang} Y. Yang, {\it Solitons in Field Theory and Nonlinear

1041: Analysis} (Springer) 2001.

1042: \bibitem{bec} C. Myatt {\it at al.}, Phys. Rev. Lett. {\bf 78}, 586 (1997); 

1043: D. Stamper-Kurn, {\it at al.}, Phys. Rev. Lett. {\bf 80}, 2027 (1998);

1044: J. Stenger {\it at al.}, Nature {\bf 396}, 345 (1998).

1045: \bibitem{ruo} J. Ruostekoski and J. Anglin, Phys. Rev. Lett.

1046: {\bf 86}, 3934 (2001); U. Al Khawaja and H. Stoof, Nature {\bf 411},

1047: 818 (2001); Phys. Rev. {\bf A64}, 043612 (2001); 

1048: H. Stoof {\it at al.}, Phys. Rev. Lett. {\bf 87}, 120407 (2001); 

1049: R. Battye, N. Cooper, and P. Sutcliffe, Phys. Rev. Lett. {\bf 88}, 

1050: 080401 (2002); C. Savage and J. Ruostekoski, Phys. Rev. Lett. {\bf 91}, 

1051: 010403 (2003).

1052: \bibitem{exp1}J. Nagamatsu et al., Nature {\bf 410}, 63 (2001);

1053: S. L. Bud'ko et al., Phys. Rev. Lett. {\bf 86}, 1877 (2001);

1054: C. U. Jung et al., Appl. Phys. Lett. {\bf 78}, 4157(2001).

1055: \bibitem{exp2}H. D. Yang et al., Phys. Rev. Lett. {\bf 87}, 167003 (2001);

1056: J. J. Tu et al., Phys. Rev. Lett. {\bf 87}, 277001 (2001).

1057: \bibitem{piet} B. Piette, B. Schroers, and W. Zakrzewski,

1058: Nucl. Phys. {\bf 439}, 205 (1995).

1059: \bibitem{plb04} Y. M. Cho, Phys. Lett. {\bf B603}, 88 (2004);

1060: hep-th/0404181.

1061: \bibitem{bec1} Y. M. Cho, cond-mat/0112325.

1062: \bibitem{bec5} Y. M. Cho, cond-mat/0409636.

1063: \bibitem{cm2} Y. M. Cho, cond-mat/0112498; cond-mat/0308182.

1064: \bibitem{plb05}Y. M. Cho, Phys. Lett. {\bf B616}, 101 (2005).

1065: \bibitem{ewknot} Y. M. Cho, hep-th/0110076.

1066: \end{thebibliography}

1067: 

1068: \end{document}

1069: