Published on

16-Apr-2017View

213Download

1

Transcript

Fortschr. Phys. 62, No. 3, 266 273 (2014) / DOI 10.1002/prop.201300034

p-wave superconductors in dilaton gravity

ZhongYing Fan

Department of Physics, Beijing Normal University, Beijing 100875, China

Received 20 January 2014, accepted 31 January 2014Published online 19 February 2014

Key words AdS/CFT correspondence, gauge/gravity duality, holographic p-wave superconductors

In this paper, we study peculiar properties of p-wave superconductors in dilaton gravity. The scale invarianceof the bulk geometry is effectively broken due to the existence of a dilaton. By coupling the dilaton to thenon-Abelian gauge field, i.e., 1

4eF aF

a , we find that the dissipative conductivity of the normalphase decreases and approaches zero at the zero frequency as increases. Intuitively, the system behavesmore and more like an insulator. When the hairy solution is turned on, the system crosses a critical point tothe superconducting phase. We find that the critical chemical potential decreases with the increasing of and the maximum height of the conductivity is suppressed gradually which are consistent with our intuitionfor insulator/supercondutor transition.

c 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1 Introduction

In recent years, AdS/CFT correspondence has been widely used to investigate strongly correlated cond-ensed-matter physics. One of remarkable achievements is holographic model of superconductors [14].The superconductors described in the weakly coupled gravitational background are dual to those stronglycoupled superconductors in the boundary theory. It is an excellent example to show the power of gauge /gravity duality since the strongly coupled superconductors still lacks an effective description in condensed-matter physics. In contrast, the AdS/CFT correspondence provides an elegant and systematic procedure tostudy the peculiar properties of superconductors. It is also possible to develop a deep understanding onvarious mysteries such as the paring mechanism and enhanced critical temperature of superconductors.

Since the boundary theory is in general not scale invariant below some dynamical scale, it is moreimportant to study non-relativistic holography with scale invariance broken in the bulk. In this paper, weintroduce a dilaton field to effectively break the scale invariance of the bulk geometry [58]. We then moveon to discuss p-wave superconductors using SU(2) non-Abelian gauge field [4]. By rescaling the gaugecoupling constant as 1/gF e/gF , we couple the dilaton to the gauge field. The parameter ischosen to be positive. As increases, we find that the real part of the conductivity monotonically decreases.The DC conductivity defined at the zero frequency approaches zero and vanishes when exceeds somecritical value. This indicates the characteristics of insulators. With the increasing of , the system behavesmore and more like an insulator in the normal phase. When the chemical potential crosses a critical point,the system begins to superconduct. The order parameter defined by A1x condenses, leading to a hair nearthe horizon of the black hole and a gap is opened in the real part of the conductivity. Moreover, thereexists a Drude-like structure at some frequency close to the zero frequency point in the conductivitiesalong the x-direction. This is similar to the result published in [4]. When increases, we find that thecritical chemical potential in general decreases and the maximum height of the dissipative conductivity issuppressed, consistent with our intuition for insulators.

The remainder of this paper is organized as follows: In Sect. 2, we briefly introduce the gravity solutionin Einstein-dilaton model. In Sect. 3, we present the holographic model of p-wave superconductors coupledto the dilaton and derive the equations of motion (Eoms). We numerically find that the order parameter

E-mail: zhyingfan@gmail.com

c 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Fortschr. Phys. 62, No. 3 (2014) 267

condenses when the chemical potential crosses a critical value. In Sect. 4, we investigate the fluctuationmodes of Yang-Mills theory. In the end, we present a short conclusion in Sect. 5.

2 Gravity solution

Let us consider the standard Einstein-dilaton action:

SGrav =1

22

d4x

g [R ()2 V ()], (1)

where 2 is the Newton constant, R is the Ricci scalar, is a real scalar field (the dilaton) and the AdSradius has been set to 1. In order to obtain the dilaton black hole solutions, the potential V () has to bechosen appropriately. It reads [5]:

V () =6

i=1

Viei, (2)

where Vi and i are constants which are given by

r3sVi = 31 2 +

, 121 22 4 , (1 + )

32 (2 4)r3s2( 2) , 3

1 + 2 , (3)

4(2 1) 32 (2 4)r3s2(2 4) , ( 1)

32 (2 4)r3s2( + 2)

, (4)

2 12

i = + 1, 1, 1, 1 , 1, 1 (5)

where the sign corresponds to 0 and < 0 respectively. The general static black hole solutionswith asymptotical AdS4 geometry have the following form:

ds2 = W (r)(f(r)dt2 + dr2

f(r)+ dx2 + dy2), = (r). (6)

The functions W (r), f(r) and (r) are given by

W (r) =2(1 + r)1

[(1 + r) 1]2 , (r) =

2 12

log (1 + r), (7)

f(r) = 1 +3r3s

{2

4 2 + (1 + r)2

[1 (1 + r)

2 + (1 + r)

2 ]}

. (8)

where is a constant, satisfying 1, rs is also a constant. Since the potential of the dilaton is symmetricunder [5]. From now on, we focus on 0.

The black hole horizon is defined by f(rh) = 0 and rh is determined by rs and . A special case is = 1. The dilaton vanishes and the dilaton black hole with broken scale invariance is reduced to thestandard scale invariant Schwarschilld black hole with W (r) = 1/r2, f(r) = 1 (r/rs)3. In this caserh = rs. Therefore, the scale invariance broken of the bulk geometry can be measured by the difference 11. In the next sections, we will set rs = 1 for convenience.

1 In fact, this difference 1 is proportional to the hyperscaling violation in the zero temperature limit, which directlymeasures how strongly the scale invariance is broken.

www.fp-journal.org c 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

268 Z.Y. Fan: p-wave superconductors in dilaton gravity

3 P-wave superconductors coupled to a dilaton

In order to investigate p-wave superconductors in holography, we start from SU(2) action

SHigg = 14g2F

d4x

geF aF a . (9)

where F a = AaAa+abcAaAa , Aa is the gauge potential, abc is the structure constant of the Lie

group. The dilaton has been coupled to the gauge field and rescales the effective gauge coupling constant.In the following, we will not consider the backreaction effect of the gauge field on the Einstein-dilatonsector by taking the probe limit 2 0. The equation of motion is obtained by variation of the action withrespect to the gauge potential

1ge (geF a) + abcAbF c = 0. (10)

To discuss the p-wave superconductors, we consider the following ansatz

A = (r)3dt+ (r)1dx. (11)

It is straightforward to derive equation of motions for (r) and (r)

2

f = 0, (12)

+(f

f

) +

2

f2 = 0. (13)

where prime denotes d/dr. Recall that the U(1) subgroup of SU(2) generated by 3 is specified to beelectromagnetic. The function (r) is interpreted as the time component of the U(1) field while (r) playsthe role of order parameters. It is more convenient to work by rescaling the radial coordinate as r r/rh.The above equations of motion remain unchanged with the fields rescaled as rh, rh. Noticethat the location of the horizon has been scaled to unity.

To solve above equations of motion, we need to impose proper boundary conditions at the horizon

(r) = 1(r 1) + ... (14)

(r) = 0 + 1(r 1) + ... (15)In the asymptotic limit, the fields are expanded as

(r) = r + ... (16)

(r) = (0) + (1)r + ... (17)

where dots ... denote higher order terms. The physical quantities such as chemical potential , chargedensity and the condensate of the order parameter O (defined by O = (1)) of the dual field theorycan be directly read off from these asymptotic behaviors. In order to obtain a stable theory with freesource, the non-normalizable mode of (r) should be set to zero, i.e., (0) = 0. This condition is realizedby choosing 1 as the shooting parameter in numerics2

2 The initial conditions imposed at the horizon have two independent parameters (1, 0). When integrating out to the bound-ary, they are mapped to the physical quantities (, , (0), (1)). In order to set the source term (0) = 0, the value of theparameter 1 should be properly chosen to satisfy this condition. This is the so-called shooting method in literatures.

c 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.fp-journal.org

Fortschr. Phys. 62, No. 3 (2014) 269

1.0 1.5 2.0 2.50

2

4

6

8

c

O

c

Fig. 1 The vector condensate of order pa-rameters for = 1.4. = 0 (blue), =1.5 (green), = 3 (orange), = 4 (red), = 6 (purple).

From Fig. 1, we find that the order parameter condenses when the chemical potential exceeds somecritical value for various value of parameter . As increases, the vector condensate measured in units ofcritical chemical potential decreases. Certainly, the critical chemical potential varies with . We find thatc = 8.9511, for = 0, c = 8.1291, for = 1.5, c = 7.4912, for = 3, c = 6.9825, for = 4and c = 3.3821, for = 6, respectively. In Fig. 2, we plot the critical chemical potential as a functionof . In general, c decreases as increases. One can expect that when is sufficiently large, the criticalchemical potential will approach to a small value, implying that the system behaves like an insulator.

In Fig. 3, we plot the free energy difference between the normal and superconducting phase for variousvalue of . The free energy difference is defined by F = TSE/V2, where T is temperature, V2 is thevolume factor of the boundary, SE is the Euclidean action of the Higgs sector Eq. (9). We immediatelysee that for each value of , the free energy of the system decreases continously as a function of chemicalpotential. This is a strong indication that the phase transition is of second order. Crossing the transitionpoint, the superconducting phase has lower free energy than the normal phase, signifying that the newphase is thermodynamically favored.

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.57.0

7.5

8.0

8.5

9.0

9.5

c

Fig. 2 The critical chemical potential as afunction of for = 1.4.

www.fp-journal.org c 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

270 Z.Y. Fan: p-wave superconductors in dilaton gravity

1.0 1.5 2.0 2.5

800

600

400

200

0

c

F

Fig. 3 The free energy difference be-tween normal and superconducting phasefor = 1.4. = 0 (blue), = 1.5(green), = 3 (orange), = 4 (red).

4 Conductivity

4.1 Eoms and boundary conditions

In order to analyse the electromagnetic perturbations of the U(1) subgroup of SU(2) generated by 3, it issufficient to obtain consistent linearized equations by considering fluctuation modes as follows

a = eit[a1t (r)1dt+ a2t (r)

2dt+ a3x(r)3dx+ a3y(r)

3dy]. (18)

From linearized Yang-Mills equations, we derive four second order differential equations

a3y+

(f

f

)a3y

+ (

2

f2

2

f)a3y = 0, (19)

a3x+

(f

f

)a3x

+

1f2

(2a3x a1t ia2t ) = 0, (20)

a1t a1t +

fa3x = 0, (21)

a2t a2t

f(a2t + ia

3x) = 0, (22)

and two first order constraints

ia1t+ a2t

a2t = 0, (23)

ia2t + a1t a1t + f(a3x a3x) = 0. (24)Notice that a3y decouples from the other modes. Since it behaves identical to the fluctuation mode of s-wavesuperconductors except the slightly difference of the last term, the conductivity along the y-direction ex-hibits similar properties of s-wave case, which is less interesting in this paper. We will focus on discussingthe coupled modes {a3x, a1t , a2t} and the resulting x-direction conductivity xx. To extract retarded corre-lator, we need to impose infalling condition at the horizon for the coupled modes

a3x = (r 1)i/[1 + a3(1)x (r 1) + a3(2)x (r 1)2 + ...], (25)

a1t = (r 1)i/[a1(1)t (r 1) + a1(2)t (r 1)2 + ...], (26)

c 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.fp-journal.org

Fortschr. Phys. 62, No. 3 (2014) 271

a2t = (r 1)i/[a2(1)t (r 1) + a2(2)t (r 1)2 + ...], (27)where, = |f (1)|. Integrating out to the boundary, the modes are expanded as

a3x = A3(0)x +A

3(1)x r + ... (28)

a1t = A1(0)t +A

1(1)t r + ... (29)

a2t = A2(0)t +A

2(1)t r + ... (30)

Since the conductivity xx is a physical quantity at the boundary, we need to do an infinitesimal SU(2)gauge transformation to construct a gauge invariant potential a3x from the coupled modes {a3x, a1t , a2t}.The details were presented in [4]. Here, we write down the final result directly

a3x = a3x +

ia2t + a1t

2 2 . (31)

Near the boundary, it behaves as

a3x = A3(0)x +

[A3(1)x +

iA2(0)t + A

1(0)t

2 2 (1)

]r + ... (32)

The conductivity xx can be directly read off from above expansion coefficients as follows

xx =1

iA3(0)x

[A3(1)x +

iA2(0)t + A

1(0)t

2 2 (1)

]. (33)

4.2 Conductivity in the normal phase

In the normal phase, 0, the fluctuation mode a3x decouples from a1t and a2t and behave identical to a3y ,resulting to xx = yy . We will drop the subscript of the conductivity in this subsection. The decoupledmode satisfies

a3x+

(f

f

)a3x

+

2

f2a3x = 0 (34)

By imposing infalling boundary condition at the horizon, the conductivity can be read off from the farfield expansion coefficients as well as the broken phase. In Fig. 4, we find that Im() = 0 at the zero fre-quency for values of . The real part of the conductivityRe() decreases as increases for all frequency.Particularly, the DC conductivity defined by Re(0) approaches zero with the increasing of . When issufficiently large, the DC conductivity becomes close enough to the zero point3, implying that the systembehaves effectively like an insulator. Intuitively, we expect that the behavior of the system will be gettingcloser to an insulator in the increasing process of .

4.3 Conductivity of the superconducting phase

We are ready to numerically solve the coupled equations of motion Eqs. (2022) with proper boundaryconditions at the horizon. In Fig. 5, we present the numerical results of xx for various value of . Forsmall , there are some common features in the real and imaginary part of conductivities. For example,when = 0, there exists a peak4 at /T = 18 in the dissipative conductivity which precisely corresponds

3 The DC conductivity can also be analytically expressed as Re(0) = e(rh)/g2F , as shown in [7].4 It should be emphasized that this peak actually contains two Drude structures i.e. one on the left of the peak and the other on

the right side. Correspondingly, the imaginary part has two wave packets in the dual frequency region.

www.fp-journal.org c 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

272 Z.Y. Fan: p-wave superconductors in dilaton gravity

0 2 4 6 8 100.0

0.2

0.4

0.6

0.8

1.0

T

Re

0.0 0.2 0.4 0.6 0.8 1.00.06

0.05

0.04

0.03

0.02

0.01

0.00

T

Im

Fig. 4 The conductivity of the normal phase for various parameter . The left plot is the real part while the right plotis the imaginary part. = 0 (black), = 1 (purple), = 3 (blue), = 5 (green), = 7 (orange), = 9 (red).

0 20 40 60 80 1000.0

0.5

1.0

1.5

2.0

2.5

3.0

T

Re

xx

0...