Viscosity in Strongly Interacting Quantum Field Theories from Black Hole Physics

arXiv:hep-th/0405231v2 17 Mar 2005

Institute for Theoretical Physics, University of California, Santa Barbara, CA 93106, USA for Nuclear Theory, University of Washington, Seattle, WA 98195-1550, USA 3 Perimeter Institute for Theoretical Physics, Waterloo, Ontario N2L 2Y5, Canada

December 2004

Abstract The ratio of shear viscosity to volume density of entropy can be used to characterize how close a given uid is to being perfect. Using string theory methods, we show that this ratio is equal to a universal value of /4kB for a large class of strongly interacting quantum eld theories whose dual description involves black holes in antide Sitter space. We provide evidence that this value may serve as a lower bound for a wide class of systems, thus suggesting that black hole horizons are dual to the most ideal uids. PACS numbers: 11.10.Wx, 04.70.Dy, 11.25.Tq, 47.75.+f

email: [email protected] email: [email protected] email: [email protected]

Introduction.It has been known since the discovery of Hawking radiation [1] that black holes are endowed with thermodynamic properties such as entropy and temperature, as rst suggested by Bekenstein [2] based on the analogy between black hole physics and equilibrium thermodynamics. In higher-dimensional gravity theories there exist solutions called black branes, which are black holes with translationally invariant horizons [3]. For these solutions, thermodynamics can be extended to hydrodynamicsthe theory that describes long-wavelength deviations from thermal equilibrium [4]. In addition to thermodynamic properties such as temperature and entropy, black branes possess hydrodynamic characteristics of continuous uids: viscosity, diusion constants, etc. From the perspective of the holographic principle [5, 6], a black brane corresponds to a certain nite-temperature quantum eld theory in fewer number of spacetime dimensions, and the hydrodynamic behavior of a black-brane horizon is identied with the hydrodynamic behavior of the dual theory. For these eld theories, in this Letter we show that the ratio of the shear viscosity to the volume density of entropy has a universal value = 6.08 1013 K s . s 4kB (1)

Furthermore, we shall argue that this is the lowest bound on the ratio /s for a wide class of thermal quantum eld theories. Viscosity and graviton absorption.Consider a thermal eld theory whose dual holographic description involves a D-dimensional black-brane metric of the form ds2 = gM N dxM dxN = f () (dx2 + dy 2 ) + g () d d .
(0)

(2)

[The O(2) symmetry of the background is required for the existence of the shear hydrodynamic mode in the dual theory, thus making the notion of shear viscosity meaningful.] One can have in mind, as an example, the near-extremal D3-brane in type IIB supergravity, dual to nite-temperature N = 4 supersymmetric SU(Nc ) YangMills theory in the limit of large Nc and large t Hooft coupling [7, 8, 9, 10], ds2 = r2 r4 1 0 R2 r4 dt2 + dx2 + dy 2 + dz 2 + R2 dr 2 , 4 r 2 (1 r0 /r 4 ) (3)

but our discussion will be quite general. All black branes have an event horizon [r = r0 for the metric (3)], which is extended along several spatial dimensions [x, y, z in the case of (3)]. The dual eld theory is at a nite temperature, equal to the Hawking temperature of the black brane. The entropy of the dual eld theory is equal to the entropy of the black brane, which is proportional to the area of its event horizon, S= A , 4G 2 (4)

where G is Newtons constant (we set = c = kB = 1). For black branes A contains a trivial innite factor V equal to the spatial volume along directions parallel to the horizon. The entropy density s is equal to a/(4G), where a = A/V . The shear viscosity of the dual theory can be computed from gravity in a number of equivalent approaches [11, 12, 13]. Here we use Kubos formula, which relates viscosity to equilibrium correlation functions. In a rotationally invariant eld theory, = lim
0

1 2

dt dx eit [Txy (t, x), Txy (0, 0)] .

(5)

Here Txy is the xy component of the stress-energy tensor (one can replace Txy by any component of the traceless part of the stress tensor). We shall now relate the right-hand side of (5) to the absorption cross section of low-energy gravitons. According to the gaugegravity duality [10], the stress-energy tensor T couples to metric perturbations at the boundary. Following Klebanov [14, 15], let us consider a graviton of frequency , polarized in the xy direction, and propagating perpendicularly to the brane. In the eld theory picture, the absorption cross section of the graviton by the brane measures the imaginary part of the retarded Greens function of the operator coupled to hxy , i.e., Txy , abs () = 22 2 Im GR () = dt dx eit [Txy (t, x), Txy (0, 0)] , (6)

where = 8G appears due to the normalization of the gravitons action. Comparing (5) and (6), we nd abs (0) abs (0) = . (7) = 2 2 16G Graviton absoprtion cross section at low energy.The absorption cross section abs is calculable classically by solving the linearised wave equation for hx . We now show that hx = y y hxy /f obeys the equation for a minimally coupled massless scalar in the background (2). This is similar to cosmological tensor perturbations on a FriedmannRobertsonWalker spacetime, which obey the equation for a massless scalar eld [16]. (0) Consider small perturbations around the metric, gM N = gM N + hM N . We assume that the only non-vanishing component of hM N is hxy , and that it does not depend on x and y: hxy = hxy (). This eld has spin 2 under the O(2) rotational symmetry in the xy plane, which implies that all other components of hM N can be consistently set to zero [12]. Einsteins equations can be written in the form RM N = TM N T gM N , D2 (8)

where the stress-energy tensor TM N and its trace T depend on other elds such as the dilaton and various forms supporting the background (2), for example, the elds appearing 3

in the low-energy type II string theory. Again, O(2) xy rotational symmetry implies that all perturbations of matter elds can be set to zero consistently. Thus when M and N are x or y, the right-hand side of Einsteins equations reads (, = x, y) T T T (0) g = L + D2 D2 ( f + h ) , (9)

where L is the Lagrange density of matter elds and T (0) is the trace of the unperturbed stress-energy tensor. TM N = gM N L + , (10) where L represents the Lagrangian density of the elds, and dots denote terms of second and higher orders in the perturbation hxy . Substituting the unperturbed metric (2) into Einsteins equations, one nds 1 2 f (f )2 T (0) =L+ . f f2 D2 (11)

Expanding Einsteins equations to linear order in hxy , one has Rxy = 1 (f )2 T (0) 1 hxy . hxy + f hxy hxy = L + 2 f 2f 2 D2 f (f )2 f hxy + 2 hxy hxy = 0 . 2 f f f (12)

Combining Eqs. (11) and (12), we obtain an equation for hxy : hxy 2 (13)

Changing the variable from hxy to hx = hxy /f , one can see that hx indeed satises the y y equation for a minimally coupled massless scalar: hx = 0. The absorption cross section of y a graviton is therefore the same as that of the scalar. The absorption cross section for the scalar is constrained by a theorem [17, 18], which states that in the low-frequency limit 0 this cross section is equal to the area of the horizon, abs = a. Since s = a/4G, one immediately nds that = , s 4kB (14)

where and kB are now restored. This shows that the ratio /s does not depend on the concrete form of the metric within the assumptions of Refs. [17, 18]. Indeed, this ratio is the same for Dp- ([11, 13]), M2- and M5- ([19]) branes and for deformations of the D3 metric [13, 20]. This fact is very surprising, given that the corresponding dual eld theories are very dierent. We do not have an explanation for the constancy of /s in these theories based on eld-theoretical arguments alone. 4

A viscosity bound conjecture.Most quantum eld theories do not have simple gravity duals. Is our result relevant in a broader setting? We speculate that the ratio /s has a lower bound (15) s 4kB for all relativistic quantum eld theories at nite temperature and zero chemical potential. The inequality is saturated by theories with gravity duals. One argument supporting the bound (15) comes from the Heisenberg uncertainty principle. The viscosity of a plasma is proportional to mft , where is the energy density and mft is the typical mean free time of a quasiparticle. The entropy density, on the other hand, 1 is proportional to the density of quasiparticles, s kB n. Therefore /s kB mft /n. Now /n is the average energy per particle. According to the uncertainty principle, the product of the energy of a quasiparticle /n and its mean free time mft cannot be smaller than , otherwise the quasiparticle concept does not make sense. Therefore we obtain, from the uncertainty principle alone, that /s /kB , which is (15) without the numerical coecient of 1/(4). We also conclude that /s is much larger than /kB in weakly coupled theories (where the mean free time is large). Another piece of evidence supporting the bound (15) comes from a recent calculation [21] of /s in the N = 4 supersymmetric SU(Nc ) YangMills theories in the regime of innite Nc and large, but nite, t Hooft coupling g 2 Nc . The rst correction in inverse powers of g 2 Nc corresponds to the rst string theory correction to Einsteins gravity. The result reads = s 4kB 1+ 135(3) + 8(2g 2Nc )3/2 (16)

where (3) 1.2020569... is Aprys constant. The correction is positive, in accordance with e (15). It is natural to assume that /s is larger than the bound for all values of the t Hooft coupling (Fig. 1). The bound (15), in contrast to the entropy bound [22] and Bekensteins bound [23], does not involve the speed of light c and hence is nontrivial when applied to nonrelativistic systems. However, the range of applicability of (15) to nonrelativistic systems is less certain. On the one hand, by subdividing the molecules of a gas to an ever-increasing number of nonidentical species one can increase the entropy density (by adding the Gibbs mixing entropy) without substantially aecting the viscosity. On the other hand, the conjectured bound is far below the ratio of /s in any laboratory liquid. For water under normal conditions /s is 380 times larger than /(4kB ). Using standard tables [24, 25] one can nd /s for many liquids and gases at dierent temperatures and pressures. Figure 2 shows temperature dependence of /s, normalized by /(4kB ), for a few substances at dierent pressures. It is clear that the viscosity bound is well satised for these substances. Liquid helium reaches the smallest 5

h 4kB 0 g 2 Nc

Figure 1: The dependence of the ratio /s on the t Hooft coupling g 2Nc in N = 4 supersymmetric YangMills theory. The ratio diverges in the limit g 2Nc 0 and approaches /4kB from above as g 2Nc . The ratio is unknown in the regime of intermediate t Hooft coupling. value of /s, but this value still exceeds the bound by a factor of about 9. We speculate that the bound (15) is valid at least for a single-component non-relativisitic gas of particles with spin 0 or 1/2. Discussion.It is important to avoid some common misconceptions which at rst sight seem to invalidate the viscosity bound. Somewhat counterintuitively, a near-ideal gas has a very large viscosity due to the large mean free path. Likewise, superuids have nite and measurable shear viscosity associated with the normal component, according to Landaus two-component theory. The bound (15) is most useful for strongly interacting systems where reliable theoretical estimates of the viscosity are not available. One of such systems is the quark-gluon plasma (QGP) created in heavy ion collisions which behaves in many respects as a droplet of a liquid. There are experimental hints that the viscosity of the QGP at temperatures achieved by the Relativistic Heavy Ion Collider is surprisingly small, possibly close to saturating the viscosity bound [26]. Another possible application of the viscosity bound is trapped atomic gases. By using the Feshbach resonance, strongly interacting Fermi gases of atoms have been created recently. These gases have been observed to behave hydrodynamically [27] and should have nite shear viscosity at nonzero temperature. It would be very interesting to test 6

200 4 hs 150 Helium 0.1MPa Nitrogen 10MPa Water 100MPa

100

50 Viscosity bound

10 T, K

100

1000

Figure 2: The viscosity-entropy ratio for some common substances: helium, nitrogen and water. The ratio is always substantially larger than its value in theories with gravity duals, represented by the horizontal line marked viscosity bound. experimentally whether the shear viscosity of these gases satises the conjectured bound. This work was supported by DOE grant DE-FG02-00ER41132, the National Science Foundation and the Alfred P. Sloan Foundation.

References
[1] S.W. Hawking, Particle creation by black holes, Commun. Math. Phys. 43, 199 (1975). [2] J.D. Bekenstein, Black holes and entropy, Phys. Rev. D 7, 2333 (1973). [3] G.T. Horowitz and A. Strominger, Black strings and p-branes, Nucl. Phys. B 360, 197 (1991). [4] L.D. Landau and E.M. Lifshitz, Fluid Mechanics (Pergamon Press, New York, 1987). [5] G. t Hooft, Dimensional reduction in quantum gravity, gr-qc/9310026. 7

[6] L. Susskind, The world as a hologram, J. Math. Phys. 36, 6377 (1995), hep-th/9409089. [7] J.M. Maldacena, The large N limit of superconformal eld theories and supergravity, Adv. Theor. Math. Phys. 2, 231 (1998), hep-th/9711200. [8] S.S. Gubser, I.R. Klebanov, and A.M. Polyakov, Gauge theory correlators from noncritical string theory, Phys. Lett. B 428, 105 (1998), hep-th/9802109. [9] E. Witten, Antide Sitter space and holography, Adv. Theor. Math. Phys. 2, 253 (1998), hep-th/9802150. [10] O. Aharony, S.S. Gubser, J.M. Maldacena, H. Ooguri, and Y. Oz, Large N eld theories, string theory and gravity, Phys. Rep. 323, 183 (2000), hep-th/9905111. [11] G. Policastro, D.T. Son, and A.O. Starinets, Shear viscosity of strongly coupled N = 4 supersymmetric Yang-Mills plasma, Phys. Rev. Lett. 87, 081601 (2001), hep-th/0104066. [12] G. Policastro, D.T. Son, and A.O. Starinets, From AdS/CFT correspondence to hydrodynamics, J. High Energy Phys. 0209, 043 (2002), hep-th/0205052. [13] P. Kovtun, D.T. Son, and A.O. Starinets, Holography and hydrodynamics: Diusion on stretched horizons, J. High Energy Phys. 0310, 064 (2003), hep-th/0309213. [14] I.R. Klebanov, World-volume approach to absorption by non-dilatonic branes, Nucl. Phys. B 496, 231 (1997), hep-th/9702076. [15] See also S. S. Gubser, I. R. Klebanov, and A. A. Tseytlin, String theory and classical absorption by three-branes, Nucl. Phys. B 499, 217 (1997), hep-th/9703040. [16] A.R. Liddle and D.H. Lyth, Cosmological Ination and Large-Scale Structure (Cambridge University Press, Cambridge, 2000). [17] S.R. Das, G.H. Gibbons, and S.D. Mathur, Universality of low energy absorption cross sections for black holes, Phys. Rev. Lett. 78, 417 (1997), hep-th/9609052. [18] R. Emparan, Absorption of scalars by extended objects, Nucl. Phys. B 516, 297 (1998), hep-th/9706204. [19] C.P. Herzog, The hydrodynamics of M-theory, J. High Energy Phys. 0212, 026 (2002), hep-th/0210126.

[20] A. Buchel and J.T. Liu, Universality of the shear viscosity from supergravity duals, Phys. Rev. Lett. 93, 090602 (2004), hep-th/0311175. [21] A. Buchel, J.T. Liu, and A.O. Starinets, Coupling constant dependence of the shear viscosity in N = 4 supersymmetric Yang-Mills theory, Nucl. Phys. B 707, 56 (2005), hep-th/0406264. [22] R. Bousso, A covariant entropy conjecture, J. High Energy Phys. 9907, 004 (1999), hep-th/9905177. [23] J.D. Bekenstein, A universal upper bound on the entropy to energy ratio for bounded systems, Phys. Rev. D 23, 287 (1981). [24] E.W. Lemmon, M.O. McLinden, and D.G. Friend, Thermophysical Properties of Fluid Systems, in NIST Chemistry WebBook, NIST Standard Reference Database Number 69, Eds. Linstrom P.G. & Mallard, W.G., March 2003 (National Institute of Standards and Technology, Gaithersburg MD, 20899, http://webbook.nist.gov). [25] J.D. Cox, D.D. Wagman, and V.A. Medvedev, CODATA Key Values for Thermodynamics (Hemisphere Publishing Corp., New York, 1989, http://www.codata.org). [26] E. Shuryak, Why does the quark-gluon plasma at RHIC behave as a nearly ideal uid? Prog. Part. Nucl. Phys. 53, 273 (2004), hep-ph/0312227. [27] K.M. OHara, S.L. Hemmer, M.E. Gehm, S.R. Granade, and J.E. Thomas, Observation of a strongly-interacting degenerate Fermi gas of atoms, Science 298, 2179 (2002), cond-mat/0212463.