RELATIONS BETWEEN TRANSPORT & CHAOS IN HOLOGRAPHIC THEORIES
Richard Davison
Heriot-Watt University
HoloTube, May 2020
RELATIONS BETWEEN TRANSPORT & CHAOS IN HOLOGRAPHIC THEORIES - - PowerPoint PPT Presentation
RELATIONS BETWEEN TRANSPORT & CHAOS IN HOLOGRAPHIC THEORIES Richard Davison Heriot-Watt University HoloTube, May 2020 OVERVIEW In some QFTs, there are connections between transport properties and underlying chaotic dynamics. First
Richard Davison
Heriot-Watt University
HoloTube, May 2020
underlying chaotic dynamics.
charge. Its relation to underlying chaotic dynamics turns out to not be very robust.
transport and chaos. I will focus on QFTs with a gravity description.
Blake
charges (e.g. energy, momentum, charge, etc.)
✴ Easy to measure ✴ Exhibit universality in interesting materials
U(1)
e.g. : change in energy density due to a small source
dispersion relations
✴ When the system is in local equilibrium,
transport is governed by simple effective theories: hydrodynamics.
✴ In this regime, transport is dominated
by a handful of gapless modes
Gεε(ω, k) ⟨ε⟩ ⟷ ω(k) ωhydro(k)
ω k
ωhydro(k) ωeq
keq
? ? ?
✴ Local thermal equilibrium
state characterized by slowly-varying
✴ Dynamics of this variable are constrained by symmetries:
Energy diffuses over long distances.
etc) depend on the details of the particular system. In a Fermi liquid, .
⟶ ε(x, t) ⟶ ∂tε = D∇2ε + Γ∇4ε + O(∇6) ωhydro(k) = − iDk2 − iΓk4 + O(k6) D, Γ D ∼ v2
Fτqp
✴ The timescale is always ✴ But the “butterfly velocity”
depends on the particular theory.
✴ In the gravity description, governed by near-horizon physics.
τL = (2πT)−1 vB
C(t, x) = − ⟨[V(t, x), W(0,0)]2⟩T C(t, x) ∼ eτ−1
L (t − |x|/vB)
Shenker, Stanford Roberts, Stanford, Susskind
energy whose dispersion relation obeys where .
vB, τL ω(k*) = iτL k2
* = − (vBτL)−2
Blake, RD, Grozdanov, Liu
D ∼ v2
BτL
Im(ω) Im(k)
(k*, ω*)
see also: Grozdanov, Schalm, Scopelliti
for And this mode is approximately diffusive up to
ωhydro(k*) = iτL k2
* = − (vBτL)−2
ωhydro(k) ∼ − iDk2 k = k*
σ
V V
W W
σ
hydrodynamic mode of energy conservation :
Blake, Lee, Liu
D ∼ v2
BτL
Arising as classical solutions of
ds2 = − f(r)dv2 + 2dvdr + h(r)dx2
d
S = ∫ dd+2x −g (R − Z(ϕ)F2 − 1 2(∂ϕ)2 + V(ϕ) − Y(ϕ)
d
∑
i=1
(∂χi)
2
)
ϕ(r) ≠ 0 Fvr(r) ≠ 0 χi = mxi
and
horizon :
∂a( −g∂aφ) − m2 −gφ = 0 φnorm(r, ω, k) φnon−norm(r, ω, k) r = r0 φingoing = a(ω, k)φnorm + b(ω, k)φnon−norm G(ω, k) = b(ω, k) a(ω, k)
r0 r
) depend in detail on the metric and matter field profiles throughout the spacetime. i.e. on many specific details of the particular QFT state.
✴
, limit where radial evolution is simple:
✴ Points in Fourier space
where the ingoing solution is not unique features of the Green’s functions that are insensitive to many details of the state
Gεε ω → 0 k → 0 d dr ( . . . φ′ (r)) = 0 (ω*, k*)
where the ingoing solution is not unique exact constraints on the spectrum
✴ Ansatz: solution that is regular at the horizon ✴ Solve iteratively for
: etc.
✴ At
both solutions are regular at the horizon
(ω*, k*) ω(k) φ(r) =
∞
∑
n=0
φn(r − r0)n φn>0 2h(r0)(2πT − iω)φ1 = (k2 + m2h(r0) + iω dh′ (r0) 2 ) φ0 (ω*, k*) ω* = − i2πT, k2
* = − (m2h(r0) + dπTh′
(r0))
Blake, RD, Vegh ; see also Grozdanov et al
yields one regular solution: But this regular solution depends on the arbitrary slope .
and by tuning : there must be a pole with dispersion relation obeying
(ω*, k*) δk/δω φnorm φnon−norm δk/δω φingoing(ω* + iδω, k* + iδk) = C (1 − vz δk δω) φnorm + (1 − vp δk δω ) φnon−norm G(ω* + iδω, k* + iδk) = C δω − vzδk δω − vpδk ω(k*) = ω*
ω = ω* + iδω k = k* + iδk φ1 φ0 = 1 4h(r0) (4ik* δk δω − dh′ (r0))
is obtained only from the near-horizon dynamics. a part of the spectrum that is independent of the rest of the spacetime
e.g.
ω(k)
e.g. Ceplak, Ramdial, Vegh a collective mode of the momentum density operator (Schwarzschild-AdS4)
exact (numerical) dispersion relation ω(k) near-horizon constraints
, the equations are seemingly more complicated. couples to other metric perturbations and to matter field perturbations e.g.
e.g. higher-derivative gravity, magnetic fields & anomalies
Gεε(ω, k) δgvv
(−iωdh′ (r0) + 2k2) δgvv(r0) − 2 (2πT + iω) (ωδgxixi(r0) + 2kδgvx(r0)) = 4h(r0)(δTvv(r0) − Tvr(r0)δgvv(r0))
ω(k*) = + i2πT k2
* = − dπTh′
(r0) ω(k*) = + iτ−1
L
k2
* = − (vBτL)−2
Grozdanov ; Abbasi, Tabatabaei
modes to obey the universal constraint. In a few cases, this has been verified.
is an excellent approximation up to
ωhydro(k) ≈ = − iDk2 k = k*
Grozdanov, Schalm, Scopelliti ; Blake, RD, Grozdanov, Liu
D ≈ v2
BτL
(k*, ω*)
Im(ω) 2πT
2πT Im(k) ω = − iDk2 exact ω(k)
“linear axion” theory at low T
ω
non-hydro modes hydro mode
Thermal conductivity is independent of the matter field profiles
Also independent of matter field profiles.
κ T c
D = κ c
κ ≡ κ − α2T σ = 4π f′ (r)h(r)d−2
d dr (f′
(r)h(r)d/2−1)
r=r0 Blake, RD, Sachdev
✴ For a large class of theories with AdS2xRd IR fixed points ✴ Generic IR fixed point has symmetry ✴ When
, diffusive approximation breaks down at .
t → Λzt, x → Λx z = 1 ω ≪ τ−1
L
T → 0
as
Blake, RD, Sachdev
D = v2
BτL
T → 0
as
Blake, Donos
D = z 2(z − 1) v2
BτL
RD, Gentle, Goutéraux
underlying chaotic dynamics.
collective modes. There is a universal constraint for collective modes of energy density
D ∼ v2
BτL
ω(k*) = + iτ−1
L
k2
* = − (vBτL)−2
✴ Effective action of holographic theories ✴ Generalisations outside holography ✴ Regime of validity of (diffusive) hydro in holographic theories?
✴ Necessary conditions in a QFT? ✴ Robustness of universal constraint at ✴ Constraints on other transport coefficients?
ω = + i2πT
e.g. Gu, Qi, Stanford ; Patel, Sachdev ; Gu, Lucas, Qi ; Grozdanov, Schalm, Scopelliti ; … e.g. Hartman, Hartnoll, Mahajan ; Lucas ; Withers ; Grozdanov et al ; ….