Global Stability for Charged Scalar Fields in Spacetimes close to - - PowerPoint PPT Presentation

global stability for charged scalar fields in spacetimes
SMART_READER_LITE
LIVE PREVIEW

Global Stability for Charged Scalar Fields in Spacetimes close to - - PowerPoint PPT Presentation

Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski Christopher Kauffman Johns Hopkins University Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski


slide-1
SLIDE 1

Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski

Christopher Kauffman Johns Hopkins University

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 1 / 42

slide-2
SLIDE 2

Table of Contents

1

Introduction

2

Background

3

The Proof

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 2 / 42

slide-3
SLIDE 3

Table of Contents

1

Introduction

2

Background

3

The Proof

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 3 / 42

slide-4
SLIDE 4

The Einstein Field Equations

As motivation for this work, we consider the Einstein field equations with zero cosmological constant, Rµν − 1 2Rgµν = 8πQµν, (1) where Rµν is the Ricci curvature associated with the metric g, R is the scalar curvature, and Q is the energy-momentum tensor associated with some set of field equations.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 4 / 42

slide-5
SLIDE 5

The Einstein Field Equations

As motivation for this work, we consider the Einstein field equations with zero cosmological constant, Rµν − 1 2Rgµν = 8πQµν, (1) where Rµν is the Ricci curvature associated with the metric g, R is the scalar curvature, and Q is the energy-momentum tensor associated with some set of field equations. The simplest such model is the Einstein vacuum equations (Qµν = 0), which can be written as Rµν = 0.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 4 / 42

slide-6
SLIDE 6

Properties of the Einstein Vacuum Equations

This system is of interest from a mathematical standpoint for several reasons:

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 5 / 42

slide-7
SLIDE 7

Properties of the Einstein Vacuum Equations

This system is of interest from a mathematical standpoint for several reasons: It admits nontrivial solutions. Two sets of famous exact solutions are the Schwarzschild and Kerr solutions, which model black hole spacetimes.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 5 / 42

slide-8
SLIDE 8

Properties of the Einstein Vacuum Equations

This system is of interest from a mathematical standpoint for several reasons: It admits nontrivial solutions. Two sets of famous exact solutions are the Schwarzschild and Kerr solutions, which model black hole spacetimes. This system linearizes to the wave equation around the Minkowski

  • spacetime. That is, for a choice of coordinates we can write the

system as gαβ∂α∂β(gµν) = Pµν(g)(∂g, ∂g), for a quadratic form Pµν(g).

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 5 / 42

slide-9
SLIDE 9

Properties of the Einstein Vacuum Equations

This system is of interest from a mathematical standpoint for several reasons: It admits nontrivial solutions. Two sets of famous exact solutions are the Schwarzschild and Kerr solutions, which model black hole spacetimes. This system linearizes to the wave equation around the Minkowski

  • spacetime. That is, for a choice of coordinates we can write the

system as gαβ∂α∂β(gµν) = Pµν(g)(∂g, ∂g), for a quadratic form Pµν(g). Finally, the background spacetime for solutions of Einstein’s field equations can often be modelled by solutions of the vacuum

  • equations. This means we can in a sense treat the equations

modelling the field and spacetime separately.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 5 / 42

slide-10
SLIDE 10

The Background Spacetime

Solutions to Einstein’s vacuum equations with small initial data can be modelled by the metric gαβ = mαβ + Mχ 1 + r δαβ + hαβ, where m is the Minkowski metric, Mχ

1+r δαβ is spherically symmetric (where

δ is the Kronecker delta and M is a small parameter corresponding to the mass), and h is a small perturbation satisfying certain decay bounds. We have by the positive mass theorem that M > 0.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 6 / 42

slide-11
SLIDE 11

The Background Spacetime

Solutions to Einstein’s vacuum equations with small initial data can be modelled by the metric gαβ = mαβ + Mχ 1 + r δαβ + hαβ, where m is the Minkowski metric, Mχ

1+r δαβ is spherically symmetric (where

δ is the Kronecker delta and M is a small parameter corresponding to the mass), and h is a small perturbation satisfying certain decay bounds. We have by the positive mass theorem that M > 0. The first two terms imply that the metric behaves like the Schwarzschild metric in the far exterior. The third term is naturally of the most interest when establishing global stability results for the Einstein Vacuum Equations.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 6 / 42

slide-12
SLIDE 12

The Maxwell-Klein-Gordon system

We choose as our set of field equations the Maxwell-Klein-Gordon system, C

g φ = DαDαφ = 0,

(2a) ∇βFαβ = ℑ(φDαφ), (2b) where, for a one-form A, we define F = dA, Dµ = ∇µ + iAµ. Additionally, we define the current vector Jα = ℑ(φDαφ).

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 7 / 42

slide-13
SLIDE 13

The Main Theorem

We are ready to state our main theorem:

Theorem

Given a background metric g satisfying bounds consistent with small-data solutions of the Einstein Vacuum Equations in harmonic gauge, the system (2) is globally well-posed for small initial data. Additionally, given some initial energy on k derivatives, Ek[F, φ], with k ≥ 11, the energy-momentum tensor, Q[F, φ]αβ = ℜ

  • DαφDβφ − 1

2gαβDγφDγφ

  • +FαγF γ

β − 1

4gαβFγδF γδ, (3) satisfies the following estimates: Q[F, φ](t, ·)L1 Ek[F, φ] (4) Q[F, φ](t, ·)L2 Ek[F, φ](1 + t)−1 (5)

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 8 / 42

slide-14
SLIDE 14

The Main Theorem: Remarks

The estimate on Q[F, φ](t, ·)L2 follows directly from decay estimates (and is sharp in time decay in only one component of Q); however, this is in a sense the most important estimate for closing the argument for the full Einstein-Maxwell-Klein-Gordon system. In particular, in an energy estimate we can generally use this to establish slowly growing energy for the metric, which is consistent with results in (Lindblad and Rodnianski 2010). Estimates of this sort have been of great use in establishing global stability for the Einstein-Vlasov system in (Lindblad and Taylor 2017)

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 9 / 42

slide-15
SLIDE 15

Comparison to Previous Work

Global Stability and Decay for the MKG system

(Eardley and Moncrief 1982) and (Klainerman and Machedon 1994): Global existence for finite initial data in Coulomb gauge. (Lindblad and Sterbenz 2006): Nice decay estimates in a gauge-free setting for small initial data. (Bieri, Miao, and Shahshahani 2017): Provided a simpler proof of results found by Lindblad and Sterbenz. (Yang 2015): Removed the smallness assumption on the Maxwell field.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 10 / 42

slide-16
SLIDE 16

Comparison to Previous Work

Global Stability and Decay for the MKG system

(Eardley and Moncrief 1982) and (Klainerman and Machedon 1994): Global existence for finite initial data in Coulomb gauge. (Lindblad and Sterbenz 2006): Nice decay estimates in a gauge-free setting for small initial data. (Bieri, Miao, and Shahshahani 2017): Provided a simpler proof of results found by Lindblad and Sterbenz. (Yang 2015): Removed the smallness assumption on the Maxwell field.

The Einstein Field Equations

(Choquet-Bruhat 1952): Local stability of the Minkowski spacetime (Christodoulou and Klainerman 1990): Global stability of the Minkowski spacetime (Lindblad and Rodnianski 2010): Global stability of the Minkowski in harmonic coordinates (Zipser 2000): Global Stability for the Einstein-Maxwell system (Loizelet 2008): Global Stability for Einstein Maxwell in Lorenz gauge (harmonic coordinates) (Speck 2012): Global Stability for Einstein-Maxwell system in a gauge-free setting (harmonic coordinates)

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 10 / 42

slide-17
SLIDE 17

Table of Contents

1

Introduction

2

Background

3

The Proof

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 11 / 42

slide-18
SLIDE 18

The Lorentz Vector Fields

It is well-known that for the vector fields L = {∂α, Ωij = xi∂j − xj∂i, Ω0i = t∂i + xi∂t, S = t∂t + r∂r}, if u = f , (Zu) = Zf + cZf , where cZ = 2 if Z = S and cZ = 0 for all

  • ther fields in L.

This follows from the calculation [Z, ]φ = −(LZm)αβ∂α∂βφ − (Z γ)∂γφ. The first term on the right is a scalar multiple of φ which comes from the Killing and conformal Killing nature of the fields Z, and the second term is 0, since all components of Z are at most first degree polynomials.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 12 / 42

slide-19
SLIDE 19

The Null Frame

One notable consequence of this is that certain derivatives of φ decay at different rates. In particular, we define L = ∂t + ∂r, L = ∂t − ∂r, as well as a set of piecewise defined derivatives S1, S2 which are tangent to the sphere. Then, roughly speaking, if φ decays like t−1, we can generally expect the following estimates for derivatives of φ: |Lφ| +

  • i

|Siφ| (1 + t)−2, |Lφ| (1 + t)−1(1 + |r − t|)−1. In particular, the latter has worse decay close to the light cone t = r.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 13 / 42

slide-20
SLIDE 20

The Null Condition and Asymptotic Systems

As an application, we consider the equation u = −(∂tu)2, in R1+3, with the initial conditions u(0, x) = ǫu0(x), ∂tu(0, x) = ǫu1(x), u0, u1 ∈ C ∞

0 .

We can rewrite this as u = −1 r LL(ru) + ∆ωu = −1 4 1 r L(ru) + 1 r L(ru) 2 . where ∆ω denotes the spherical Laplacian. If we assume all derivatives except L are negligible, we get the equation LL(ru) = 1 4r |L(ru)|2. Solutions of this equation, and the 1 + 3-dimensional equation it models, can blow up in finite time, no matter how small the initial data!

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 14 / 42

slide-21
SLIDE 21

The Null Condition and Asymptotic Systems (Cont.)

For certain classes of quasilinear equations, we have global existence for small initial data. One famous example is the equation u = (∂tu)2 − |∇xu|2, which has global solutions for sufficiently small ǫ (cf. (Klainerman 1986)). In our null decomposition, the right hand side is equal to Lu · Lu − | / ∇u|2, i.e. the problematic term, |Lu|2, is absent! The asymptotic system for this equation is LL(ru) = 0, which of course has global existence for all time. This in general holds for systems where the right hand side is composed of null forms Q0[φ, ψ] = mαβ∂αφ∂βψ, Qαβ[φ, ψ] = ∂αφ∂βψ − ∂αψ∂βφ.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 15 / 42

slide-22
SLIDE 22

The Weak Null Condition and Global Stability of Minkowski

For all quadratic nonlinearities satisfying the null condition, we have the asymptotic system LL(ru) = 0, which leads to global existence. However, this is not the only case when the asymptotic system doesn’t blow up. Consider the system ψ1 = 0, (6a) ψ2 = (∂tψ1)2. (6b) This has the asymptotic system LL(rψ1) = 0, LL(rψ2) = 1 4r |L(rψ1)|2.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 16 / 42

slide-23
SLIDE 23

The Weak Null Condition and Global Stability of Minkowski (Cont.)

This has global solutions, with solutions growing like (rψ1) ≈ 1, (rψ2) ≈ ln(r). It is known that solutions to this reduction also model the asymptotic behavior of the full system (6). It was shown in the celebrated work (Lindblad and Rodnianski 2010), that the accuracy of this model extends to solutions to Einstein’s Equations.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 17 / 42

slide-24
SLIDE 24

The Modified Frame

Due to the Mχ

1+r δαβ term in the metric g, we cannot reasonably expect gLL

to decay any better than ǫt−1 along the light cone t = r. This error term leads to slowly increasing energy in the standard energy estimate (cf. (Lindblad and Rodnianski 2010)), and growing energy in the conformal Morawetz estimate. We account for this in two steps: First, we use the modified radial coordinate r∗ = r + Mχ ln(1 + r), t∗ = t and null frame L∗ = ∂t∗ + ∂r∗, L∗ = ∂t∗ − ∂r∗, S∗

i = r

r∗ Si, where Si are piecewise defined orthonormal vectors tangent to the sphere. The use of r∗ here comes from similar analysis on the Schwarzschild metric in the exterior, and will be of use in bounding certain error terms stemming from the massive part of the metric.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 18 / 42

slide-25
SLIDE 25

The Modified Frame

Due to the Mχ

1+r δαβ term in the metric g, we cannot reasonably expect gLL

to decay any better than ǫt−1 along the light cone t = r. This error term leads to slowly increasing energy in the standard energy estimate (cf. (Lindblad and Rodnianski 2010)), and growing energy in the conformal Morawetz estimate. We account for this in two steps: First, we use the modified radial coordinate r∗ = r + Mχ ln(1 + r), t∗ = t and null frame L∗ = ∂t∗ + ∂r∗, L∗ = ∂t∗ − ∂r∗, S∗

i = r

r∗ Si, where Si are piecewise defined orthonormal vectors tangent to the sphere. The use of r∗ here comes from similar analysis on the Schwarzschild metric in the exterior, and will be of use in bounding certain error terms stemming from the massive part of the metric. Second, we carry out our energy estimates with respect to the fractional acceleration field, K s

0 = 1

2(1 + |t + r∗|2s)L∗ + 1 2(1 + |t − r∗|2s)L∗.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 18 / 42

slide-26
SLIDE 26

The Modified Lorentz Fields

In addition to the modified radial coordinate and frame, we also define x∗i = r∗ωi and consequently the Lorentz fields L∗ = {∂x∗α, Ω∗

ij = x∗i∂x∗j − x∗j∂x∗i, Ω∗ 0i = t∂x∗i + x∗i∂t∗, S∗ = t∂t∗ + r∗∂r∗}.

We generally use Z to refer to any of these fields. These are analogous to the Lorentz fields in Minkowski space (and in fact the rotation fields are unchanged). This modification of the fields is in particular necessary in making sure null components of Lie derivatives like [Z, L∗] have the right decay along the light cone.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 19 / 42

slide-27
SLIDE 27

The Maxwell-Klein-Gordon Equations: The Weak Null Condition

We recall the potential Aµ. We have some freedom in the choice of gauge; in particular, we can assume A satisfies the Lorenz gauge condition, ∇ · A = 0. In Minkowski space, this reduces the Maxwell-Klein-Gordon system to a system of semilinear wave equations, Aµ = ℑ(φDµφ), (7) φ = −2iAµ∂µφ + AµAµφ. (8) Intuitively, every time a bad derivative or component of A appears on the right of (8), it is paired with a good derivative or component. In (7), this holds everywhere except for the component AL, for which we can similarly expect decay like (1 + t)−1 ln(1 + t) along the light cone.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 20 / 42

slide-28
SLIDE 28

Properties of the Metric

We recall the assumption that the metric is of the form gαβ = mαβ + Mχ 1 + r δαβ + hαβ. Here m is the Minkowski metric, M ≤ ǫ, and h satisfies the L∞ estimates |LI

Zh| ≤ ǫτ −1+ι +

, |LI

ZhLT | ≤ ǫτ −1−γ′+ι +

τ γ′

− ,

for τ+ = |t + r∗| , τ− = |t − r∗| , L ∈ {L∗}, T = {L∗, S∗

1.S∗ 2}. We take as

well the L2 estimates

  • τ −1/2−ι

+

(|∂LI

Zh| + τ −1 − |LI Zh|)w1/2 1

  • L2(t,x) ≤ ǫ,
  • τ −1/2

τ −ι

+ |(∂LI Zh)LL| + |∂LI Zh| + τ −1 − |(LI Zh)LL|)w1/2 1

  • L2(t,x) ≤ ǫ,

for the simplified weight w1 = (r∗ − t)+, and Z ∈ L∗.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 21 / 42

slide-29
SLIDE 29

Properties of the Metric (Cont.)

The nicer estimates on hLT and hLL are a consequence of the harmonic coordinate condition. In particular, from the harmonic coordinate condition we have the rough identification L(gLT ) ≈ ∂g These in particular give nicer estimates for the Lie derivative (LK s

0 g)L∗L∗,

which are necessary to control the certain components in the energy estimate. We can say equivalently that u∗ = t − r∗ is an approximate optical function, in the sense that we have nice decay for the quantity gαβ∂αu∗∂βu∗. This decay implies that we should look at the geometry as a small perturbation of Schwarzschild in the exterior, rather than of Minkowski, in

  • rder to get the best control over our field quantities.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 22 / 42

slide-30
SLIDE 30

Preliminaries: The Weights

In addition to the fractional acceleration field, our proof also heavily relies

  • n the use of the following weights:

w =

  • 1

r − t < 0 r − t2γ r − t > 0 w ′ =

  • r − t−1−2ι

r − t < 0 r − t2γ−1 r − t > 0 wι =

  • r − t2ι

r − t < 0 r − t2γ r − t > 0

We in particular use a weight ˜ w, which we do not define here, such that the following hold: ˜ w ≈ w, L( ˜ w) ≈ w′, L( ˜ w) ≈ t − r∗ t + r∗ 1+2ι w′.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 23 / 42

slide-31
SLIDE 31

Optical Weights and the Null Decomposition of F

We additionally define the optical weights τ+ = t + r∗ = (1 + (t + r∗)2)1/2, τ− = t − r∗ = (1 + (t − r∗)2)1/2, τ0 = τ−/τ+. Additionally, we define the null decomposition of a two-form F: αi[F] = FL∗S∗

i

αi[F] = FL∗S∗

i

ρ[F] = 1 2FL∗L∗ σ[F] = FS1S2 We have the following rough identification: αi[F] ∼ DL∗(r∗φ) r∗ , αi[F] ∼ DL∗φ, ρ[F] ∼ DSiφ, σ[F] ∼ DSiφ, which comes from the fact that

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 24 / 42

slide-32
SLIDE 32

Table of Contents

1

Introduction

2

Background

3

The Proof

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 25 / 42

slide-33
SLIDE 33

The Energy Estimate: F

We define the energy-momentum tensor for an arbitrary 2-form G as follows: Q[G]αβ = GαγG γ

β − 1

4gαβGγδG γδ. For a general two-form G with sufficient decay, we can define the time-slice energy on G E0[G](T) = sup

0≤t≤T

  • Σt
  • τ 2s

+

  • |α|2 + |ρ|2 + |σ|2

+ τ 2s

− |α|2

w dx, the interior energy S0[G](T) =

  • [0,t]×R3
  • τ 2s

+ |α|2 + τ 1+2ι

  • τ 2s

+ (|ρ|2 + |σ|2) + τ 2s − |α|2

w′ dx dt, and the light-cone energy C0[G](T) = sup

u∗

  • {C(u∗)}∩{t∈[0,T]}

Q(K s

0, ∇u)w dVC(u∗).

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 26 / 42

slide-34
SLIDE 34

The Energy Estimate: F (Cont.)

We have the following preliminary energy estimate:

Theorem

Given a 2-form Fαβ such that ∇βFαβ = Jα, and defining the norm JL2[w] =

  • τ s

+τ −1/2−ι

τ 1/2

− JL∗w1/2 ι

  • 2 +
  • τ s

+τ 1/2 − |JS∗|w1/2 ι

  • 2 +

+

  • τ s−1/2−ι

τ s+1/2

|JL∗|w1/2

ι

  • 2 ,

we have the estimate E0[F](T) + S0[F](T) + C0[F](T) E0[F](0) + J[F]2

L2[w] .

(9) We briefly sketch the proof.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 27 / 42

slide-35
SLIDE 35

The Energy Estimate: F (Proof)

Proof.

We follow a standard divergence estimate, applied to the momentum density Pα[F] = −Q[F]αβK s

βw.

First, we take the divergence theorem over either time slabs of the form [0, T] × R3, or time slabs exterior to some light cone t − r = C. The boundary terms coming from these identities in particular contain the E0 and C0 terms respectively. We now consider the interior. First, we have −(∇αw)K s

βQ[F]αβ ≈ −S0[F].

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 28 / 42

slide-36
SLIDE 36

The Energy Estimate: F (Proof Cont.)

Proof.

Next, we can discard certain terms coming from the deformation tensor

(K s

0)π, as they have the right sign due to the choice of field. The terms

coming from the interior energy therefore behave like

  • FK s

0αJαw

  • L1(t,x) + ǫ

T E[F](t) (1 + t)1+ι . We use H¨

  • lder’s inequality to deal with the first term on the left, splitting

it into a term which can be subtracted off of S0[F](T) times a term corresponding to the current norm, and note that the second term can be bounded by an arbitrarily small constant times the energy.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 29 / 42

slide-37
SLIDE 37

The Charged Portion of F

One issue that immediately crops up is that the energy norm E0[F](0) is in general not finite. In particular, we look at the elliptic consideration in the Lorenz gauge ∆A0 = ∂j(∂tAj) + J0, which follows almost directly from the Lorenz gauge condition. In particular, it is not possible for A0 to decay better than r−1, or certain components of F to decay better than r−2, unless J0 integrates to 0. Our method for accounting for this follows from (Lindblad and Sterbenz 2006), and involves subtracting off a well-defined charge quantity, F 0i = ωi q 4π χ(r∗ − t − 2)∂r(r∗) r∗2

  • .

and running subsequent analysis on the quantity F = F − F.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 30 / 42

slide-38
SLIDE 38

The Energy Estimate: φ

We can take an energy estimate on the quantity φ analogous to that for

  • F. We take the energies

E0[φ](T) = sup

0≤t≤T

  • Σt
  • τ 2s

+

  • DL∗(r ∗φ)

r ∗

  • 2

+ | / Dφ|2 +

  • φ

r ∗

  • 2

+ τ 2s

− |DL∗φ|2

  • w dx,

as well as

S0[φ](T) = T

  • Σt
  • τ 2s

+

  • DL∗(r ∗φ)

r ∗

  • 2

w ′ dx dt+ + T

  • Σt
  • τ 1+2ι
  • τ 2s

+

  • | /

Dφ|2 +

  • φ

r ∗

  • 2

+ τ 2s

  • |DL∗φ|2
  • w ′ dx dt

and

C0[φ](T) = sup

u∗

  • C(u∗)

Q(∇α

m∗u∗, K s 0)w dC(u∗).

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 31 / 42

slide-39
SLIDE 39

The Energy Estimate: φ (Cont.)

We now have the following:

Theorem

For a function φ with sufficient decay, and sufficient L∞ decay for the field F, we have the estimate E0[φ](T)+S0[φ](T)+C0[φ](T) E0[φ](0)+

  • τ s

+τ 1/2 − (C g φ)w1/2 ι

  • 2

2 . (10)

We note that the L∞ decay for F is consistent with the bounds in the next section, and only depends on the quantity F. In particular, we can establish this estimate on higher derivatives of φ without requiring higher derivatives on F.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 32 / 42

slide-40
SLIDE 40

The Energy: Combined

We define the combined energies E0[F](T) = q[F]2 + E0[ F](T) + S0[ F](T) + C0[ F](T), E0[φ](T) = E0[φ](T) + S0[φ](T) + C0[φ](T). Additionally, for a collection of vector fields Z ∈ L∗, we define Ek[F](T) = q[F]2 +

  • |I|≤k

E0[LI

Z

F](T), Ek[φ](T) =

  • |I|≤k

E0[DI

Zφ](T),

and the total energy Ek(T) = Ek[F](T) + Ek[φ](T)

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 33 / 42

slide-41
SLIDE 41

The Decay Estimates: F

We now establish decay for F. There are two parts to these calculations. First, we take decay for the charge-free portion F using the energy estimate and weighted Klainerman-Sobolev estimates. In particular we get, for |I| ≤ k − 5, |α[LI

Z

F]| E1/2

k

(T)t + r−3/2−sw−1/2, |ρ[LI

Z

F]| E1/2

k

(T)t + r−1−st − r−1/2w−1/2, |σ[LI

Z

F]| E1/2

k

(T)t + r−1−st − r−1/2w−1/2, |α[LI

Z

F]| E1/2

k

(T)t + r−1t − r−1/2−sw−1/2. The estimates on ρ, σ, α come from weighted Klainerman-Sobolev estimates over the time-slice energy, E0. The estimate on α comes from the conical energy C0. Additionally, there is an L2(t)L∞(x) estimate for α which we do not state here, which comes from the interior energy S0.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 34 / 42

slide-42
SLIDE 42

The Decay Estimates: F (Cont.)

We additionally have the following estimates on F, for all sets of vector fields I (and the implicit constant in depending on |I|): |α[LI

ZF]| |q| · t + r−3t − rH(r − t),

|ρ[LI

ZF]| |q| · t + r−2H(r − t),

|σ[LI

ZF]| |q| · t + r−2H(r − t),

|α[LI

ZF]| |q| · t + r−2H(r − t).

Here, H is the Heaviside function, meaning that F is supported outside the light cone. Here we recall that q is the charge.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 35 / 42

slide-43
SLIDE 43

The Decay Estimates: φ

We can establish analogous estimates for φ. We have |φ| E1/2

5

(T)t + r−1t − r1/2−sw−1/2, |DL∗φ| E1/2

5

(T)t + r−1t − r−1/2−sw−1/2, |DS∗φ| E1/2

5

(T)t + r−1−st − r−1/2w−1/2, |DL∗φ| E1/2

5

(T)t + r−1−st − r−1/2w−1/2,

  • χDL∗(r∗φ)

r∗

  • E1/2

5

(T)t + r−3/2−sw−1/2. The proof for these is similar to analogous estimates for F. In particular, the first four follow from the Klainerman-Sobolev inequality applied to the time slice, with a Poincare-type estimate to handle the terms where no derivatives fall on φ. The last estimate follows from a Klainerman-Soblev type estimate using the conical energy C0.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 36 / 42

slide-44
SLIDE 44

Dealing With the Commutator Terms: F

We now must bound the right hand side of our energy estimates for Lie derivatives of F. We can write J[LI

Z

F] = [J, LI

Z]

F + LI

ZJ[F] − LI ZJ[F].

We must bound these terms separately. Roughly speaking, the first term consists of error terms which behave like ∇αLI

Z(∇ · Z)(LJ ZF)αβ.

The second term can be written in terms of φ and its derivatives. The analysis of this follows closely that in (Lindblad and Sterbenz 2006), and uses L2, L∞ estimates along with the identity ℑ

  • φDαψ + ψDαφ
  • = ℑ
  • φDα(r∗ψ)

r∗ + ψDα(r∗φ) r∗

  • ,

along with similar symmetries.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 37 / 42

slide-45
SLIDE 45

Dealing With the Commutator Terms: F (Cont.)

Finally, for the last term, we use the approximation ∇βF

αβ ≈ q

4π χ′(r∗ − t − 2) r∗2 L∗α, such that Lie derivatives of this quantity (and associated commutators) can be explicitly defined and bounded, and error terms depend on the deviation of the metric from Minkowski.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 38 / 42

slide-46
SLIDE 46

Dealing With the Commutator Terms: φ

The energy estimates for derivatives of φ can be solved using a similar

  • method. It suffices to show that for |I| ≤ k we have the estimate
  • τ s

+τ− 1/2C g (DI Xφ)w1/2 ι

  • L2(t,x) Ek(T).

first using the identity [C

g , DY ]φ = −Dαφ∇α(∇·Y )+Dα

  • (Y )παβDβφ
  • −i(∇αFY αφ+2FY αDαφ).

We iterate this to establish the full commutator [C

g , DI Y ]φ. Again, we

look by this term-by-term.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 39 / 42

slide-47
SLIDE 47

Dealing With the Commutator Terms: φ

The energy estimates for derivatives of φ can be solved using a similar

  • method. It suffices to show that for |I| ≤ k we have the estimate
  • τ s

+τ− 1/2C g (DI Xφ)w1/2 ι

  • L2(t,x) Ek(T).

first using the identity [C

g , DY ]φ = −Dαφ∇α(∇·Y )+Dα

  • (Y )παβDβφ
  • −i(∇αFY αφ+2FY αDαφ).

We iterate this to establish the full commutator [C

g , DI Y ]φ. Again, we

look by this term-by-term. The first term is again an error term which scales with the metric, and we deal with it in a similar way to the analogous term in the current norm.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 39 / 42

slide-48
SLIDE 48

Dealing With the Commutator Terms: φ

The energy estimates for derivatives of φ can be solved using a similar

  • method. It suffices to show that for |I| ≤ k we have the estimate
  • τ s

+τ− 1/2C g (DI Xφ)w1/2 ι

  • L2(t,x) Ek(T).

first using the identity [C

g , DY ]φ = −Dαφ∇α(∇·Y )+Dα

  • (Y )παβDβφ
  • −i(∇αFY αφ+2FY αDαφ).

We iterate this to establish the full commutator [C

g , DI Y ]φ. Again, we

look by this term-by-term. The first term is again an error term which scales with the metric, and we deal with it in a similar way to the analogous term in the current norm. For the second term, Dα (Y )παβDβφ

  • , we take advantage of the fact that

this is approximately equal to cY C

g φ. We must be careful when bounding

error quantities like ((Y )παβ − cZgαβ)DαDβφ.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 39 / 42

slide-49
SLIDE 49

Dealing With the Commutator Terms: φ (Cont.)

The third term requires great care even in the Minkowski metric. We first decompose (∇αFY α)φ+2FY αDαφ = JY φ+2FY α Dα(r∗φ) r∗ +

  • ∇αY β − 2Y β∇α(r∗)
  • Fβαφ.

The analysis of the first two terms is straightforward, where we take advantage of the L2(t)L∞(x) estimate for the nice components of F, D(r∗φ) and match it with our L∞(t)L2(x) on our bad components. For the third term, we must establish the bound

  • ∇αY β − 2Y β∇α(r∗)
  • Fβα α[F] + ρ[F] + σ[F] + τ −s

0 α[F].

This nice cancellation in the bad components was noted in (Shu 1991).

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 40 / 42

slide-50
SLIDE 50

Wrapping it All Up

We can now combine our energy results. In particular, for sufficiently small Ek(0), it follows that Ek(T) Ek(0) + ǫk[g]2. (11) The main theorem follows. We note that the initial energy is equivalent to selecting initial conditions ˙ φ, Dφ0, B0, E0, respectively the initial time, and bounding certain Sobolev norms on the first three quantities as well as the divergence-free part of E0.

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 41 / 42

slide-51
SLIDE 51

Conclusion

Thanks for listening!

Christopher Kauffman Johns Hopkins University Global Stability for Charged Scalar Fields in Spacetimes close to Minkowski 42 / 42