general relativity - universidad de...
TRANSCRIPT
General RelativityIV Escuela Mexicana
de Cuerdas y Supersimetría
Gustavo NizJune 2015
General Relativity
Based on:
●Sean Carrol's notes (short ones)● John Stewart course/book●Personal notes/talks
GR as a Cooking Book
● Ingredients– Physics & Maths
● Cookware– Equations
● Recipes– Solutions
If time allows it … ”BEYOND GR”
Ingredients
● Why GR?
Newtonian gravity
Vs
Einstein's gravity
Particle moves in straight line until hit by a force
Gravity is not a force but a result of space-time geometry
Ingredients
● Why GR?
Newtonian gravity
Vs
Einstein's gravity(GR)
Particle moves in straight line until hit by a force
Gravity is not a force but a result of space-time geometry
Keywords
Ingredients
Need to understand
a) Space and time together
b) Curved space
Ingredients
Need to understand
a) Space and time together
b) Curved space
TensorNotation
Ingredients: Special Relativity
● Speed of light is constant in any frame (c=1)● Space and time are entangled
Coordinates of space-time
Ingredients: Special Relativity
Special Relativity (SR) lives in
Minkowski space-time
4-vector
Event P
Ingredients: Special Relativity
Minkowski metric
Distance (dot product)
Infinitesimal distance (line element)
Summation convention
Same information as in the metric
Ingredients: Special Relativity
Lorentzian signature ( - , + , + , + )
[ sometimes (+,-,-,-) ]
C.f. Euclidean 4d space
Euclidean signature ( + , + , + , + )
Ingredients: Special Relativity
Lorentz group
– Transformations that leave ds unchanged
e.g. x-directionboost
Exercise
Ingredients: Special Relativity
For a particle moving only
in time (xi=const.) the elapsed time is
Define proper time as
Ingredients: Special Relativity
Spacetime diagram
Ingredients: Special Relativity
Spacetime diagram
Null vector
Defines light-cone
Ingredients: Special Relativity
Spacetime diagram
Null vector
Timelike vector
Spacelike vector
Ingredients: Special Relativity
Trajectory specified by
is timelike/spacelike/null
if tangent vector
is timelike/spacelike/null
Ingredients: Special Relativity
● Null trajectories represent light or massless particles
● Spacelike trajectories are causally disconnected points (or particles with v>1)
● Timelike trajectories describe massive particles (observers)
More convenient to
use proper time
Ingredients: Special Relativity
Massive particles
4-velocity
4-momentum
m rest frame mass
Ingredients: Special Relativity
Massive particles
4-velocity 4-momentum
Energy is , which in the rest frame, is
Ingredients: Special Relativity
Massive particles
One can boost the frame with a Lorentz transf., say in the x-dir.
Which for small v reduces to
Rest mass + kinetic energy
Newtonian momentum
Ingredients: Curved Space
Minkoswki Manifold
Cartesian coords. Cannot use Cartesian
coords. globally
Ingredients: Curved Space
Def. A topological space M is an n-dimensional
manifold if there is a collection (atlas, )
of maps (charts, ) , such that each map
is continuous, bijective and
invertible
Manifold
Ingredients: Curved Space
In simple words
Ingredients: Curved Space
Equivalence Principle.
A choice of chart (coordinates) is arbitrary and the physics should not depend on this!
A convenient way to describe chart-independent equations is to use tensors, which can be thought of generalisations of vectors, with possibly more indices
Ingredients: Tensors
Under a coordinate transformation a
tensor changes in the following way
(summation convention)
Ingredients: Tensors
Under a coordinate transformation a
tensor changes in the following way
(n,m) - Tensor
(summation convention)
Ingredients: Tensors
Observations
(0,0) – tensor scalar (function)
(0,1) – tensor vector
(1,0) – tensor co-vector
Defin
ition
s
Ingredients: Tensors
Observations
Trace over indices
Symmetric
Antisymmetric
Mat
hs
Ingredients: Tensors
Observations
An equation that holds in one
coordinate system holds in all
coordinate systems
Equivalence Principle!
Physi
cs
Ingredients: Tensors
In GR the most important tensor is the metric
a generalisation of Minkowski's metric
which encodes the geometrical information of
spacetime
Ingredients: Tensors
At a given point, there is always a coordinate system such that
and also first derivatives of vanish!
Second derivatives cannot be made to vanish,
a manifestation of curvature
i.e. the spacetime
looks locally flat
Ingredients: Tensors
Dot product
Suggests the concept of raising/lowering indices, namely
where is the inverse of
so
Ingredients: Tensors
Important objects in physics which are not tensors
1) determinants
Under transforms as
Ingredients: Tensors
Important objects in physics which are not tensors
2) volume factor
Transforms as
1) and 2) are call tensor densities because transform as some powers of the Jacobian
Ingredients: Tensors
Important objects in physics which are not tensors
Notice
Is invariant under a change of coordinate, since f(x) is a scalar
Ingredients: Tensors
Important objects in physics which are not tensors
Notice
Is invariant under a change of coordinate, since f(x) is a scalar
Therefore the right way of writing integrals is using
Ingredients: Tensors
Important objects in physics which are not tensors
3) Partial derivatives
On scalar are OK
...but on vectors,
Ingredients: Tensors
Important objects in physics which are not tensors
3) Partial derivatives
On scalar are OK
...but on vectors,
Ingredients: Tensors
Define the Covariant Derivative
Connection transforms to cancel this
Ingredients: Tensors
transforms a tensor and
defines parallel transport
Ingredients: Tensors
transforms a tensor and
defines parallel transport
Similarly for lower indices
Ingredients: Tensors
Christoffel Symbols
There is a particular choice of connection that
In components, it is
Ingredients: Curvature
Information about the curvature is contained in the metric tensor, and it is the Riemann tensor which explicitly accounts for it
(note: it has second derivatives of the metric)
Ingredients: Curvature
Geometrical interpretation
In Euclidean space if a vector is parallel transported around a closed loop, it returns unchanged.
In curved space, this is not necessarily true.
The Riemann tensormeasures the
difference!
Ingredients: Curvature
Properties of
1)
2)
(Bianchi identity)
3)
Antisymmetric
Symmetric
Ingredients: Curvature
Other tensors derived from
1) Ricci tensor
It satisfies
2) Ricci scalar
Ingredients: Curvature
Other tensors derived from
3) Einstein tensor
It obeys
Cookware: Geodesic Equation
A geodesic is a curve that extremises
The path is a geodesic if it obeys
where is the affine parameter
Cookware: Geodesic Equation
In GR test particles always move along geodesics.
– Massless particles follow null geodesics
(null trajectory)– Massive bodies follow timelike geodesics
(timelike trajectory)
Objects move along geodesics until a force knocks them off!
Cookware: Einstein Equations
Einstein (1915)
Cookware: Einstein Equations
Non-linear equations of the metric components!
EnergyMomentum
tensor
Cookware: Einstein Equations
Dimension-full parameters
Newton's Constant
If c=h=1 then
Cookware: Einstein Equations
Geometría del Materia/Energía
Espacio-tiempo
Cookware: Einstein Equations
El espacio-tiempo es como una “manta” invisible
deformado por la materia o energía
Cookware: Einstein Equations
Cauchy Problem. e.g. Maxwell eqns.
No t derivatives Constraints
1st order t der. Evolution eqns.
Give suitable initial data so that the system has only one solution. We need E(0,x), B(0,x) which satisfy the constraints, then E(t,x), B(t,x) are obtained from the evolution equations
Cookware: Einstein Equations
Cauchy Problem. Similar for RG
1) No t derivatives Constraints
2) 1st order t der. Constraints
3) 2nd order t der. Evolution eqns.
Find for , which satisfy the constraints initially, and use the evolution equations to solve for
The equations are linear in second derivatives!
Cookware: Einstein Equations
Relevant components of the metric
has 10 independent components (remember it is symmetric), but there are 6 evolution equations to determine them (?).
Actually, we have an arbitrary choice of coordinates. Therefore, there are only 10-4=6 variables to determine, which can be found using the evolution equations.
Cookware: Einstein Equations
Energy momentum tensor
A popular choice is a perfect fluid
– Fluid with no viscosity, or heat flow, and isotropic in its rest frame
– Completely specified by energy density and pressure
where is the unitary 4-velocity
– Bianchi identity implies conservation
Cookware: Einstein Equations
Newtonian limit
Recall Poission equation for the Newtonian potential
2nd order diff.operator over
a field
Cookware: Einstein Equations
Newtonian limit
Recall Poission equation for the Newtonian potential
2nd order diff.operator over
a field
In a relativistic theorywe expect the energy
momentum tensor
Cookware: Einstein Equations
Newtonian limit
Recall Poission equation for the Newtonian potential
2nd order diff.operator over
a field
In a relativistic theorywe expect the energy
momentum tensor
Einstein equations
Cookware: Einstein Equations
Newtonian limit
Now formally consider (perturbation around Minkowski)
Geodesic Equation + slow motion
Newton's 2nd law
Cookware: Einstein Equations
Newtonian limit
Now formally consider (perturbation around Minkowski)
Geodesic Equation + slow motion
00-component of Einstein eqns Poisson's eq.
Newton's 2nd law
Cookware: Hilbert-Einstein action
Einstein equations can be obtained from
Einstein equations
Cookware: Hilbert-Einstein action
Can include a Cosmological Constant
Cookware: Alternative approaches
1) á la Palatini
Consider the connection and the metric as independent
The metricconnection
Cookware: Alternative approaches
2) Hamiltonian (ADM or 3+1 formalism)
Consider the space-time splitting
N lapse function
Ni shift function
gij spatial metric
Schematically
Cookware: Alternative approaches
2) Hamiltonian (ADM or 3+1 formalism)
Extrinsic curvature
Then one gets the Hamiltonian Canonical momenta
Hamiltonian
Momentum
Cookware: Alternative approaches
2) Hamiltonian (ADM or 3+1 formalism)
Observations
● Only is dynamical
● are Lagrange multipliers which lead to (in vacuum)
Dynamical D.O.F.
6 – (1+3) = 2 two polarization modes
of the graviton
Recipes: content
● Maximally symmetric solutions
– Penrose Diagrams● Black Hole solutions
– Schwarzschild– Others– Thermodynamics
● Cosmology. In other course(s)!!
Recipes: Maximally symmetric solns
Consider vacuum with a cosmological constant
There are 3 maximally symmetric solutions
– Minkowski
– De Sitter (dS)
– Anti-de Sitter (AdS)
Recipes: Penrose diagrams
● Causal and topological structure
● Infinite is brought to a finite distance
● Light rays propagate at 45°
● Conformal factors do not change
the lightcone structure, so can be dropped
● Useful to study black holes
Recipes: Penrose diagrams
Consider Minkowski spacetime
Change variables to
Recipes: Penrose diagrams
Bring infinite to fine distance with
Recipes: Penrose diagrams
Back to the Cartesian coords.
Conformal metric
Recipes: Penrose diagrams
Half of3-sphere
Recipes: Penrose diagrams
Take half of triangle
and attacha full
3-sphere ateach point
Recipes: Penrose diagrams
Future timelike infinity
Future null infinity
Spatial infinity
Past null infinity
Past timelike infinity
Recipes: de Sitter (dS)
Easier to picture in 5d
Recipes: de Sitter (dS)
By choosing one can write the metric in the following useful frames:
Static coordinates
Cosmological coordinates
Exponential expansion
Recipes: de Sitter (dS)
Penrose diagram
Both timelike and null future infinity
Recipes: de Sitter (dS)
Penrose diagram
Recipes (detour): Horizons
Particle horizon
Event horizon
Recipes: Anti de Sitter (AdS)
Maximal symmetry
5d picture
Two times
Recipes: Anti de Sitter (AdS)
Closed Timelike Curves (travel back in time?)
On the Universal covering space we avoid them!
Recipes: Anti de Sitter (AdS)
By choosing one can write the metric as
Static coordinates
Penrose diagram: cannot be drawn well...
Relevant for AdS/CFT
Recipes: Solutions so far...
Solution Symmetry
Minkowski 0 0 Maximal
de Sitter >0 0 Maximal
Anti de Sitter <0 0 Maximal
Schwarzschild 0 Spherical
Recipes: Schwarzschild solution
Birkhoff's theorem
If one assumes spherical symmetry in GR in vacuum
(using the gauge freedom can fix ), then
● The solution is unique
● Static (i.e. do not depend on time)
● Characterised by one integration constant alone (M)
It is called the Schwarzschild solution
Recipes: Schwarzschild solution
Schwarzschild solution is
Since it is unique, it describes from
black holes to planets!
BHor
planetVacuum Schwarzschild sol.
Mass inside Ro
Ro
Recipes: Schwarzschild solution
Test particle in the plane
For the Schwarzschild metric the potential is
Particle of unitmass and energyE in 1d potential
Kinetic potential energy conservation
Const . Newtonian angular potential momentum
Only GRcorrection(Mercury)
Recipes: Schwarzschild solution
Re-write V(r) as
Particle cannot scape the gravitational potential unless is traveling faster than light
For a star one hits the surface before
A black hole is such that matter has collapse inside its Schwarzschild radius,
Recipes: Schwarzschild solution
Two singularities
1) True singularity diverge
2) Coordinate singularity bad frame
If
Nothing happens at
Recipes: Schwarzschild solution
However, something special still happens at
Event horizon
Sees (1) moving slowerand slower, taking infinitetime to reach
Recipes: Schwarzschild solution
However, something special still happens at
Event horizon
Sees (1) moving slowerand slower, taking infinitetime to reach
(1) takes a finite timeto cross then r and t swap
roles and (1) goes straight into r=0
Recipes: Schwarzschild solution
However, something special still happens at
Recipes: Schwarzschild solution
Penrose diagram
Recipes: Schwarzschild solution
Penrose diagram (completion of spacetime)
Recipes: Kerr BH
Add rotation
Recipes: Kerr BH
Momentum bound
Extremal case
(BPS, stable, in nature?)
Recipes: Reissner–Nordström BH
Electric charge
Extremal case when
Charged+Rotating Kerr-Newman solution
Recipes: Kerr, KN or RN
Recipes: Kerr, KN or RN
Recipes: Kerr, KN or RN
Black holes have no hair
If matter collapses into a black hole, then all information would be lost, since the later is described by its Mass, Charge and Angular Momentum only.
Moreover, due to quantum processes, it emits isotropic (Hawking) radiation in a rate proportional to its mass (M). Therefore, it looses energy and evaporates in a time of the order
The associated radiation temperature is related to its surface gravity (acceleration of observer on its horizon).
Recipes: BH thermodynamics
Thermodynamics Black Holes
0th – law Constant T Constant surface
in thermal eq. gravity (at horizon)
1st – law
2nd – law
3rd – law No system with S=0 No BH with
in finite series of steps
Recipes: BH thermodynamics
Implication
Temperature Surface gravity
Entropy Area of horizon
Beyond GR ...
● Desser's construction. Bottom up!
● Lovelock's theorem
● Problems with GR
● UV modifications
● IR modifications
Bottom up approach
Is GR unique?
(Kraichnan, Feynman, Weinberg, Boulware, Deser, etc.)
● Electromagnetism
– The vector field (spin 1) couples to conserved current, but it does not contribute to it (photons do not carry charge!)
● Gravity
– The graviton field (spin 2) couples to conserved current (energy-momentum tensor), and it does contribute to it (c.f. Yang-Mills)
Bottom up approach
Basic idea
● To recover Einstein's theory from linear theory of gravitons coupled to the gravitational energy momentum tensor
● Assumptions:
– Symmetric stress-energy tensor
– Second order derivatives
– Total derivatives (boundary terms) irrelevant
– Key one: local Lorentz symmetry● Infinite series in metric field, but finite in the right variables
Deser, gr-qc/0411023, 0910.2975
Bottom up approach
Linear Gravity(Written in )
à la Palatini
Generates its own
Energy-MomentumTensor
(Belifante)
FULLGR
EOMWritten in
Are the same!!!
Bottom up approach
Start from (Fierz Pauli) action
with independent Minkowski tensors
Want to recover
Totalderivative
Bottom up approach
Start from (Fierz Pauli) action
with independent Minkowski tensors
Want to recover
Totalderivative
Full GR only needs an extra cubic term!!!
Bottom up approach
Add coupling
With Belifante-Rosenfield Procedure (1940)
● Covariantise linear theory w.r.t. fictitious metric and derive associated energy momentum tensor. Set at the end.
Defines newaction
Bottom up approach
Finally, one can check the EOM from
are
LL L
Theorems
Lovelock's Theorem (1971-72)
If we want second order differential field equations for the metric with a 4d space-time, the only form is
To recover Newtonian physics
Theorems
Lovelock's Theorem
Implications:
1) Terms in the Lagrangian such as
which give second order field equations in 5d, are topological in 4d.
Theorems
Lovelock's Theorem
2) To modify Einstein's theory we need
● Consider more fields (beyond the spin-2)
● Accept higher order derivatives in field equations
● Higher dimensions
● Give up rank (0,2) field eqns, symmetry of the indices in eqns, or divergence free equations
● Give up locality
Why do we need to modify GR
● Non-renormalisable (perturbatively)● Essential singularities
– Black holes
– Big bang singularity
● Early Universe physics– Hot big bang problems (horizon, Universe size,
monopoles/defects, flatness)
in the UV ?
Why do we want to modify it
Cosmology's dark sector– Dark matter (particle physics?)
– Dark energy ● Small observed value● Why today (coincidence problem)● Vacuum energy (“old cosmological constant problem”)
in the IR ?
Why do we want to modify it
Cosmology's dark sector– Dark matter (particle physics?)
– Dark energy ● Small observed value● Why today (coincidence problem)● Vacuum energy (“old cosmological constant problem”)
in the IR ?
Non-renormalisable (perturbatively) in the UV
IRideas
Dark Sector in Cosmology
Observational evidence of dark energy
Simple solution: Λ (Vacuum energy?)
Complicated solution:New gravitational degrees of freedom!
Theoryconsistency ObservationsBeauty
New physics?
Dark Energy
Cosmological Constant (CC) & vacuum energy
Vacuum energy gravitates as a CC
QFT
Lorentz invariance
Energy-Momentum tensor
Cosmological Constant (CC) & vacuum energy
To get one needs
an important cancellation
The effective CC is
QFT
Dark Energy
Cosmological Constant (CC) & vacuum energy
Estimate perturbatively
By adding the normal modes of a free scalar with mass m we can see the problem
QFT
Dark Energy
Cosmological Constant (CC) & vacuum energy
Estimate perturbatively
By adding the normal modes of a free scalar with mass m, we can see the problem
QFT
Dark Energy
Cosmological Constant (CC) & vacuum energy
For some many species of matter
Fermionicnumber
E.g. electrons
QFT
Dark Energy
Cosmological Constant (CC) & vacuum energy
One of the most important problems in theoretical physics
Dark Energy
Possible ways outCancellation very specific among contributions (eg. SUSY)
Relaxation of its value as Universe expands is impossible
(“no-go” theorem by Weinberg, 1989)
Dark Energy
Possible ways outCancellation very specific among contributions (eg. SUSY)
Relaxation of its value as Universe expands is impossible
(“no-go” theorem by Weinberg, 1989)
Dark Energy
Modified
?
Baker et al, 2014
Where canbe modified?
Eot-Wash
Baker et al, 2014
Baker et al, 2014
Baker et al, 2014, Koyama's talk 2015
Screening mechanisms
EOM for scalar d.o.f.
Kineticterm
Potential(mass)
MatterInteraction
Only consider a coupling to non-relativistic matter
Screening mechanisms
Chameleon
Screening mechanisms
Symmetron
Screening mechanisms
Vainshtein effect (Galileon)
Screening mechanisms
Horndeski Lagrangian
Horndeski (70's)
Most general scalar-tensor theory with second order EOMS
Fab Four
Charmousis et al, 2011
Sector of Horneski's Lagrangian which
1) Minkowski is a vacuum solution independently of the bare cosmological constant (c.f. De Sitter in GR)
2) Solution remains flat after phase transition (change in the bare C.C.)
3) Composed by four different terms (Paul, Ringo, John, George)
UVideas
UV: Non-renormalisable
Where is the problem?
Consider
Re-scaling
Lorentz invariance
UV: Non-renormalisable
Potential
Diverges in the UV if n>4
UV: Non-renormalisable
Dimensional analysis
Propagator
UV: Non-renormalisable
In General Relativity the linear theory is
Propagator (schematically)
Differential operator of 2nd order
Dimensionfulcoupling
UV: Non-renormalisable
● We need an infinite number of counter-terms
● Due to Lorentz symmetry the have to be powers of curvature tensors
UV: Non-renormalisable
So what's wrong with ?
Propagator (schematically)
...
UV: Non-renormalisable
But
Propagator
Masslessgraviton
Massiveghost!!
UV: Non-renormalisable
But
Propagator
Masslessgraviton
Massiveghost!!
Usual problem withhigher derivatives
(Ostrogradski)
UV: Non-renormalisable
Possible ways out
1)SUSY SUGRA (helps but not a solution)
2)Lorentz
● Horava-Liftshitz gravity
3)Non-perturbative methods
● Asymptotic safety
4)Path integral
● Wave function of the Universe● Spin-foams
UV: Non-renormalisable
Possible ways out
4)Canonical quantization (background independent)
● Wheeler-deWitt● Loop quantum gravity
5)New physics/mathematics
● Algebraic structures (Kac-Moody)● BCJ amplitudes?
● String Theory● Etc.