Asymptotic Flatness at Spatial Infinity  

In General > s.a. [solutions of general relativity]; ADM; canonical general relativity; initial value formulation; multipole moments.
* Idea: One gets information on conserved quantities of spacetime, but there are no equations, the dynamics is not recorded.
* History: It took about ten years more to develop than the structure at , mainly because of the intricate differentiable structure.
@ Early work: in Lichnerowicz 39 [for stationary spacetimes].
@ General: Geroch JMP(72), in(77); Sommers JMP(78); Ashtekar in(80), FP(85); Beig & Schmidt CMP(82); Ashtekar & Magnon JMP(84); Beig PRS(84); Winicour FP(85); Chrusciel JMP(89), JMP(89); Petrov IJMPD(95), IJMPD(97); Hayward PRD(03)gq.
@ With symmetries: Beig GRG(80) [static]; Beig & Simon GRG(80), Kennefick & Ó Murchadha CQG(95)gq/93 [stationary]; Beig & Chrusciel JMP(96)gq/95; > s.a. initial value formulation.
@ Existence: Reula CMP(89); Lindblad & Rodnianski CMP(05)m.AP/03 [wave coordinates].
@ Symmetries, conservation laws: Goldberg PRD(90); Perng JMP(99)gq/98; Lusanna & de Pietri gq/99; Szabados CQG(03)gq.
@ Other: Friedrich CMP(88) [radiativity condition]; Ashtekar & Romano CQG(92) [i0 as boundary]; Thiemann CQG(95)gq/93 [Ashtekar variables]; Herberthson CQG(98)gq/97 [diff]; Husa gq/01-in [numerical]; Finster & Kraus CJM(05)m.DG/03 [curvature estimates]; Shiromizu & Tomizawa PRD(04)gq [D > 4]; Zenginoglu gq/06-in [numerical].

Spi Formalism > s.a. Penrose Diagram; Ripple.
$ Def: A spacetime (M, g) is said to be asymptotically flat at spatial infinity if there exists a spacetime (M', g'), which is Cinfty everywhere except at a point i0, where M' is C>1 and g' is C>0, together with an imbedding i: MM, such that
(1) J-bar(i0) = M' \ M, i.e., i0 is at spatial infinity;
(2) : M' → R, C2 at i0, Cinfty elsewhere, g'ab|M = 2 gab, |i^0 = 'a |i^0 = 0,'a'b |i^0 = 2 g'ab ( –2);
(3) Rab admits a regular direction-dependent limit at i0 (matter sources fall off like –4).
* Re boundary conditions: If the metric were C1 at i0, the mass would vanish; If the metric were C0, the mass would not be defined; R'abcd blows up at i0, but 1/2 R'abcdRabcd(), a regular direction-dependent limit; Its Weyl part is coded in the electric and magnetic parts, with potentials provided by the Ricci part.
* Relationships: Implies asymptotic flatness at spatial infinity in the ADM sense.
* Hyperboloid : The most useful construction at spi is the hyperboloid of unit timelike vectors, the "space of directions of approach to i0"; It has a natural metric hab = gabab.
* First order structure: Universal, the C>1 manifold with well-defined tangent space and metric.
* Connections: They correspond to equivalence classes of connections on spacetime; They need not always be the same.
* Higher order structure: Not meaningful, reflected in the fact that the curvature blows up at spi.
@ References: Bergmann & Smith PRD(93) [structure].

Ambiguities > s.a. asymptotic flatness.
* Non-uniqueness: Equivalent conformal completions may be obtained by supertranslations, inequivalent ones by a 4-parameter family of logarithmic transformations; If we use ' = , where = 1 and C>0 at i0, i.e., = 1 + 1/2 , where ...; If is C0 at i0, we get an inequivalent completion (e.g., related by a log translation).
* Logarithmic transformations: An ambiguity in the choice of flat metric (in addition to supertranslations); If is one such metric, with Cartesian chart xm, then ' with x'm:= xm + Cm ln , for all Cm, will also do; In some cases there is a preferred or asymptotic frame; In the spi framework, a 4-parameter family of inequivalent, logarithmically related completions which give the same physical answers, and can be considered as gauge.
@ Logarithmic transformations: Bergmann PR(61); Beig & Schmidt CMP(82); Ashtekar FP(85).

Symmetries: The Spi Group
* Idea: The set of all diffeomorphisms that leave the spi structure invariant, modulo those which generate the identity at i0 and leave each ripple fixed, = /I.
* Structure: Similar to the BMS group, a semidirect product of the Lorentz group and the supertranslation group; It has a preferred translation subgroup, but not a preferred Lorentz (and hence Poincaré) subgroup.
* Generators: Vector fields Xa such that at i0, Xa C>0 and X a = 0 (not to move i0), '(aXb) = 0 (not to change the metric, asymptotic Killing vector fields), and 'a '(b Xc) = 'a g'bc, for some (= 1 at i0) (so X can be associated with some change in conformal factor); If a = 0, then X generates the identity at i0; It belongs to I.
* Remark: X defines a vector field on by X' a:= lim –1/2 X a (tangential to since (a Xb) = 0).

Energy-Momentum > s.a. ADM formalism; energy in general relativity [and generalization].
$ Def: The energy-momentum component along a tangent vector Va at i0 is

Va pa = (1/8G) C Eab Vb dSa ,

where Eab is the electric part of the Weyl tensor on the hyperboloid at i0, and C is any cross section of .
@ Relationship with ADM: Sniatycki RPMP(89); Huang & Zhang gq/05, gq/06.


Main pageAbbreviationsJournalsCommentsOther sitesAcknowledgements
Send feedback and suggestions to bombelli at olemiss.edu – Modified 22 jun 2008