Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology Torsten Asselmeyer-Maluga 2 GermanAerospaceCenter,Berlinand 1 LoyolaUniversity,NewOrleans,LA,USA 0 2 E-mail: [email protected] n a Jerzy Kr´ol J UniversityofSilesia,InstituteofPhysics,ul. Uniwesytecka 4,40-007Katowice, 8 Poland 1 E-mail: [email protected] ] c q Abstract. Sincethefirstworkonexoticsmoothness inphysics,itwasfolklore - to assume a direct influence of exotic smoothness to quantum gravity. In the r second paper, we calculate the “smoothness structure” part of the path integral g in quantum gravity for the exotic R4 as non-compact manifold. We discuss the [ influence of the “sum over geometries to the “sum over smoothness structure. There are two types of exotic R4: large (no smooth embedded 3-sphere) and 2 small(smooth embedded 3-sphere). Alargeexotic R4 canbeproduced byusing v topologically slice but smoothly non-slice knots whereas a small exotic R4 is 2 constructed by a 5-dimensional h-cobordism between compact 4-manifolds. The 8 resultsareappliedtothecalculationofexpectationvalues,i.e. wediscussthetwo 8 observables,volumeandWilsonloop. Thentheappearanceofnakedsingularities 4 is analyzed. By using Mostow rigidity, we obtain a justification of area and 2. volume quantization again. Finally exotic smoothness of the R4 produces in all 1 cases(smallorlarge)acosmological constant. 1 1 : v PACSnumbers: 04.60.Gw,02.40.Ma,04.60.Rt i X r a Submitted to: Class. Quantum Grav. Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology2 1. Introduction Since the first papers about exotic smoothness it was folklore to state an influence of exoticsmoothnessonthestatesum(orpathintegral)forquantumgravity. Inourfirst paper[Asselmeyer-Maluga(2010)]wecalculatedthe “exoticsmoothness”contribution tothepathintegralforaspecialclassofcompact4-manifoldsincludingtheK3surface. Theexoticsmoothnessstructurewasconstructedbyknotsurgery. Similarresultswere obtained by Duston [Duston(2009)] for branched covers. The calculation of the path integral has to formally include the exotic smoothness [Pfeiffer(2004)] to relate it to smooth invariants of 4-manifolds. We demonstrated it in our previous paper (to get the Chern-Simons invariant). Unfortunately the most interesting and physically important case of an exotic R4 is also the most complicated one which strongly relies oninfiniteconstructions(Cassonhandlesetc.). Theappearanceoftwoclassesofexotic R4, large and small, complicates the situation. These two classes have their origin in the two mainfailures in 4-dimensionaldifferentialtopologywhich staysin contrastto the topological theory: the smooth h-cobordism theorem and the large class of non- smoothable, topological 4-manifolds. If there is a smooth embedding of a 3-sphere into the exoticR4 then one callsit asmall exotic R4 andifnot itis a large exoticR4. In this paper we will study the effect of the exotic R4 on the functional integral for the Einstein-Hilbert action. In the next section we present some of the physical assumptions as well the definition of an exotic R4. Then we discuss the existence of a Lorentz metric for an exotic R4. Formally the existence of a Lorentz metric is a purely topologicalquestion which can be answered positively. On the other hand, global hyperbolicity and all its consequences depend on the standard smoothness of R4 = R3 R (as smooth product). Therefore an exotic R4 must contain naked singularities.×We will further analyze these singularities in section 8 to obtain a pairwise structure by the failure of the Whitney’s trick. The main part of this paper is formed by the sections 4 to 6. It starts with a description of large and small exotic R4. Then we discuss the splittingoftheactionfunctional(usingthediffeomorphisminvarianceoftheEinstein- Hilbert action) according to these descriptions. Finally we calculate the functional integral at first for a particular exotic R4 and then for the whole continuous (radial) family. The discussion of observables, like volume and the Wilson loop, in section 7 completesthe picture. Especiallyweconfirmthe resultsofLoopquantumgravity,i.e. the quantization of area and volume. All exotic R4 have one common property: the appearance of a cosmologicalconstant. 2. Physical Motivation and model assumptions Einsteins insight that gravity is the manifestation of geometry leads to a new view on the structure of spacetime. From the mathematical point of view, spacetime is a smooth 4-manifold endowed with a (smooth) metric as basic variable for general relativity. Later on, the existence question for Lorentz structure and causality problems (see [Hawking and Ellis(1994)]) gave further restrictions on the 4-manifold: causalityimplies non-compactness,Lorentzstructureneedsa codimension-1foliation. Usually,onestartswithagloballyfoliated,non-compact4-manifoldΣ Rfulfillingall × restrictions where Σ is a smooth 3-manifold representing the spatial part. But other non-compact4-manifoldsarealsopossible,i.e. itisenoughto assumeanon-compact, smooth 4-manifold endowed with a codimension-1 foliation. Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology3 All these restrictions on the representationof spacetime by the manifold concept are clearly motivated by physical questions. Among the properties there is one distinguished element: the smoothness. Usually one assumes a smooth, unique atlas ofchartscoveringthemanifoldwherethesmoothnessisinducedbytheuniquesmooth structure on R. But as discussed in the introduction, that is not the full story. Even in dimension 4, there are an infinity of possible other smoothness structures (i.e. a smooth atlas) non-diffeomorphic to each other. In the following we will specialize to the R4: Definition 1 The smoothness structure of R4 is called an exotic smoothness structureorexoticR4 ifitisnon-diffeomorphictothestandardsmoothnessstructure (induced from the smooth product R R R R). × × × The implications for physics seem to be obvious because we rely on the smooth calculus to formulate equations of any field theory. Thus different smoothness structurescouldrepresentdifferentphysicalsituationsleadingtodifferentmeasurable results. But it should be stressed that exotic smoothness is not exotic physics! Exotic smoothness is a mathematical possibility which should be further explored to understand its physical relevance. 3. Lorentz metric and global hyperbolicity Before we start with the construction of the various exotic R4’s (large and small), we will discuss some physical implications which are independent of these constructions. Firstly we consider the existence of a Lorentz metric, i.e. a 4-manifold M (the spacetime) admits a Lorentz metric if (and only if) there is a non-vanishing vector field. In case of a compact 4-manifold M we can use the Poincare-Hopf theorem to state: a compact 4-manifold admits a Lorentz metric if the Euler characteristic vanishes χ(M) = 0. But in a compact 4-manifold there are closed time-like curves (CTC) contradicting the causality or more exactly: the chronology violating set of a compact 4-manifold is non-empty (Proposition 6.4.2 in [Hawking and Ellis(1994)]). Non-compact 4-manifold M admits always a Lorentz metric and a special class of these 4-manifolds have an empty chronology violating set. If is an acausal S hypersurface in M (i.e., a topological hypersurface of M such that no pair of points ofM canbe connectedby meansofacausalcurve),thenD+( ) isthe future Cauchy S development (or domain of dependence) of , i.e. the set of all points p of M such S that any past-inextensible causal curve through p intersects . Similarly D−( ) is S S the past Cauchy development of . If there are no closed causal curves, then is S S a Cauchy surface if D+( ) D−( ) = M. But then M is diffeomorphic to R [Bernal and Sa´ncheSz(2∪00S3)∪]. TheSexistence of a Cauchy surface implies global S × hyperbolicity, i.e. a spacetime manifold M without boundary is said to be globally hyperbolic if the following two conditions hold: (i) Absence of naked singularities: For every pair of points p and q in M, the space of all points that can be both reached from p along a past-oriented causal curve and reached from q along a future-oriented causal curve is compact. (ii) Chronology: No closed causal curves exist (or ”Causality” holds on M). Usually condition 2 above is replaced by the more technical condition ”Strong causalityholdsonM”butasshownin[Bernal and Sa´anchez(2007)]insteadof”strong Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology4 causality”, one can write simply the condition ”causality” (and strong causality will hold under causality plus condition 1 above). Then all (non-compact) 4-manifolds R are the only 4-manifolds which admit S× a globally hyperbolic Lorentz metric, where the product has to be a smooth × product not only by physical reasons but also because of the claimed result in [Bernal and Sa´nchez(2003)]. But more is true [Bernal and Sa´anchez(2005)], Theorem 1 If a spacetime (M,g) is globally hyperbolic, then it is isometric to (R , f dτ2 + g ) with a smooth positive function f : R R and a smooth τ × S − · → family of Riemannian metrics g on varying with τ. Moreover, each t is a τ S { }×S Cauchy slice. Furthermore in [Bernal and Sa´nchez(2006)] it was shown: If a compact spacelike submanifold with boundary of a globally hyperbolic • spacetime is acausal then it can be extended to a full Cauchy spacelike hypersurface of M, and S for any Cauchy spacelike hypersurface there exists a function as in Th. 1 such • S that is one of the levels τ =constant. S So, what about exotic 4-manifolds? At first the existence of the Lorentz metric is a purely topological condition which will be fulfilled by all non-compact 4-manifolds independentofthesmoothnessstructure. Byconsideringtheglobalhyperbolicity,the picture changes. An exotic spacetime homeomorphicto R is not diffeomorphic to R. The Cauchy surface is a 3-manifold with uniqSue×smoothness structure (up tSo×diffeomorphisms), the stanSdard structure. So, the smooth product R has also the standardsmoothness structure. But the diffeomorphism to R iSs n×ecessaryfor globalhyperbolicity. Thereforeanexotic R is never globallyShy×perbolic but admits a Lorentz metric. Generally we have an eSxo×tic R with a Lorentz metric suchthat the projection R R is a time-function (Sth×at is, a ontinuous function which is strictly increasiSng×on →future directed causal curves). But then the exotic R has S × no closed causal curves and must contain naked singularities . We will later see the ‡ source of these singularities. 4. Exotic R4 InthissectionwewillgivesomeinformationabouttheconstructionofexoticR4. The existence of a smooth embedding S3 R4 in the exotic R4 splits all exotic R4 into → two classes, large (no embedding) or small. 4.1. Preliminaries: Slice and non-slice knots At first we start with some definitions from knot theory. A (smooth) knot K is a smooth embedding S1 S3. In the following we assume every knot to be smooth. Secondly we exclude wi→lderness of knots, i.e the knot is equivalent to a polygonin R3 or S3 (tame knot). Furthermore, the n-disk is denoted by Dn with ∂Dn =Sn−1. Definition 2 Smoothly Slice Knot: A knot in ∂D4 =S3 is smoothly slice if there exists a two-disk D2 smoothly embedded in D4 such that the image of ∂D2 =S1 is K. ‡ Anynon-compact manifoldM admits stably causal metrics (that is,those witha timefunction). So, ifM isnot diffeomorphic tosomeproductS×R, allthese (causally wellbehaved) metricsmust containnakedsingularities. WethankM.S´anchezfortheexplanationofthisresult. Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology5 Figure 1. Stevedore’s knot61 Figure 2. pretzel knot(−3,5,7) An example of a slice knot is the so-called Stevedore’s Knot (in Rolfson notation 6 , 1 see Fig. 1). Definition 3 Flat Topological Embedding: Let X be a topological manifold of dimension n and Y atopological manifold of dimension m where n<m. Atopological embeddingρ:X Y isflatifitextendstoatopologicalembeddingρ:X Dm−n Y. → × → Topologically Slice Knot: A knot K in ∂D4 is topologically slice if there exists a two-disk D2 flatly topologically embedded in D4 such that the image of ∂D2 is K. Here we remark that the flatness condition is essential. Any knot K S3 is the ⊂ boundary of a disc D2 embedded in D4, which can be seen by taking the cone over the knot. But the vertex of the cone is a non-flat point (the knot is crashed to a point). The difference between the smooth and the flat topological embedding is the key for the following discussion. This innocent looking difference seem to imply that both definitions are equivalent. But deep results from 4-manifold topology gave a negative answer: there are topologically slice knots which are not smoothly slice. An example is the pretzel knot ( 3,5,7) (see Fig. 2). In [Freedman(1982a)], Freedman − gave a topological criteria for topological sliceness: the Alexander polynomial (t) K △ (the best known knot invariant, see [Rolfson(1976)]) of the knot K has to be one, (t)=1. An example how to measure the smooth sliceness is given by the smooth K △ 4-genusg (K)oftheknotK,i.e. theminimalgenusofasurfaceF smoothlyembedded 4 in D4 with boundary ∂F = K the knot. This surface F is called the Seifert surface. Therefore,ifthe smooth4-genusvanishesg (K)=0thenthe knotK boundsa2-disk 4 D2 (surface of genus 0) given by the smooth embedding D2 D4 so that the image → of ∂D2 ∂D4 is the knot K. → Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology6 4.2. Large exotic R4 and non-slice knots Large exotic R4 can be constructed by using the failure to arbitrarily split of a compact, simple-connected 4-manifold. For every topological 4-manifold one knows how to split this manifold topologically into simpler pieces using the work of Freedman [Freedman(1982b)]. But as shown by Donaldson [Donaldson(1983)], some of these 4-manifolds do not exist as smooth 4-manifolds. This contradiction between the continuous and the smooth case produces the first examples of exotic R4[Gompf(1983)]. Unfortunately, the construction method is rather indirect and therefore useless for the calculation of the path integral contribution of the exotic R4. But as pointed out by Gompf (see [Gompf(1985)] or [Gompf and Stipsicz(1999)] Exercise 9.4.23 on p. 377ff and its solution on p. 522ff), large exotic R4 can be also constructed by using smoothly non-slice but topologically slice knots. Especially one obtains an explicit construction which will be used in the calculations later. Let K be a knot in ∂D4 and X the two-handlebody obtained by attaching a K two-handletoD4 alongK withframing0. Thatmeans: onehasatwo-handleD2 D2 × whichisgluedtothe0-handleD4alongitsboundaryusingamapf :∂D2 D2 ∂D4 × → so that f(., x) = K x S3 = ∂D4 for all x D2 (or the image im(f) = K D2 is the solid knotted t×orus⊂). Let ρ : X R4 b∈e a flat topological embedding (×K is K → topologically slice). For K a smoothly non-slice knot, the open 4-manifold R4 = R4 intρ(X ) X (1) \ K ∪∂XK K (cid:0) (cid:1) where intρ(X ) is the interior of ρ(X ), is homeomorphic but non-diffeomorphic to K K R4 with the standardsmoothness structure (both pieces are glued along the common boundary ∂X ). The proof of this fact (R4 is exotic) is given by contradiction, K i.e. let us assume R4 is diffeomorphic to R4. Thus, there exists a diffeomorphism R4 R4. The restriction of this diffeomorphism to X is a smooth embedding K X →֒ R4. However, such a smooth embedding exists if and only if K is smoothly K → slice (see [Gompf and Stipsicz(1999)]). But, by hypothesis, K is not smoothly slice. Thus by contradiction, there exists a no diffeomorphism R4 R4 and R4 is exotic, homeomorphicbutnotdiffeomorphictoR4. Finally,wehavet→oprovethatR4islarge. X , by construction, is compact and a smooth submanifold of R4. By hypothesis, K K isnotsmoothlysliceandthereforeX cannotsmoothlyembedinR4. Byrestriction, K D4 X and also ∂D4 = S3 can not smoothly embed and therefore R4 is a large K exot⊂ic R4. 4.3. Small exotic R4 and Casson handles Small exotic R4’s are again the result of anomalous smoothness in 4-dimensional topology but of a different kind than for large exotic R4’s. In 4-manifold topology [Freedman(1982b)], a homotopy-equivalence between two compact, closed, simply-connected 4-manifolds implies a homeomorphism between them (a so- called h cobordism). But Donaldson [Donaldson(1987)] provided the first smooth counterexample, i.e. both manifolds are generally not diffeomorphic to each other. The failure can be localized in some contractible submanifold (Akbulut cork) so that an open neighborhood of this submanifold is a small exotic R4. The whole procedure implies that this exotic R4 can be embedded in the 4-sphere S4. The idea of the construction is simply given by the fact that every such smooth h-cobordism between non-diffeomorphic 4-manifolds can be written as a product cobordism except for a compact contractible sub-h-cobordism V, the Akbulut cork. Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology7 An open subset U V homeomorphic to [0,1] R4 is the corresponding sub-h- cobordism between t⊂wo exotic R4’s. These exotic×R4’s are called ribbon R4’s. They have the important property of being diffeomorphic to open subsets of the standard R4. To be more precise, consider a pair (X ,X ) of homeomorphic, smooth, closed, + − simply-connected 4-manifolds. Theorem 2 Let W be a smooth h-cobordism between closed, simply connected 4- manifolds X and X . Then there is an open subset U W homeomorphic to − + [0,1] R4withacompactsubsetC U suchthatthepair(W C⊂,U C)isdiffeomorphic × ⊂ \ \ to a product [0,1] (X C,U X C). The subsets R =U X (homeomorphic − − ± ± to R4) are diffeom×orphic\to ope∩n subs\ets of R4. If X and X a∩re not diffeomorphic, − + then there is no smooth 4-ball in R containing the compact set Y =C R , so both ± ± ± R are exotic R4’s. ∩ ± Thus, remove a certain contractible, smooth, compact 4-manifold Y X − − ⊂ (called an Akbulut cork) from X , and re-glue it by an involution of ∂Y , i.e. a − − diffeomorphism τ : ∂Y ∂Y with τ τ = Id and τ(p) = p for all p ∂Y . − − − → ◦ 6 ± ∈ This argument was modified above so that it works for a contractible open subset R X with similar properties, such that R will be an exotic R4 if X is not − − − + ⊂ diffeomorphictoX . FurthermoreR liesinacompactset,i.e. a4-sphereorR isa − − − smallexoticR4. Inthenextsubsectionwewillseehowthisresultsintheconstruction of handlebodies of exotic R4. In [DeMichelis and Freedman(1992)] Freedman and DeMichelis constructed also a continuous family of small exotic R4. Now we are ready to discuss the decomposition of a small exotic R4 by Bizaca and Gompf [Bi˘zaca and Gompf(1996)] by using special pieces, the handles forming a handle body. Every 4-manifold can be decomposed (seen as handle body) using standardpiecessuchasDk D4−k,theso-calledk-handleattachedalong∂Dk D4−k × × to the boundary S3 = ∂D4 of a 0 handle D0 D4 = D4. The construction of the − × handle body can be divided into two parts. The first part is known as the Akbulut cork,acontractable4-manifoldwithboundaryahomology3-sphere(a3-manifoldwith the same homology as the 3-sphere). The Akbulut cork A is given by a linking cork betweena1-handleanda2-handleofframing0. ThesecondpartistheCassonhandle CH which will be considered now. Let us start with the basic construction of the Casson handle CH. Let M be a smooth, compact, simple-connected 4-manifold and f : D2 M a (codimension-2) → mapping. By using diffeomorphisms of D2 and M, one can deform the mapping f to get an immersion (i.e. injective differential) generically with only double points (i.e. #f−1(f(x)) = 2) as singularities [Golubitsky and Guillemin(1973)]. But to | | incorporate the generic location of the disk, one is rather interesting in the mapping of a 2-handle D2 D2 induced by f id:D2 D2 M from f. Then every double × × × → point (or self-intersection) of f(D2) leads to self-plumbings of the 2-handle D2 D2. × A self-plumbing is an identification of D2 D2 with D2 D2 where D2,D2 D2 0 × 1 × 0 1 ⊂ are disjointsub-disks ofthe firstfactordisk . Consider the pair(D2 D2,∂D2 D2) § × × and produce finitely many self-plumbings away from the attaching region ∂D2 D2 × to get a kinky handle (k,∂−k) where ∂−k denotes the attaching region of the kinky handle. A kinky handle (k,∂−k) is a one-stage tower (T ,∂−T ) and an (n+1)-stage 1 1 tower (T ,∂−T ) is an n-stage tower union kinky handles n (T ,∂−T ) where n+1 n+1 ℓ=1 ℓ ℓ S § Incomplexcoordinatestheplumbingmaybewrittenas(z,w)7→(w,z)or(z,w)7→(w¯,z¯)creating eitherapositiveornegative (respectively)doublepointonthediskD2×0(thecore). Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology8 two towers are attached along ∂−T . Let T− be (interiorT ) ∂−T and the Casson ℓ n n ∪ n handle CH = T− ℓ ℓ[=0 is the union of towers (with direct limit topology induced from the inclusions T ֒ n → T ). n+1 The main idea of the construction above is very simple: an immersed disk (disk with self-intersections) can be deformed into an embedded disk (disk without self- intersections) by sliding one part of the disk along another (embedded) disk to kill the self-intersections. Unfortunately the other disk can be immersed only. But the immersion can be deformed to an embedding by a disk again etc. In the limit of this processone”shiftsthe self-intersectionsintoinfinity” andobtains the standardopen 2-handle (D2 R2,∂D2 R2). k × × A Casson handle is specified up to (orientation preserving) diffeomorphism (of pairs) by a labeled finitely-branching tree with base-point *, having all edge paths infinitely extendable away from *. Each edge should be given a label + or . Here − is the construction: tree CH. Each vertex corresponds to a kinky handle; the → self-plumbing number of that kinky handle equals the number of branches leaving the vertex. The sign on each branch corresponds to the sign of the associated self plumbing. The whole process generates a tree with infinite many levels. In principle,everytreewithafinitenumberofbranchesperlevelrealizesacorresponding Casson handle. Each building block of a Casson handle, the “kinky” handle with n kinks , is diffeomorphic to the n times boundary-connected sum ♮ (S1 D3) (see n ¶ − × appendixAppendix A)withtwoattachingregions. Technicallyspeaking,oneregionis atubularneighborhoodofbandsumsofWhiteheadlinksconnectedwiththeprevious block. The other region is a disjoint union of the standard open subsets S1 D2 in × # S1 S2 =∂(♮ S1 D3) (this is connected with the next block). n n × × 5. The action functional In this section we will discuss the Einstein-Hilbert action functional S (M)= R√gd4x (2) EH Z M of the 4-manifold M and fix the Ricci-flat metric g as solution of the vacuum field equations of the exotic 4-manifold. The main part of our argumentationis additional contribution to the action functional coming from exotic smoothness. 5.1. Large exotic R4 In case of the large exotic R4, we consider the decompositions R4 = R4 intρ(X ) X (3) \ K ∪∂XK K R4 = (cid:0)R4 intρ(X )(cid:1) ρ(X ) (4) \ K ∪∂XK K (cid:0) (cid:1) k IntheproofofFreedman[Freedman(1982b)],themaincomplicationscomefromthelackofcontrol aboutthisprocess. ¶ Thenumberofend-connectedsumsisexactlythenumberofselfintersectionsoftheimmersedtwo handle. Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology9 leading to a sum in the action S (R4)= R√gd4x= R√gd4x+ R√gd4x EH Z Z Z R4 R4\intρ(XK) ρ(XK) =S (R4 intρ(X ))+S (ρ(X )) . EH K EH K \ Because of diffeomorphism invariance of the Einstein-Hilbert action, this decomposi- tiondo notdepend onthe concreterealizationwithrespecttoanycoordinatesystem. Therefore we obtain the relation S (R4 intρ(X ))=S (R4) S (ρ(X )) (5) EH K EH EH K \ − and get a similar relation using (3) between the action S (R4) of the standard R4 EH and the action S (R4) of the large exotic R4 EH S (R4)=S (R4 intρ(X ))+S (X ) EH EH K EH K \ =S (R4)+S (X ) S (ρ(X )) . (6) EH EH K EH K − The knot is topologically slice (ρ is a flat topological embedding). Therefore the restriction of ρ to the 2-handle D2 D2 in X is a topological embedding defining K × an embedding ρ′ : D2 D4 with ρ′(∂D2) = K. From the topological point of → view, the Seifert surface of K is the disc D2 with genus 0. Then we obtain using X =D4 (D2 D2) K ∪ × S (ρ(X ))=S (ρ(D4))+S (ρ(D2 D2))=S (ρ(D2 D2)) EH K EH EH EH × × assumingtheflatnessofthe0 handleD4. Theproductmetric(blockdiagonalmetric) − ds2 =g dx2+g dx2 D1 1 D2 2 on the image ρ(D2 D2)=D D of the 2-handle with ∂D =K induces 1 2 1 × × S (ρ(D2 D2))=vol(D ) R √g d2x +vol(D ) R √g d2x EH × 2 ·Z D1 D1 1 1 ·Z D2 D2 2 D1 D2 with the curvature scalars R ,R . The 2-dimensional integrals D1 D2 Z RD1√gD1d2x1 =2π·χ(D1)=2π D1 Z RD2√gD2d2x2 =2π·χ(D2)=2π D2 arebydefinitiontheEulercharacteristicsχ(D )=1,χ(D )=1usingthetopologically 1 2 sliceness of the knot K =∂D . Finally we obtain 1 S (ρ(X ))=2π vol(D )+2π vol(D ) . EH K 2 1 · · Now we consider the other action S (X ) where we use a non-flat embedding EH K X ֒ R4. Remember the knot K is smoothly not slice. But then we can only K → choose the embedding so that the minimal genus g (K) of the Seifert surface F is 4 non-zero, i.e. one obtains for the Euler characteristics R √g d2x=2π (1 2g (K)) . F F 4 Z · − F Exotic Smoothness and Quantum Gravity II: exotic R4, singularities and cosmology10 This genus g (K) is an invariant of the knot also known as smooth 4-genus. 4 Importantly the Seifert surface F has negative curvature for g (K) > 0. A similar 4 argumentation leads to the result S (X )=2π vol(D ) (1 2g (K))+2π vol(F) EH K 2 4 · · − · and finally we have the relation using (6) and the results above S (R4)=S (R4) 4π vol(D ) g (K)+2π (vol(F) vol(D ))(7) EH EH 2 4 1 − · · · − as the correction to the action S (R4) of the large exotic R4. The two surfaces F EH and D have the same boundary (the knot K) and differ only by the embedding. So, 1 it seems natural to assume the same volume, i.e. vol(F) = vol(D ). Finally we will 1 write this relation in the usual units 1 1 vol(D ) S (R4)= S (R4) 2 4π2 g (K) (8) ~ EH ~ EH − L2 · · 4 P Thisexpressionlooksverysimplebutthecomplicationislocatedatthe4-genusg (K). 4 Currently there is no simple expression for the calculation. All results show only the existence g (K) = 0 but never calculate the value. So, we are not satisfied with 4 6 the expression above. We would expect that g (K) is related to the map ρ which is 4 certainlyrelatedtoinfinite constructionslikethe Cassonhandle. Ifthis speculationis correctthenone caninterpretthe expressionaboveas anon-perturbativecalculation. 5.2. Small exotic R4 As explained above, a small exotic R4 can be decomposed into a compact subset A (Akbulutcork)andaCassonhandle(see[Bi˘zaca and Gompf(1996)]). Especially cork this exotic R4 depends strongly on the Casson handle, i.e. non-diffeomorphic Casson handles lead to non-diffeomorphic R4’s. Thus we have to understand the analytical propertiesofaCassonhandle. In[Kato(2004)],theanalyticalpropertiesoftheCasson handle were discussed. The main idea is the usage of the theory of end-periodic manifolds, i.e. an infinite periodic structure generated by W glued along a compact set A to get for the interior cork R4 =int(A W W ) θ cork∪N ∪N ∪N ··· the end-periodic manifold. The definition of an end-periodic manifold is very formal (see[Taubes(1987)])andweomitithere. AllCassonhandlesgeneratedbyabalanced tree have the structure of end-periodic manifolds as shown in [Kato(2004)]. By using thetheoryofTaubes[Taubes(1987)]onecanconstructametricon W W N N N by using the metric on W. Then a metric g in R4 transforms to a··p·e∪riodic∪func∪tion··gˆ· θ on the infinite periodic manifold Y˜ = W W W N −1 N 0 N 1 N ···∪ ∪ ∪ ∪ ··· where W is the building block W at the ith place. Then the action of R4 can be i θ divided into two parts S (R4)=S (A )+ S (W ) (9) EH θ EH cork EH i Xi and we start with the discussion of the compact part A . This part A is cork cork formally given by a so-called plumbing of two spheres A,B with trivial normal bundles having the algebraic intersection number A B = 1 but an extra pair of · intersections (with numbers +1 and 1). The whole construction can be simplified −