The Cartan, Choquet and Kellogg properties for the fine topology on metric spaces
aa r X i v : . [ m a t h . A P ] O c t The Cartan, Choquet and Kellogg properties for thefine topology on metric spaces
Anders Bj¨orn
Department of Mathematics, Link¨opings universitet,SE-581 83 Link¨oping, Sweden ; [email protected] Jana Bj¨orn
Department of Mathematics, Link¨opings universitet,SE-581 83 Link¨oping, Sweden ; [email protected] Visa Latvala
Department of Physics and Mathematics, University of Eastern Finland,P.O. Box 111, FI-80101 Joensuu, Finland ; visa.latvala@uef.fi Abstract . We prove the Cartan and Choquet properties for the fine topology on a com-plete metric space equipped with a doubling measure supporting a p -Poincar´e inequality,1 < p < ∞ . We apply these key tools to establish a fine version of the Kellogg property,characterize finely continuous functions by means of quasicontinuous functions, and showthat capacitary measures associated with Cheeger supersolutions are supported by the fineboundary of the set. Key words and phrases : Capacitary measure, Cartan property, Choquet property, doublingmeasure, fine Kellogg property, finely continuous, finely open, fine topology, metric space,nonlinear potential theory, p -harmonic, Poincar´e inequality, quasicontinuous, quasiopen,thin.Mathematics Subject Classification (2010): Primary: 31E05; Secondary: 30L99, 31C40,31C45, 35J92, 49Q20.
1. Introduction
The aim of this paper is to establish the Cartan and Choquet properties for thefine topology on a complete metric space X equipped with a doubling measure µ supporting a p -Poincar´e inequality, 1 < p < ∞ . These properties are crucial fordeep applications of the fine topology in potential theory. As applications of thesekey tools we establish the fine Kellogg property and characterize finely continuousfunctions by means of quasicontinuous functions. We also show that capacitary mea-sures associated with Cheeger p -supersolutions are supported by the fine boundaryof the set (not just by the metric boundary).The classical fine topology is closely related to the Dirichlet problem for theLaplace equation. Wiener [53] showed in 1924 that a boundary point of a domainis irregular if and only if the complement is thin at that point in a certain capacitydensity sense, cf. Definition 6.1. In 1939 Brelot [22], [23] characterized thinnessby a condition, which is nowadays called the Cartan property. The reason for thisname is that Cartan (in a letter to Brelot in 1940, see Brelot [24, p. 14]) connected1 Anders Bj¨orn, Jana Bj¨orn and Visa Latvala the notion of thinness to the coarsest topology making all superharmonic functionscontinuous. Cartan [26] coined the name fine topology for such a topology.Nonlinear potential theory associated with p -harmonic functions has been stud-ied since the 1960s. For extensive treatises and notes on the history, see the mono-graphs Adams–Hedberg [1], Heinonen–Kilpel¨ainen–Martio [37] and Mal´y–Ziemer [49].Starting in the 1990s a lot of attention has been given to analysis on metric spaces,see e.g. Haj lasz–Koskela [32], Heinonen [35] and Heinonen–Koskela [38]. Around2000 this initiated studies of p -harmonic and p -superharmonic functions on metricspaces without a differentiable structure, see e.g. Bj¨orn–Bj¨orn [9], Bj¨orn–Bj¨orn–Shanmugalingam [15], Kinnunen–Martio [43], Kinnunen–Shanmugalingam [44] andShanmugalingam [52].The classical linear fine potential theory and fine topology (the case p = 2) havebeen systematically studied since the 1960s. Let us here just mention Brelot [25],Fuglede [30], [31] and Lukeˇs–Mal´y–Zaj´ıˇcek [48], which include most of the theoryand the main references. Some of these works are written in large generality includ-ing topological spaces, general capacities and families of functions, and some resultsthus apply also to the nonlinear theory. At the same time, many other results relyindirectly on a linear structure, e.g. through potentials, integral representations andconvex cones of superharmonic functions, which are in general not available in thenonlinear setting.The nonlinear fine potential theory started in the 1970s on unweighted R n ,see the notes to Chapter 12 in [37] and Section 2.6 in [49]. For the fine poten-tial theory associated with p -harmonic functions on unweighted R n , see [49] andLatvala [47]. The monograph [37] is the main source for fine potential theory onweighted R n (note that Chapter 21, which is only in the second addition, containssome more recent results). The study of fine potential theory on metric spacesis more recent, see e.g. Bj¨orn–Bj¨orn [10], Bj¨orn–Bj¨orn–Latvala [11], J. Bj¨orn [20],Kinnunen–Latvala [42] and Korte [45]. For further references to nonlinear and finenonlinear potential theory, see the introduction to [11].Recently, in [11], we established the so-called weak Cartan property , which saysthat if E ⊂ X is thin at x / ∈ E , then there exist a ball B ∋ x and superharmonicfunctions u, u ′ on B such that v ( x ) < lim inf E ∋ x → x v ( x ) , where v = max { u, u ′ } .The superharmonic functions considered in [11] were based on upper gradients,and because of the lack of a differential equation, we did not succeed in obtainingthe full Cartan property as in R n , where v itself can be chosen superharmonic,cf. Theorem 1.1 below. Indeed, the proof of the full Cartan property seems to beas hard as the proof of the Wiener criterion, which is also open in the nonlinearpotential theory based on upper gradients, but is known to hold in the potentialtheory based on Cheeger gradients, see J. Bj¨orn [19]. Nevertheless, the weak Cartanproperty in [11] was enough to conclude that the fine topology is the coarsest onemaking all superharmonic functions continuous.Here, we instead focus on Cheeger superharmonic functions based on Cheeger’stheorem yielding a vector-valued Cheeger gradient. In this case we do have anequation available and this enables us to establish the following full Cartan property. Theorem 1.1. (Cartan property)
Suppose that E is thin at x ∈ E \ E . Then thereis a bounded positive Cheeger superharmonic function u in an open neighbourhoodof x such that u ( x ) < lim inf E ∋ x → x u ( x ) . he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 3 For a Newtonian function, the minimal p -weak upper gradient and the modulusof the Cheeger gradient are comparable. Thus the corresponding capacities are com-parable to each other, and the fine topology, as well as thinness (and thickness), isthe same in both cases. Superminimizers, superharmonic and p -harmonic functionsare however different. Hence, using the Cheeger structure we can study thinness andthe fine topology, but not e.g. the superharmonic and p -harmonic functions basedon upper gradients. Only Cheeger p -(super)harmonic functions can be treated.We use the Cartan property to establish the following important Choquet prop-erty. Theorem 1.2. (Choquet property)
For any E ⊂ X and any ε > there is an openset G containing all the points in X at which E is thin, such that C p ( E ∩ G ) < ε . The Choquet property was first established by Choquet [28] in 1959. In thenonlinear theory on unweighted R n it was later established by Hedberg [33] andHedberg–Wolff [34] in connections with potentials (also for higher-order Sobolevspaces). The Cartan property for p -superharmonic functions on unweighted R n was obtained by Kilpel¨ainen–Mal´y [41] as a consequence of their pointwise Wolff-potential estimates. In fact, Kilpel¨ainen and Mal´y used the Cartan property toestablish the necessity in the Wiener criterion. In Mal´y–Ziemer [49], the authorsdeduce the Choquet property from the Cartan property. The proof of the Cartanproperty was extended to weighted R n by Mikkonen [50, Theorem 5.8] and can alsobe found in Heinonen–Kilpel¨ainen–Martio [37, Theorem 21.26 (which is only in thesecond edition)]; in both places they however refrained from deducing consequencessuch as the Choquet property.Our proof of the Choquet property follows the one in [49], but we have some extracomplications due to the fact that we can simultaneously have some points with zerocapacity and others with positive capacity. Note that Fuglede [30] contains a proofof the Choquet property, in an axiomatic setting, assuming that Corollary 1.3 andpart (a) in Theorem 1.4 are true. We have a converse approach, since our proofs ofCorollary 1.3 and Theorem 1.4 are based on the Choquet property. Corollary 1.3. (Fine Kellogg property)
For any E ⊂ X we have C p ( { x ∈ E : E is thin at x } ) = 0 . (1.1)The fine Kellogg property has close connections with boundary regularity, seeRemark 7.3. The implications ⇒ in the following result were already obtained inBj¨orn–Bj¨orn–Latvala [11], but now we are able to complete the picture. Theorem 1.4. (a)
A set U ⊂ X is quasiopen if and only if U = V ∪ E for somefinely open set V and for a set E of capacity zero. (b) An extended real-valued function on a quasiopen set U is quasicontinuous in U if and only if u is finite q.e. and finely continuous q.e. in U . It is pointed out in Adams–Lewis [2, Proposition 3] that (a) for unweighted R n follows from the Choquet property established in Hedberg–Wolff [34]. Also (b) thenfollows by modifying the earlier axiomatic argumentation of Fuglede [30, Lemma,p. 143]. The proof of Theorem 1.4 in unweighted R n is given in Mal´y–Ziemer [49,p. 146]. For the reader’s convenience, we include the proof of Theorem 1.4 althoughthe proof essentially follows [49]. In Section 8, we use Theorem 1.4 to extend andsimplify some recent results from Bj¨orn–Bj¨orn [10].We end the paper with another application of the Cartan property in Section 9,which contains results on capacitary measures related to Cheeger supersolutions, seeTheorem 9.1 and Corollaries 9.6 and 9.7. In particular, we show that the capacitarymeasure only charges the fine boundary of the corresponding set. Anders Bj¨orn, Jana Bj¨orn and Visa Latvala
Acknowledgement.
The first two authors were supported by the Swedish Re-search Council. Part of this research was done during several visits of the thirdauthor to Link¨opings universitet in 2012–2014 and while all three authors visitedInstitut Mittag-Leffler in the autumn of 2013. We thank both institutions for theirhospitality and support.
2. Notation and preliminaries
We assume throughout the paper that 1 < p < ∞ and that X = ( X, d, µ ) is ametric space equipped with a metric d and a positive complete Borel measure µ such that 0 < µ ( B ) < ∞ for all (open) balls B ⊂ X . It follows that X is separable.The σ -algebra on which µ is defined is obtained by the completion of the Borel σ -algebra. We also assume that Ω ⊂ X is a nonempty open set.We say that µ is doubling if there exists a doubling constant C > B = B ( x , r ) := { x ∈ X : d ( x, x ) < r } in X ,0 < µ (2 B ) ≤ Cµ ( B ) < ∞ . Here and elsewhere we let δB = B ( x , δr ). A metric space with a doubling measureis proper (i.e. closed and bounded subsets are compact) if and only if it is complete.See Heinonen [35] for more on doubling measures.A curve is a continuous mapping from an interval, and a rectifiable curve isa curve with finite length. We will only consider curves which are nonconstant,compact and rectifiable. A curve can thus be parameterized by its arc length ds .We follow Heinonen and Koskela [38] in introducing upper gradients as follows (theycalled them very weak gradients). Definition 2.1.
A nonnegative Borel function g on X is an upper gradient of anextended real-valued function f on X if for all nonconstant, compact and rectifiablecurves γ : [0 , l γ ] → X , | f ( γ (0)) − f ( γ ( l γ )) | ≤ Z γ g ds, (2.1)where we follow the convention that the left-hand side is ∞ whenever at least oneof the terms therein is infinite. If g is a nonnegative measurable function on X andif (2.1) holds for p -almost every curve (see below), then g is a p -weak upper gradient of f .Here we say that a property holds for p -almost every curve if it fails only fora curve family Γ with zero p -modulus, i.e. there exists 0 ≤ ρ ∈ L p ( X ) such that R γ ρ ds = ∞ for every curve γ ∈ Γ. Note that a p -weak upper gradient need not bea Borel function, it is only required to be measurable. On the other hand, everymeasurable function g can be modified on a set of measure zero to obtain a Borelfunction, from which it follows that R γ g ds is defined (with a value in [0 , ∞ ]) for p -almost every curve γ . For proofs of these and all other facts in this section werefer to Bj¨orn–Bj¨orn [9] and Heinonen–Koskela–Shanmugalingam–Tyson [39].The p -weak upper gradients were introduced in Koskela–MacManus [46]. It wasalso shown there that if g ∈ L p loc ( X ) is a p -weak upper gradient of f , then one canfind a sequence { g j } ∞ j =1 of upper gradients of f such that g j − g → L p ( X ). If f has an upper gradient in L p loc ( X ), then it has a minimal p -weak upper gradient g f ∈ L p loc ( X ) in the sense that for every p -weak upper gradient g ∈ L p loc ( X ) of f we have g f ≤ g a.e., see Shanmugalingam [52]. The minimal p -weak upper gradientis well defined up to a set of measure zero in the cone of nonnegative functions in L p loc ( X ). Following Shanmugalingam [51], we define a version of Sobolev spaces onthe metric measure space X . he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 5 Definition 2.2.
Let for measurable f , k f k N ,p ( X ) = (cid:18)Z X | f | p dµ + inf g Z X g p dµ (cid:19) /p , where the infimum is taken over all upper gradients of f . The Newtonian space on X is N ,p ( X ) = { f : k f k N ,p ( X ) < ∞} . The space N ,p ( X ) / ∼ , where f ∼ h if and only if k f − h k N ,p ( X ) = 0, is aBanach space and a lattice, see Shanmugalingam [51]. In this paper we assumethat functions in N ,p ( X ) are defined everywhere (with values in R := [ −∞ , ∞ ]),not just up to an equivalence class in the corresponding function space. For ameasurable set E ⊂ X , the Newtonian space N ,p ( E ) is defined by considering( E, d | E , µ | E ) as a metric space on its own. We say that f ∈ N ,p loc ( E ) if for every x ∈ E there exists a ball B x ∋ x such that f ∈ N ,p ( B x ∩ E ). If f, h ∈ N ,p loc ( X ),then g f = g h a.e. in { x ∈ X : f ( x ) = h ( x ) } , in particular g min { f,c } = g f χ { f The Sobolev capacity of an arbitrary set E ⊂ X is C p ( E ) = inf u k u k pN ,p ( X ) , where the infimum is taken over all u ∈ N ,p ( X ) such that u ≥ E .The Sobolev capacity is countably subadditive. We say that a property holds quasieverywhere (q.e.) if the set of points for which the property does not hold hasSobolev capacity zero. The Sobolev capacity is the correct gauge for distinguishingbetween two Newtonian functions. If u ∈ N ,p ( X ), then u ∼ v if and only if u = v q.e. Moreover, Corollary 3.3 in Shanmugalingam [51] shows that if u, v ∈ N ,p ( X )and u = v a.e., then u = v q.e.A set U ⊂ X is quasiopen if for every ε > G ⊂ X suchthat C p ( G ) < ε and G ∪ U is open. A function u defined on a set E ⊂ X is quasicontinuous if for every ε > G ⊂ X such that C p ( G ) < ε and u | E \ G is finite and continuous. If u is quasicontinuous on a quasiopen set U ,then it is easily verified that { x ∈ U : u ( x ) < a } and { x ∈ U : u ( x ) > a } arequasiopen for all a ∈ R , cf. Proposition 3.3 in Bj¨orn–Bj¨orn–Mal´y [13]. Definition 2.4. We say that X supports a p -Poincar´e inequality if there existconstants C > λ ≥ B ⊂ X , all integrable functions f on X and all upper gradients g of f , Z B | f − f B | dµ ≤ C diam( B ) (cid:18)Z λB g p dµ (cid:19) /p , (2.2)where f B := R B f dµ := R B f dµ/µ ( B ).In the definition of Poincar´e inequality we can equivalently assume that g is a p -weak upper gradient.In R n equipped with a doubling measure dµ = w dx , where dx denotes Lebesguemeasure, the p -Poincar´e inequality (2.2) is equivalent to the p -admissibility of theweight w in the sense of Heinonen–Kilpel¨ainen–Martio [37], cf. Corollary 20.9 in [37]and Proposition A.17 in [9].If X is complete and supports a p -Poincar´e inequality and µ is doubling, thenLipschitz functions are dense in N ,p ( X ), see Shanmugalingam [51]. Moreover, all Anders Bj¨orn, Jana Bj¨orn and Visa Latvala functions in N ,p ( X ) and those in N ,p (Ω) are quasicontinuous, see Bj¨orn–Bj¨orn–Shanmugalingam [16]. This means that in the Euclidean setting, N ,p ( R n ) is therefined Sobolev space as defined in Heinonen–Kilpel¨ainen–Martio [37, p. 96], seeBj¨orn–Bj¨orn [9, Appendix A.2] for a proof of this fact valid in weighted R n . Thisis the main reason why, unlike in the classical Euclidean setting, we do not need torequire the functions competing in the definitions of capacity to be 1 in a neigh-bourhood of E . For recent related progress on the density of Lipschitz functionssee Ambrosio–Colombo–Di Marino [3] and Ambrosio–Gigli–Savar´e [4].In Section 6 the fine topology is defined by means of thin sets, which in turnuse the variational capacity cap p . To be able to define the variational capacity wefirst need a Newtonian space with zero boundary values. We let, for an arbitraryset A ⊂ X , N ,p ( A ) = { f | A : f ∈ N ,p ( X ) and f = 0 on X \ A } . One can replace the assumption “ f = 0 on X \ A ” with “ f = 0 q.e. on X \ A ” withoutchanging the obtained space N ,p ( A ). Functions from N ,p ( A ) can be extended byzero in X \ A and we will regard them in that sense if needed. Definition 2.5. The variational capacity of E ⊂ Ω with respect to Ω iscap p ( E, Ω) = inf u Z X g pu dµ, where the infimum is taken over all u ∈ N ,p (Ω) such that u ≥ E .If C p ( E ) = 0, then cap p ( E, Ω) = 0. The converse implication is true if µ isdoubling and supports a p -Poincar´e inequality.In Section 9 we will need the following simple lemma. For the reader’s conve-nience we provide the short proof. Lemma 2.6. If u, v ∈ N ,p ( X ) are bounded, then uv ∈ N ,p ( X ) .Proof. We can assume that | u | and | v | are bounded by 1. Then | uv | ≤ | u | andhence uv ∈ L p ( X ). By the Leibniz rule (Theorem 2.15 in Bj¨orn–Bj¨orn [9]), g := | u | g v + | v | g u is a p -weak upper gradient of uv . As g ≤ g v + g u ∈ L p ( X ) we see that uv ∈ N ,p ( X ).Throughout the paper, the letter C will denote various positive constants whosevalues may vary even within a line. We also write A ≃ B if C − A ≤ B ≤ CA . 3. Cheeger gradients Throughout the rest of the paper, we assume that X is complete and supports a p -Poincar´e inequality, and that µ is doubling. In addition to upper gradients we will also use Cheeger gradients in this paper.Their existence is based on the following deep result of Cheeger. Theorem 3.1. (Theorem 4.38 in Cheeger [27]) There exists N and a countablecollection ( U α , X α ) of pairwise disjoint measurable sets U α and Lipschitz “coordi-nate” functions X α : X → R k ( α ) , 1 ≤ k ( α ) ≤ N , such that µ (cid:0) X \ S α U α (cid:1) = 0 andfor every Lipschitz function f : X → R there exist unique bounded vector-valuedfunctions d α f : U α → R k ( α ) such that for a.e. x ∈ U α ,lim r → sup y ∈ B ( x,r ) | f ( y ) − f ( x ) − d α f ( x ) · ( X α ( y ) − X α ( x )) | r = 0 , where · denotes the usual inner product in R k ( α ) . he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 7 Cheeger further shows that for a.e. x ∈ U α , there is an inner product norm | · | x on R k ( α ) such that for all Lipschitz f ,1 C g f ( x ) ≤ | d α f ( x ) | x ≤ Cg f ( x ) , (3.1)where C is independent of f and x , see p. 460 in Cheeger [27]. As Lipschitz functionsare dense in N ,p ( X ), the “gradients” d α f extend uniquely to the whole N ,p ( X ),by Franchi–Haj lasz–Koskela [29, Theorem 10] or Keith [40]. Moreover, (3.1) holdseven for functions in N ,p ( X ).From now on we drop α and set Df := d α f in U α . On a metric space there is some freedom in choosing the Cheeger structure.On R n we will however always make the natural choice Df = ∇ f and let theinner product norm in (3.1) be the Euclidean norm. Here ∇ f denotes the Sobolevgradient from Heinonen–Kilpel¨ainen–Martio [37], which equals the distributionalgradient if the weight on R n is a Muckenhoupt A p weight. In this case, | Df | = g f ,by Proposition A.13 in Bj¨orn–Bj¨orn [9]. 4. Supersolutions and superharmonic functions In the literature on potential theory in metric spaces one usually studies the follow-ing (super)minimizers based on upper gradients. Definition 4.1. A function u ∈ N ,p loc (Ω) is a ( super ) minimizer in Ω if Z { ϕ =0 } g pu dµ ≤ Z { ϕ =0 } g pu + ϕ dµ for all (nonnegative) ϕ ∈ Lip c (Ω) . A p -harmonic function is a continuous minimizer.Here Lip c (Ω) = { ϕ ∈ Lip( X ) : supp ϕ ⋐ Ω } and E ⋐ Ω if E is a compact subsetof Ω.Minimizers were first studied by Shanmugalingam [52], and superminimizers byKinnunen–Martio [43]. For various characterizations of minimizers and supermini-mizers see A. Bj¨orn [6]. If u is a superminimizer, then its lsc-regularization u ∗ ( x ) := ess lim inf y → x u ( y ) = lim r → ess inf B ( x,r ) u (4.1)is also a superminimizer and u ∗ = u q.e., see [43] or Bj¨orn–Bj¨orn–Parviainen [14].If u is a minimizer, then u ∗ is continuous, (by Kinnunen–Shanmugalingam [44] orBj¨orn–Marola [17]), and thus p -harmonic. For further discussion and references onthe topics in this section see [9].In this paper, we consider Cheeger ( super ) minimizers and Cheeger p -harmonicfunctions defined by replacing g u and g u + ϕ in Definition 4.1 by | Du | and | D ( u + ϕ ) | ,respectively, where |·| is the inner product norm in (3.1). Due to the vector structureof the Cheeger gradient one can also make the following definition. (There is nocorresponding notion for upper gradients.) Definition 4.2. A function u ∈ N ,p loc (Ω) is a ( super ) solution in Ω if Z Ω | Du | p − Du · Dϕ dµ ≥ ϕ ∈ Lip c (Ω) , where · is the inner product giving rise to the norm in (3.1). Anders Bj¨orn, Jana Bj¨orn and Visa Latvala It can be shown that a function is a (super)solution if and only if it is a Cheeger(super)minimizer, the proof is the same as for Theorem 5.13 in Heinonen–Kilpe-l¨ainen–Martio [37]. In weighted R n , with the choice Df = ∇ f , we have g f = | Df | = |∇ f | a.e. which implies that (super)minimizers, Cheeger (super)minimizersand (super)solutions coincide, and are the same as in [37].Let G ⊂ X be a nonempty bounded open set with C p ( X \ G ) > 0. We considerthe following obstacle problem in G . Definition 4.3. For f ∈ N ,p ( G ) and ψ : G → R let K ψ,f ( G ) = { v ∈ N ,p ( G ) : v − f ∈ N ,p ( G ) and v ≥ ψ q.e. in G } . A function u ∈ K ψ,f ( G ) is a solution of the K ψ,f ( G ) -Cheeger obstacle problem if Z G | Du | p dµ ≤ Z G | Dv | p dµ for all v ∈ K ψ,f ( G ) . A solution to the K ψ,f ( G )-Cheeger obstacle problem is easily seen to be aCheeger superminimizer (i.e. a supersolution) in G . Conversely, a supersolution u in Ω is a solution of the K u,u ( G )-Cheeger obstacle problem for all open G ⋐ Ωwith C p ( X \ G ) > K ψ,f ( G ) = ∅ , then there is a solution u of the K ψ,f ( G )-Cheeger obstacleproblem, and this solution is unique up to equivalence in N ,p ( G ). Moreover, u ∗ is the unique lsc-regularized solution. Conditions for when K ψ,f ( G ) = ∅ can befound in Bj¨orn–Bj¨orn [10]. If the obstacle ψ is continuous, as a function with valuesin [ −∞ , ∞ ), then u ∗ is also continuous. These results were obtained for the uppergradient obstacle problem by Kinnunen–Martio [43], where superharmonic functionsbased on upper gradients were also introduced. As with most of the results in themetric theory their proofs work verbatim for the Cheeger case considered here. Sincemost of the theory has been developed in the setting of upper gradients, we will oftenjust refer to the upper gradient equivalents of results for Cheeger (super)minimizers.For f ∈ N ,p ( G ), we let H G f denote the continuous solution of the K −∞ ,f ( G )-Cheeger obstacle problem. This function is Cheeger p -harmonic in G and has thesame boundary values (in the Sobolev sense) as f on ∂G , and hence is also calledthe solution of the (Cheeger) Dirichlet problem with Sobolev boundary values. Definition 4.4. A function u : Ω → ( −∞ , ∞ ] is Cheeger superharmonic in Ω if(i) u is lower semicontinuous;(ii) u is not identically ∞ in any component of Ω;(iii) for every nonempty open set Ω ′ ⋐ Ω with C p ( X \ Ω ′ ) > v ∈ Lip( X ), we have H Ω ′ v ≤ u in Ω ′ whenever v ≤ u on ∂ Ω ′ .This definition of Cheeger superharmonicity is equivalent to the one in Heino-nen–Kilpel¨ainen–Martio [37], see A. Bj¨orn [5]. A locally bounded Cheeger super-harmonic function is a supersolution, and all Cheeger superharmonic functions arelsc-regularized. Conversely, any lsc-regularized supersolution is Cheeger superhar-monic. See Kinnunen–Martio [43]. Definition 4.5. The Cheeger capacitary potential u E of a set E ⊂ G in G is thelsc-regularized solution of the K χ E , ( G )-Cheeger obstacle problem.The Cheeger variational capacity of E ⊂ G is defined asCh-cap p ( E, G ) = Z X | Du E | p dµ = inf u Z X | Du | p dµ, (4.2)where the infimum is taken over all u ∈ N ,p ( G ) such that u ≥ E .By (3.1), we have Ch-cap p ( E, G ) ≃ cap p ( E, G ) . (4.3) he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 9 5. Supersolutions and Radon measures In this section we assume that Ω is a nonempty bounded open set with C p ( X \ Ω) > . It was shown in Bj¨orn–MacManus–Shanmugalingam [21, Propositions 3.5 and 3.9]that there is a one-to-one correspondence between supersolutions in Ω and Radonmeasures in the dual N ,p (Ω) ′ . A Radon measure is a positive complete Borelmeasure which is finite on every compact set. Proposition 5.1. For every supersolution u in Ω there is a Radon measure ν ∈ N ,p (Ω) ′ such that for all ϕ ∈ N ,p (Ω) , T u ( ϕ ) := Z Ω | Du | p − Du · Dϕ dµ = Z Ω ϕ dν, (5.1) where · is the inner product giving rise to the norm in (3.1) .Conversely, if ν ∈ N ,p (Ω) ′ is a Radon measure on Ω then there exists a uniquelsc-regularized u ∈ N ,p (Ω) satisfying T u = ν in the sense of (5.1) for all ϕ ∈ N ,p (Ω) . Moreover, u is a nonnegative supersolution in Ω . Remark 5.2. This result is always false if we drop the assumption C p ( X \ Ω) > u is a nonnegative lsc-regularized supersolution in Ω, then u is Cheegersuperharmonic in Ω. If C p ( X \ Ω) = 0, then u has a Cheeger superharmonicextension to X , by Theorem 6.3 in A. Bj¨orn [7] (or Theorem 12.3 in [9]), which mustbe constant, by Corollary 9.14 in [9]. (Note that if Ω is bounded and C p ( X \ Ω) = 0,then also X must be bounded.) On the other hand, there are nonzero Radonmeasures in N ,p (Ω) ′ , so the existence of a corresponding supersolution fails. Proof of Proposition . See [21, Propositions 3.5 and 3.9], where the result wasstated under stronger assumptions than here, but the proof of this result is validunder our assumptions. In particular, as C p ( X \ Ω) > 0, the coercivity of themap T follows from the Poincar´e inequality for N ,p (also called the p -Friedrichs’inequality), whose proof can be found e.g. in Corollary 5.54 in [9].In [21] the uniqueness was shown up to equivalence between supersolutions.The pointwise uniqueness for lsc-regularized supersolutions then follows from (4.1).That u is nonnegative follows from Lemma 5.3 below, as u ≡ Lemma 5.3. Let ν , ν ∈ N ,p (Ω) ′ be Radon measures such that ν ≤ ν . If u , u ∈ N ,p (Ω) are the corresponding lsc-regularized supersolutions given by Proposition ,then u ≤ u in Ω .Proof. Inserting ϕ = ( u − u ) + ∈ N ,p (Ω) into the equation (5.1) for u and u gives 0 ≤ Z Ω ϕ dν − Z Ω ϕ dν (5.2)= Z Ω ( | Du | p − Du − | Du | p − Du ) · Dϕ dµ = Z { u >u } ( | Du | p − Du − | Du | p − Du ) · ( Du − Du ) dµ ≤ Z { u >u } ( | Du | p − | Du | + | Du | p − | Du | − | Du | p − | Du | p ) dµ. The Young inequality shows that | Du | p − | Du | + | Du | p − | Du | ≤ p − p | Du | p + 1 p | Du | p + p − p | Du | p + 1 p | Du | p = | Du | p + | Du | p . (5.3)Inserting this into (5.2) shows that equality must hold in (5.2) and (5.3) for a.e. x such that u ( x ) > u ( x ). This implies that Du ( x ) = k ( x ) Du ( x ) for some k ( x ) ≥ | Du ( x ) | = | Du ( x ) | (by equality in theYoung inequality) for a.e. x with u ( x ) > u ( x ). It follows that Du ( x ) = Du ( x )for a.e. such x and hence Dϕ = 0 a.e. in Ω. The Poincar´e inequality for N ,p (e.g.Corollary 5.54 in [9]) then yields Z Ω ϕ p dµ ≤ C Ω Z Ω | Dϕ | p dµ = 0 . Hence ϕ = 0 a.e. in Ω, i.e. u ≤ u a.e. in Ω. As u and u are lsc-regularized, itfollows that u ≤ u everywhere in Ω. Remark 5.4. Note that if u E is the Cheeger capacitary potential of E in Ω, givenby Definition 4.5, then u E is the lsc-regularized solution of the K ψ, (Ω)-Cheegerobstacle problem, where ψ = 1 in E and ψ = −∞ otherwise. Hence, for every ϕ ∈ N ,p (Ω \ E ) and every t > 0, the function u E + tϕ ∈ K ψ, (Ω) and thus0 ≤ Z Ω ( | Du E + tDϕ | p − | Du E | p ) dµ. Dividing by t and letting t → Z Ω | Du E | p − Du E · Dϕ dµ ≥ , (5.4)see (2.8) in Mal´y–Ziemer [49]. Applying this also to − ϕ shows that equality musthold in (5.4). Consequently, the measure ν E = T u E satisfies Z Ω ϕ dν E = 0 for every ϕ ∈ N ,p (Ω \ E ) . (5.5)We will need the following lemma when proving the Cartan property (Theo-rem 1.1). Later, in Theorem 9.1, we will generalize this lemma to quasiopen setsand as a consequence obtain that the measure ν E is supported on the fine boundary ∂ p E ; that it is supported on the boundary ∂E is well known. Lemma 5.5. Let E ⊂ Ω be such that cap p ( E, Ω) < ∞ and let u E be the Cheegercapacitary potential of E in Ω , with the corresponding Radon measure ν E = T u E .If G ⊂ Ω is open and v ∈ N ,p (Ω) is bounded and such that v = 1 q.e. in G ∩ E then Z G v dν E = Z G u E dν E . (5.6) In particular, ν E ( G ) = R G u E dν E , and if C p ( G ∩ E ) = 0 then ν E ( G ) = 0 .Proof. For every η ∈ Lip c ( G ) with 0 ≤ η ≤ η ( v − u E ) ∈ N ,p (Ω \ E ).Thus, (5.5) yields that Z G η ( v − u E ) dν E = 0 . Since v − u E and G are bounded, dominated convergence and letting η ր χ G imply (5.6). For the last part, apply this to v = 1 and v = 0 respectively. he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 11 6. Thinness and the fine topology We now define the fine topological notions which are central in this paper. Definition 6.1. A set E ⊂ X is thin at x ∈ X if Z (cid:18) cap p ( E ∩ B ( x, r ) , B ( x, r ))cap p ( B ( x, r ) , B ( x, r )) (cid:19) / ( p − drr < ∞ . (6.1)A set U ⊂ X is finely open if X \ U is thin at each point x ∈ U .It is easy to see that the finely open sets give rise to a topology, which is calledthe fine topology . Every open set is finely open, but the converse is not true ingeneral.In the definition of thinness, we make the convention that the integrand is 1whenever cap p ( B ( x, r ) , B ( x, r )) = 0. This happens e.g. if X = B ( x, r ), but neverif r < diam X . Note that thinness is a local property. Because of (4.3), thinnesscan equivalently be defined using the Cheeger variational capacity Ch-cap p . Definition 6.2. A function u : U → R , defined on a finely open set U , is finelycontinuous if it is continuous when U is equipped with the fine topology and R with the usual topology.Since every open set is finely open, the fine topology generated by the finelyopen sets is finer than the metric topology. In fact, it is the coarsest topologymaking all (Cheeger) superharmonic functions finely continuous, by J. Bj¨orn [20,Theorem 4.4], Korte [45, Theorem 4.3] and Bj¨orn–Bj¨orn–Latvala [11, Theorem 1.1].See [9, Section 11.6] and [11] for further discussion on thinness and the fine topology. 7. The Cartan, Choquet and Kellogg properties We start this section by proving the Cartan property (Theorem 1.1). The proofcombines arguments in Kilpel¨ainen–Mal´y [41, p. 155] with those in Section 2.1.5in Mal´y–Ziemer [49]. As in [41], the pointwise estimate (7.1) is essential here.However, to obtain the estimate ν k ( B j ) ≤ cap( E j , B j − ), in [41] they use the dualcharacterization of capacity as the supremum of measures on E j with potentialsbounded by 1. A similar estimate follows also from Theorem 2.45 in [49]. Here weinstead use a direct derivation of ν k ( B j ) ≤ cap( E j , B j − ) based on (5.1), Remark 5.4and Lemma 5.5. Proof of Theorem . By Lemma 4.7 in Bj¨orn–Bj¨orn–Latvala [11], we may assumethat E is open. Let B j = B ( x , r j ), r j = 2 − j , E j = E ∩ B j and u j be the Cheegercapacitary potential of E j with respect to B j − , j = 1 , , ... . As E j is open, wehave u j = 1 in E j . Let k ≥ B k < diam X , and let ν k = T u k be the Radon measure in N ,p ( B k − ) ′ ,given by Proposition 5.1.Since u k = 1 in E k , it remains to show that u k ( x ) < k . ByRemark 5.4 in Kinnunen–Martio [43] (or Proposition 8.24 in [9]), x is a Lebesguepoint of u k . Hence, Proposition 4.10 in Bj¨orn–MacManus–Shanmugalingam [21]shows that u k ( x ) ≤ c (cid:18)Z B k u pk dµ (cid:19) /p + c ∞ X j = k − (cid:18) r pj ν k ( B j ) µ ( B j ) (cid:19) / ( p − . (7.1) The first term in the right-hand side can be estimated using the Sobolev inequality[9, Theorem 5.51] and the fact that cap p ( B k , B k − ) ≃ r − pk µ ( B k ) (by Lemma 3.3 inJ. Bj¨orn [18] or Proposition 6.16 in [9]) as Z B k u pk dµ ≤ µ ( B k ) Z B k − u pk dµ ≤ Cr pk µ ( B k ) Z B k − | Du k | p dµ ≃ cap p ( E k , B k − )cap p ( B k , B k − ) . (7.2)Here we have also used (4.2) and (4.3).As for the second term in (7.1), let v j be the lsc-regularized solution of T v j = ν k | B j in B k − , j ≥ k . Lemma 5.3 shows that v j ≤ u k ≤ B k − . Thus, with v j as a test function in (5.1), we have Z B k − | Dv j | p dµ = Z B j v j dν k ≤ Z B j u k dν k . (7.3)Using Lemma 5.5 (for the first equality below) and (5.1) with u j as a test function(for the third equality) we obtain that Z B j u k dν k = Z B j u j dν k = Z B k − u j dν k | B j = Z B k − | Dv j | p − Dv j · Du j dµ ≤ (cid:18)Z B k − | Dv j | p dµ (cid:19) − /p (cid:18)Z B k − | Du j | p dµ (cid:19) /p . (7.4)Together with (7.3) this implies that Z B k − | Dv j | p dµ ≤ Z B k − | Du j | p dµ = Ch-cap p ( E j , B j − ) , where Ch-cap p denotes the Cheeger variational capacity. Inserting this into (7.4)yields, Z B j u k dν k ≤ Ch-cap p ( E j , B j − ) , which together with the last part of Lemma 5.5 and (4.3) shows that ν k ( B j ) = Z B j u k dν k ≤ Ch-cap p ( E j , B j − ) ≃ cap p ( E j , B j − ) . Hence using cap p ( B j , B j − ) ≃ r − pj µ ( B j ) again we obtain ∞ X j = k − (cid:18) r pj ν k ( B j ) µ ( B j ) (cid:19) / ( p − ≤ C ∞ X j = k − (cid:18) cap p ( E j , B j − )cap p ( B j , B j − ) (cid:19) / ( p − . (7.5)Since E is thin at x , both (7.2) and (7.5) can be made arbitrarily small by choosing k large enough. Thus u k ( x ) < k .We now turn to the proof of the Choquet property (Theorem 1.2). The followingnotation is common in the literature. The base b p E of a set E ⊂ X consists of allpoints x ∈ X where E is thick , i.e. not thin, at x . Using this notation, the Choquetproperty can be formulated as follows. Theorem 7.1. (Choquet property) For any E ⊂ X and any ε > there is an openset G so that G ∪ b p E = X and C p ( E ∩ G ) < ε. he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 13 Proof. Let { B j } ∞ j =1 be a countable covering of X by balls such that every point iscovered by arbitrarily small balls. Such a covering exists as X is separable. Choose ε > 0. For each j , let u j be the Cheeger capacitary potential of E ∩ B j withrespect to 2 B j . Since each u j is quasicontinuous, there is an open set G ′ j with C p ( G ′ j ) < − j ε such that the set G j := { x ∈ B j : u j ( x ) < } ∪ G ′ j (7.6)is open. We set G := S ∞ j =1 G j and will show that G ∪ b p E = X .Choose a point z ∈ X \ b p E . If dist( z, E \ { z } ) > 0, then there is B j ∋ z suchthat B j ∩ E is either empty or { z } . If B j ∩ E = ∅ , then u j ≡ 0. If B j ∩ E = { z } , thenthe thinness of E at z together with Proposition 1.3 in Bj¨orn–Bj¨orn–Latvala [11]shows that C p ( { z } ) = 0, and hence u j ≡ z ∈ B j ⊂ G j ⊂ G .We can therefore assume that z ∈ E \ { z } . By Theorem 1.1 (applied to E \{ z } ), there is a bounded positive Cheeger superharmonic function v in an openneighbourhood of z such that v ( z ) < < lim inf E ∋ x → z v ( x ) . Hence we may fix a ball B j ∋ z so that v is Cheeger superharmonic in 3 B j and v ≥ B j ∩ E . Since v is the lsc-regularized solution of the K v,v (2 B j )-Cheeger obstacleproblem and u j is the lsc-regularized solution of the K χ Bj ∩ E , (2 B j )-Cheeger obstacleproblem, the comparison principle in Lemma 5.4 in Bj¨orn–Bj¨orn [8] (or Lemma 8.30in [9]) yields u j ≤ v in 2 B j . It follows that u j ( z ) < 1, and thus z ∈ G j ⊂ G .It remains to prove that C p ( E ∩ G ) < ε . For any j , we have u j ≥ E ∩ B j , and thus (7.6) implies C p ( E ∩ G j ) ≤ C p ( { x ∈ E ∩ B j : u j ( x ) < } ) + C p ( G ′ j ) = C p ( G ′ j ) < − j ε. By the countable subadditivity of the capacity we obtain C p ( E ∩ G ) < ε .As a consequence of the Choquet property we can now deduce Corollary 1.3.Because of Remark 7.3 below, we find the name fine Kellogg property natural. Corollary 7.2. (Fine Kellogg property) For any E ⊂ X we have C p ( E \ b p E ) = 0 . Proof. For every ε > 0, Theorem 7.1 provides us with an open set G such that G ∪ b p E = X and C p ( E ∩ G ) < ε . Then E \ b p E ⊂ E ∩ G , and therefore C p ( E \ b p E ) <ε . Letting ε → Remark 7.3. Let Ω ⊂ X be a bounded open set with C p ( X \ Ω) > 0. Choosing E = X \ Ω in Corollary 7.2 gives C p ( ∂ Ω \ b p ( X \ Ω)) ≤ C p (( X \ Ω) \ b p ( X \ Ω)) = 0 . (7.7)On the other hand, a boundary point x ∈ ∂ Ω is regular (both for p -harmonicfunctions defined through upper gradients and for Cheeger p -harmonic functions)whenever X \ Ω is thick at x , by the sufficiency part of the Wiener criterion, seeBj¨orn–MacManus–Shanmugalingam [21] and J. Bj¨orn [19], [20] (or Theorem 11.24in [9]). Hence (7.7) yields that the set of irregular boundary points of Ω is ofcapacity zero. This result was obtained by a different method (and called theKellogg property) in Bj¨orn–Bj¨orn–Shanmugalingam [15, Theorem 3.9]. To clarify that the above proof of the Kellogg property is not using circularreasoning let us explain how the results we use here are obtained in [9]. Here weonly need results up to Chapter 9 therein plus the results in Sections 11.4 and 11.6.They in turn only rely on results up to Chapter 9 plus the implication (b) ⇒ (a) inTheorem 10.29, which can easily be obtained just using comparison. Hence we arenot relying on the Kellogg property obtained in Section 10.2 in [9]. 8. Finely open and quasiopen sets We start this section by using the Choquet property to prove Theorem 1.4, i.e. wecharacterize quasiopen sets and quasicontinuity by means of the corresponding finetopological notions. We then proceed by giving several immediate applications ofthis characterization.Note that if C p ( { x } ) = 0, then { x } is quasiopen, but not finely open. Thus thezero capacity set in Theorem 1.4 (a) cannot be dropped. Proof of Theorem . (a) That each quasiopen set U is of the form U = V ∪ E for some finely open set V and for a set E of capacity zero, was recently shown inBj¨orn–Bj¨orn–Latvala [11, Theorem 4.9].For the converse, assume that U = V ∪ E , where V is finely open and C p ( E ) = 0.Let ε > 0. By the Choquet property (Theorem 7.1), applied to X \ V , there is anopen set G such that G ∪ b p ( X \ V ) = X and C p ( G \ V ) < ε. The capacity C p is an outer capacity, by Corollary 1.3 in Bj¨orn–Bj¨orn–Shanmuga-lingam [16] (or Theorem 5.31 in [9]), so there is an open set e G ⊃ ( G \ V ) ∪ E suchthat C p ( e G ) < ε . Since V is finely open, we have V ⊂ X \ b p ( X \ V ) ⊂ G , and thus U ∪ e G = V ∪ e G = G ∪ e G is open, i.e. U is quasiopen.(b) If u is quasicontinuous, then it is finite q.e., by definition, and finely contin-uous q.e., by Theorem 4.9 in Bj¨orn–Bj¨orn–Latvala [11].Conversely, assume that there is a set Z with C p ( Z ) = 0 such that u is finiteand finely continuous on V := U \ Z . By (a), we can assume that V is finely open.Let ε > { ( a j , b j ) } ∞ j =1 be an enumeration of all open intervals with rationalendpoints and set V j := { x ∈ V : a j < u ( x ) < b j } . By the fine continuity of u , the sets V j are finely open. Hence by (a), V j arequasiopen, and thus there are open sets G j and G U with C p ( G j ) < − j ε and C p ( G U ) < ε such that V j ∪ G j and U ∪ G U are open. Also, as C p is an outercapacity, there is an open set G Z ⊃ Z with C p ( G Z ) < ε . Then G := G Z ∪ G U ∪ ∞ [ j =1 G j is open, C p ( G ) < ε , and u | U \ G is continuous since V j ∪ G are open sets.Theorem 1.4 leads directly to the following improvements of the results in Bj¨orn–Bj¨orn [10]. Corollary 8.1. Every finely open set is quasiopen, measurable and p -path open. A set U is p -path open if for p -almost every curve γ : [0 , l γ ] → X , the set γ − ( U )is (relatively) open in [0 , l γ ]. he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 15 Proof. By Theorem 1.4 (a) every finely open set is quasiopen. Hence the resultfollows from Remark 3.5 in Shanmugalingam [52] and Lemma 9.3 in [10].An important consequence is that the restriction of a minimal p -weak uppergradient to a finely open set remains minimal. This was shown for measurable p -path open sets in [10, Corollary 3.7]. We restate this result, in view of Corollary 8.1.In order to do so in full generality, we need to introduce some more notation.We define the Dirichlet space D p ( X ) = { u : u is measurable and u has an upper gradient in L p ( X ) } . As with N ,p ( X ) we assume that functions in D p ( X ) are defined everywhere (withvalues in R := [ −∞ , ∞ ]). For a measurable set E ⊂ X , the spaces D p ( E ) and D p loc ( E ) are defined similarly. For u ∈ D p loc ( E ) we denote the minimal p -weakupper gradient of u taken with E as the underlying space by g u,E . Its existence isguaranteed by Theorem 2.25 in [9]. Corollary 8.2. Let U be quasiopen and u ∈ D p loc ( X ) . Then g u,U = g u a.e. in U .In particular this holds if U is finely open.Proof. By Remark 3.5 in Shanmugalingam [52] and Lemma 9.3 in [10] every qua-siopen set is p -path open and measurable, whereas Theorem 1.4 (a) shows that everyfinely open set is quasiopen. Hence the result follows from Corollary 3.7 in [10].In [10], the fine topology turned out to be important for obstacle problems onnonopen measurable sets, i.e. when minimizing the p -energy integral Z E g u,E dµ (8.1)on an arbitrary bounded measurable set E among all functions u ∈ K ψ ,ψ ,f ( E ) := { v ∈ D p ( E ) : v − f ∈ N ,p ( E ) and ψ ≤ v ≤ ψ q.e. in E } . Knowing that finely open sets are measurable and p -path open, we are now able toimprove and simplify some of the results therein. We summarize these improvementsin the following theorem, which follows directly from [10, Theorems 1.2 and 8.3,and Corollaries 3.7 and 7.4] and Corollary 8.1. We denote the fine interior of E byfine-int E . Theorem 8.3. Let E ⊂ X be a bounded measurable set such that C p ( X \ E ) > ,and let f ∈ D p ( E ) and ψ j : E → R , j = 1 , , be such that K ψ ,ψ ,f ( E ) = ∅ . Alsolet E = fine-int E .Then K ψ ,ψ ,f ( E ) = K ψ ,ψ ,f ( E ) , and the solutions of the minimization prob-lem for (8.1) with respect to K ψ ,ψ ,f ( E ) and K ψ ,ψ ,f ( E ) coincide. Moreover, g u,E = g u,E a.e. in E and if µ ( E \ E ) = 0 then also the p -energies associatedwith these two minimization problems coincide.Furthermore, if f ∈ D p (Ω) for some open set Ω ⊃ E , then g u,E = g u,E = g u a.e. in E and the above solutions coincide with the solutions of the corresponding K ψ ′ ,ψ ′ ,f (Ω) -obstacle problem, where ψ ′ j is the extension of ψ j to Ω \ E by f , j = 1 , . We also obtain the following consequence of Lemma 3.9 and Theorem 7.3 in [10],which generalizes Theorem 2.147 and Corollary 2.162 in Mal´y–Ziemer [49] to metricspaces and to arbitrary sets. See also Remark 2.148 in [49] for another descriptionof W ,p (Ω) in R n . Proposition 8.4. (Cf. Proposition 9.4 in [10].) Let E ⊂ X be arbitrary and u ∈ N ,p ( E p ) , where E p is the fine closure of E . Then u ∈ N ,p ( E ) if and only if u = 0 q.e. on the fine boundary ∂ p E := E p \ fine-int E of E . 9. Support of capacitary measures We can now bootstrap Lemma 5.5 to quasiopen sets and in particular show thatthe capacitary measure ν E only charges the fine boundary ∂ p E := E p \ fine-int E of E , where E p is the fine closure of E . This observation seems to be new even inunweighted R n . Theorem 9.1. Let Ω be a nonempty bounded open set with C p ( X \ Ω) > . Let E ⊂ Ω , u E and ν E = T u E be as in Lemma . Let U ⊂ Ω be quasiopen and v ∈ N ,p (Ω) . Then the following are true :(a) If u ∈ N ,p (Ω) and either u is bounded from below or belongs to L ( ν E ) , and u = v q.e. in U ∩ E , then Z U u dν E = Z U v dν E . (9.1)(b) If v = 1 q.e. in U ∩ E , then ν E ( U ) = Z U v dν E = Z U u E dν E . (c) If C p ( U ∩ E ) = 0 , then ν E ( U ) = 0 . Remark 9.2. We shall see in Corollary 9.6 below that the set U ∩ E in (a), (b)and (c) above can be replaced by U ∩ ∂ p E and that the assumption v ∈ N ,p (Ω) inTheorem 9.1 can be omitted in that case.To prove Theorem 9.1 we need the following quasi-Lindel¨of principle, whoseproof in unweighted R n is given in Theorem 2.3 in Heinonen–Kilpel¨ainen–Mal´y [36].This proof, which relies on the fine Kellogg property, extends to metric spaces, seeBj¨orn–Bj¨orn–Latvala [12]. Theorem 9.3. (Quasi-Lindel¨of principle) For each family V of finely open sets thereis a countable subfamily V ′ such that C p (cid:18) [ V ∈V V \ [ V ′ ∈V ′ V ′ (cid:19) = 0 . We also need the following lemmas. Lemma 9.4. Let U be finely open and let x ∈ U . Then there exists a finely openset V ⋐ U containing x and a function v ∈ N ,p ( U ) such that v = 1 on V and ≤ v ≤ everywhere.Proof. Since U is finely open, E := X \ U is thin at x . By the Cartan property(Theorem 1.1), there are a ball B ∋ x and a lower semicontinuous finely continuous u ∈ N ,p ( B ) such that 0 ≤ u ≤ B , u ( x ) < u = 1 in E ∩ B . Let η ∈ Lip c ( B ) be such that 0 ≤ η ≤ B and η = 1 in B . Then w := η (1 − u ) ∈ N ,p ( U )is upper semicontinuous and finely continuous in X and w ( x ) = 1 − u ( x ) > v = min { , w/w ( x ) } ∈ N ,p ( U ) and V = (cid:8) x ∈ U : w ( x ) > w ( x ) (cid:9) . The finecontinuity and upper semicontinuity of w imply that V is finely open and V ⋐ U .Moreover x ∈ V and v = 1 on V . Lemma 9.5. Let U ⊂ X be quasiopen. Then U = W ∪ E = W \ E , (9.2) where W and W are Borel sets and E and E are of capacity zero. Moreover, wemay choose W to be of type F σ and W to be of type G δ . he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 17 Not all finely open sets are Borel. Let for instance V = G \ A , where G is openand A ⊂ G is a non-Borel set with C p ( A ) = 0. Then V is a non-Borel finely openset. To be more specific, we may let A ⊂ G ⊂ R n be any non-Borel set of Hausdorffdimension < n − p . Proof. By definition, for each j = 1 , , ... there is an open set G j such that U ∪ G j is open and C p ( G j ) < /j . Then U = (cid:18) U \ ∞ \ j =1 G j (cid:19) ∪ (cid:18) U ∩ ∞ \ j =1 G j (cid:19) = ∞ [ j =1 ( U \ G j ) ∪ ∞ \ j =1 ( U ∩ G j )= ∞ [ j =1 (( U ∪ G j ) \ G j ) ∪ ∞ \ j =1 ( U ∩ G j ) =: W ∪ E . The second equality in (9.2) follows by choosing W = T ∞ j =1 ( U ∪ G j ) and E = W \ U . The last two claims follow from the choices above. Proof of Theorem . By Theorem 1.4, we can find a finely open set V ⊂ U suchthat C p ( U \ V ) = 0. For every x ∈ V , Lemma 9.4 provides us with a finely openset V x ⋐ V containing x and a function v x ∈ N ,p ( V ) such that v x = 1 on V x and 0 ≤ v x ≤ C p ( U \ V ) = 0, we can out of these choose V j = V x j and v j = v x j , j = 1 , , ... , sothat U = S ∞ j =1 V j ∪ Z , where C p ( Z ) = 0. For k = 1 , , ... , set η k = χ X \ Z max j =1 , ,...,k v j ∈ N ,p ( U ) . Since ν E is a complete Borel measure which, by Lemma 5.5 (or Lemma 3.8 inBj¨orn–MacManus–Shanmugalingam [21]), is absolutely continuous with respect tothe capacity C p , it follows from Lemma 9.5 that U is ν E -measurable and ν E ( Z ) = 0.We are now ready to prove (a)–(c).(a) First, assume that u and v are bounded. Then η k ( u − v ) ∈ N ,p ( U ), byLemma 2.6, and Lemma 2.37 in Bj¨orn–Bj¨orn [9] shows that η k ( u − v ) ∈ N ,p ( U ).Since u = v q.e. in U ∩ E , it follows that η k ( u − v ) ∈ N ,p ( U \ E ). Hence (5.5)yields that Z U η k ( u − v ) dν E = 0 . Since η k ր χ U \ Z in U , dominated convergence and the fact that ν E ( Z ) = 0 implythat Z U ( u − v ) dν E = Z U \ Z ( u − v ) dν E = 0 , and (9.1) follows.Next, assume that u and v are bounded from below. Then, by monotone con-vergence and the bounded case, Z U u dν E = lim k →∞ Z U min { u, k } dν E = lim k →∞ Z U min { v, k } dν E = Z U v dν E . Finally, applying this to the positive and negative parts of u and v gives Z U u + dν E = Z U v + dν E and Z U u − dν E = Z U v − dν E , and hence Z U u dν E = Z U u + dν E − Z U u − dν E = Z U v + dν E − Z U v − dν E = Z U v dν E , where the assumptions on u guarantee that the subtractions are well defined (i.e.not ∞ − ∞ ).(b) By applying (a) to u = u E and v we have R U v dν E = R U u E dν E . Choosing v ≡ ν E ( U ) = R U u E dν E .(c) This follows by applying (b) to v ≡ Corollary 9.6. Let Ω , E , u E and ν E be as in Theorem . Then ν E (Ω \ ∂ p E ) = 0 , i.e. ν E is supported on the fine boundary ∂ p E := E p \ fine-int E of E .Proof. First, the fine exterior V = Ω \ E p is finely open and V ∩ E = ∅ , whence ν E ( V ) = 0 by Theorem 9.1 (c).Next, the fine interior E := fine-int E is finely open and as in the proof of The-orem 9.1 we can use the quasi-Lindel¨of principle to find nonnegative η k ∈ N ,p ( E )such that η k ր χ E \ Z as k → ∞ , where C p ( Z ) = 0. Since u E = 1 q.e. in E , we have Du E = 0 a.e. in E and henceby (5.1) Z Ω η k dν E = Z Ω | Du E | p − Du E · Dη k dµ = 0 . Dominated convergence then shows that ν E ( E \ Z ) = 0. Since ν E ( Z ) = 0 byLemma 5.5 (or Lemma 3.8 in [21]), the proof is complete. Corollary 9.7. Let Ω , E ⊂ Ω , u E and ν E = T u E be as in Lemma . Let U ⊂ Ω be quasiopen. Then the following are true :(a) If u is a function on Ω such that R U ∩ ∂ p E u dν E is well-defined and v is afunction on U such that v = u q.e. in U ∩ ∂ p E , then Z U v dν E = Z U u dν E . (b) If v = 1 q.e. in U ∩ ∂ p E , then ν E ( U ) = Z U v dν E = Z U u E dν E . (c) If C p ( U ∩ ∂ p E ) = 0 , then ν E ( U ) = 0 .Proof. (c) This follows directly from Corollary 9.6 and the fact that ν E is absolutelycontinuous with respect to the capacity C p (by Lemma 5.5).(a) By Corollary 9.6 and the absolute continuity of ν E with respect to C p again,we see that Z U v dν E = Z U ∩ ∂ p E v dν E = Z U ∩ ∂ p E u dν E = Z U u dν E . (b) This follows from (a) by choosing u ≡ u = u E , respectively.We end with a simple example showing that the fine boundary can be muchsmaller than the metric boundary. A much more involved example in the samespirit is given in Section 9 in Bj¨orn–Bj¨orn [10]. Example 9.8. Let B be an open ball in R n , 1 < p ≤ n , and let E = B \ Q n . Theset E is finely open and has fine closure E p = B . Hence ∂ p E = ∂B ∪ ( B ∩ Q n ),while ∂E = B . he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 19 References Adams, D. R. and Hedberg, L. I., Function Spaces and Potential Theory, Springer, Berlin–Heidelberg, 1996.2. Adams, D. R. and Lewis, J. L., Fine and quasiconnectedness in nonlinearpotential theory, Ann. Inst. Fourier ( Grenoble ) :1 (1985), 57–73.3. Ambrosio, L., Colombo, M. and Di Marino, S., Sobolev spaces in metricmeasure spaces: reflexivity and lower semicontinuity of slope, Preprint , 2012, arXiv:1212.3779 .4. Ambrosio, L., Gigli, N. and Savar´e, G., Density of Lipschitz functions andequivalence of weak gradients in metric measure spaces, Rev. Mat. Iberoam. (2013), 969–986.5. Bj¨orn, A., Characterizations of p -superharmonic functions on metric spaces, Studia Math. (2005), 45–62.6. Bj¨orn, A., A weak Kellogg property for quasiminimizers, Comment. Math.Helv. (2006), 809–825.7. Bj¨orn, A., Removable singularities for bounded p -harmonic and quasi(super)-harmonic functions on metric spaces, Ann. Acad. Sci. Fenn. Math. (2006),71–95.8. Bj¨orn, A. and Bj¨orn, J., Boundary regularity for p -harmonic functions andsolutions of the obstacle problem, J. Math. Soc. Japan (2006), 1211–1232.9. Bj¨orn, A. and Bj¨orn, J., Nonlinear Potential Theory on Metric Spaces, EMSTracts in Mathematics , European Math. Soc., Z¨urich, 2011.10. Bj¨orn, A. and Bj¨orn, J., Obstacle and Dirichlet problems on arbitrarynonopen sets in metric spaces, and fine topology, to appear in Rev. Mat.Iberoam. Bj¨orn, A., Bj¨orn, J. and Latvala, V., The weak Cartan property for the p -fine topology on metric spaces, to appear in Indiana Univ. Math. J. Bj¨orn, A., Bj¨orn, J. and Latvala, V., Quasicontinuity, fine gradients andSobolev spaces on quasiopen sets, In preparation .13. Bj¨orn, A., Bj¨orn, J. and Mal´y, J., Quasiopen and p -path open sets, andcharacterizations of quasicontinuity, In preparation .14. Bj¨orn, A., Bj¨orn, J. and Parviainen, M., Lebesgue points and the fun-damental convergence theorem for superharmonic functions on metric spaces, Rev. Mat. Iberoam. (2010), 147–174.15. Bj¨orn, A., Bj¨orn, J. and Shanmugalingam, N., The Dirichlet problemfor p -harmonic functions on metric spaces, J. Reine Angew. Math. (2003),173–203.16. Bj¨orn, A., Bj¨orn, J. and Shanmugalingam, N., Quasicontinuity ofNewton–Sobolev functions and density of Lipschitz functions on metric spaces, Houston J. Math. (2008), 1197–1211.17. Bj¨orn, A. and Marola, N., Moser iteration for (quasi)minimizers on metricspaces, Manuscripta Math. (2006), 339–366.18. Bj¨orn, J., Boundary continuity for quasiminimizers on metric spaces, IllinoisJ. Math. (2002), 383–403.19. Bj¨orn, J., Wiener criterion for Cheeger p -harmonic functions on metric spaces,in Potential Theory in Matsue , Advanced Studies in Pure Mathematics , pp.103–115, Mathematical Society of Japan, Tokyo, 2006.20. Bj¨orn, J., Fine continuity on metric spaces, Manuscripta Math. (2008),369–381.21. Bj¨orn, J., MacManus, P. and Shanmugalingam, N., Fat sets and point-wise boundary estimates for p -harmonic functions in metric spaces, J. Anal.Math. (2001), 339–369. Brelot, M., Sur la th´eorie moderne du potentiel, C. R. Acad. Sci. Paris (1939), 828–830.23. Brelot, M., Points irr´eguliers et transformations continues en th´eorie du po-tentiel, J. Math. Pures Appl. (1940), 319–337.24. Brelot, M., Sur les ensembles effils, Bull. Sci. Math. (1944), 12–36.25. Brelot, M., On Topologies and Boundaries in Potential Theory, LectureNotes in Math. , Springer, Berlin–Heidelberg, 1971.26. Cartan, H., Th´eorie g´en´erale du balayage en potentiel newtonien, Ann. Univ.Grenoble. Sect. Sci. Math. Phys. (1946), 221–280.27. Cheeger, J., Differentiability of Lipschitz functions on metric measure spaces, Geom. Funct. Anal. (1999), 428–517.28. Choquet, G., Sur les points d’effilement d’un ensemble. Application `a l’´etudede la capacit´e, Ann. Inst. Fourier ( Grenoble ) (1959), 91–101.29. Franchi, B., Haj lasz, P. and Koskela, P., Definitions of Sobolev classeson metric spaces, Ann. Inst. Fourier ( Grenoble ) (1999), 1903–1924.30. Fuglede, B., The quasi topology associated with a countably subadditive setfunction, Ann. Inst. Fourier ( Grenoble ) :1 (1971), 123–169.31. Fuglede, B., Finely Harmonic Functions, Springer, Berlin–New York, 1972.32. Haj lasz, P. and Koskela, P., Sobolev met Poincar´e, Mem. Amer. Math.Soc. :688 (2000).33. Hedberg, L. I., Non-linear potentials and approximation in the mean by an-alytic functions, Math. Z. (1972), 299–319.34. Hedberg, L. I. and Wolff, T. H., Thin sets in nonlinear potential theory, Ann. Inst. Fourier ( Grenoble ) :4 (1983), 161–187.35. Heinonen, J., Lectures on Analysis on Metric Spaces, Springer, New York,2001.36. Heinonen, J., Kilpel¨ainen, T. and Mal´y, J., Connectedness in fine topolo-gies, Ann. Acad. Sci. Fenn. Ser. A I Math. (1990), 107–123.37. Heinonen, J., Kilpel¨ainen, T. and Martio, O., Nonlinear Potential Theoryof Degenerate Elliptic Equations, Heinonen, J. and Koskela, P., Quasiconformal maps in metric spaces withcontrolled geometry, Acta Math. (1998), 1–61.39. Heinonen, J., Koskela, P., Shanmugalingam, N. and Tyson, J. T., Sobolev Spaces on Metric Measure Spaces : an Approach Based on Upper Gra-dients, In preparation.40. Keith, S., Measurable differentiable structures and the Poincar´e inequality, Indiana Univ. Math. J. (2004), 1127–1150.41. Kilpel¨ainen, T. and Mal´y, J., The Wiener test and potential estimates forquasilinear elliptic equations, Acta Math. (1994), 137–161.42. Kinnunen, J. and Latvala, V., Fine regularity of superharmonic functionson metric spaces, in Future Trends in Geometric Function Theory RNC Work-shop Jyv¨askyl¨a 2003 , Rep. Univ. Jyv¨askyl¨a Dep. Math. Stat. , pp. 157–167,University of Jyv¨askyl¨a, Jyv¨askyl¨a, 2003.43. Kinnunen, J. and Martio, O., Nonlinear potential theory on metric spaces, Illinois Math. J. (2002), 857–883.44. Kinnunen, J. and Shanmugalingam, N., Regularity of quasi-minimizers onmetric spaces, Manuscripta Math. (2001), 401–423.45. Korte, R., A Caccioppoli estimate and fine continuity for superminimizers onmetric spaces, Ann. Acad. Sci. Fenn. Math. (2008), 597–604.46. Koskela, P. and MacManus, P., Quasiconformal mappings and Sobolevspaces, Studia Math. (1998), 1–17.47. Latvala, V., Finely Superharmonic Functions of Degenerate Elliptic Equa-tions, Ann. Acad. Sci. Fenn. Ser. A I Math. Dissertationes (1994). he Cartan, Choquet and Kellogg properties for the fine topology on metric spaces 21 Lukeˇs, J., Mal´y, J. and Zaj´ıˇcek, L., Fine Topology Methods in Real Anal-ysis and Potential Theory, Lecture Notes in Math. , Springer, Berlin–Heidelberg, 1986.49. Mal´y, J. and Ziemer, W. P., Fine Regularity of Solutions of Elliptic PartialDifferential Equations, Amer. Math. Soc., Providence, RI, 1997.50. Mikkonen, P., On the Wolff Potential and Quasilinear Elliptic Equations In-volving Measures, Ann. Acad. Sci. Fenn. Math. Diss. (1996).51. Shanmugalingam, N., Newtonian spaces: An extension of Sobolev spaces tometric measure spaces, Rev. Mat. Iberoam. (2000), 243–279.52. Shanmugalingam, N., Harmonic functions on metric spaces, Illinois J. Math. (2001), 1021–1050.53. Wiener, N., The Dirichlet problem, J. Math. Phys.3