N=2 Einstein-Yang-Mills's BPS solutions
Mechthild Huebscher, Patrick Meessen, Tomas Ortin, Silvia Vaula

TL;DR
This paper characterizes all supersymmetric solutions in N=2, d=4 Einstein-Yang-Mills theories, linking known flat solutions to supergravity configurations, including monopoles and black holes with non-Abelian hair, and discusses the attractor mechanism.
Contribution
It provides the complete classification of supersymmetric solutions in N=2 Einstein-Yang-Mills theories, connecting flat space solutions to supergravity embeddings and analyzing regular and irregular solutions.
Findings
Solutions constructed from flat spacetime Bogomol'nyi solutions
Regular monopole embeddings in supergravity
Black holes with non-Abelian hair
Abstract
We find the general form of all the supersymmetric configurations and solutions of N=2,d=4 Einstein-Yang-Mills theories. In the timelike case, which we study in great detail, giving many examples, the solutions to the full supergravity equations can be constructed from known flat spacetime solutions of the Bogomol'nyi equations. This allows the regular supersymmetric embedding in supergravity of regular monopole solutions ('t Hooft-Poyakov's, Weinberg's, Wilkinson and Bais's) but also embeddings of irregular solutions to the Bogomol'nyi equations which turn out to be regular black holes with different forms of non-Abelian hair once the non-triviality of the spacetime metric is taken into account. The attractor mechanism is realized in a gauge-covariant way. In the null case we determine the general equations that supersymmetric configurations and solutions must satisfy but we do not…
Click any figure to enlarge with its caption.
Figure 1
Figure 2Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
IFT-UAM/CSIC-08-32
arXiv:0806.1477
June ,
** Einstein-Yang-Mills’s BPS solutions**
Mechthild Hübscher 111E-mail: [email protected], Patrick Meessen 222E-mail: [email protected], Tomás Ortín 333E-mail: [email protected] and Silvia Vaulà 444E-mail: [email protected]
*Instituto de Física Teórica UAM/CSIC
Facultad de Ciencias C-XVI, C.U. Cantoblanco, E-28049-Madrid, Spain
Abstract
We find the general form of all the supersymmetric configurations and solutions of Einstein-Yang-Mills theories. In the timelike case, which we study in great detail, giving many examples, the solutions to the full supergravity equations can be constructed from known flat spacetime solutions of the Bogomol’nyi equations. This allows the regular supersymmetric embedding in supergravity of regular monopole solutions (’t Hooft-Poyakov’s, Weinberg’s, Wilkinson and Bais’s) but also embeddings of irregular solutions to the Bogomol’nyi equations which turn out to be regular black holes with different forms of non-Abelian hair once the non-triviality of the spacetime metric is taken into account. The attractor mechanism is realized in a gauge-covariant way.
In the null case we determine the general equations that supersymmetric configurations and solutions must satisfy but we do not find relevant new supersymmetric solutions.
Contents
-
5.1 Killing equations for the vector bilinears and first consequences
-
A Gauging holomorphic isometries of special Kähler manifolds
Introduction
Supersymmetric solutions of supergravity theories are playing a curcial rôle in may of the developments that Superstring Theory has seen in the last few years. The knowledge of all the possible solutions can lead to new interesting models from which we can learn more about the possible vacua of the theory, their potential holographic relations with CFTs etc. Achieving a complete characterization and classification of all the supersymmetric solutions of supergravity theories is, thus, an important goal with may potential spin-offs.
Most of the work done so far in this subject has been focussed on higher-dimensional ungauged theories. The 4-dimensional theories are equally interesting, though, since they admit solutions such as the much-studied families of charged extreme black holes found in Refs. [1] in ungauged supergravity and in Refs. [2, 3] in ungauged supergravity coupled to vector multiplets.
The systematic study and classification of supersymmetric solutions of 4-dimensional supergravities was pioneered by Tod 25 years ago in Ref. [4], in which he completeley solved the problem in pure, ungauged, supergravity. Apart from another work on supergravity [5], the subject was not reanimated until quite recently: the problem was solved for pure, gauged supergravity in Refs. [6, 7, 8], for ungauged supergravity coupled to vector supermultiplets in Ref. [9], and for the same theory with a gauging in Ref. [10]. Ungauged supergravity coupled to vector supermultiplets and hypermultiplets was dealt with in Ref. [11]. Finally, the problem was solved for pure, ungauged supergravity in [12] and for matter-coupled supergravity in Ref. [13] for the ungauged case (without superpotential but with non-trivial kinetic matrix) and in Ref. [14] for the gauged case with superpotential but without kinetic matrix.
The cases considered so far (the above list) only include non-Abelian gauge groups in the case, which does not admit supersymmetric black-hole-type nor static monopole-like solutions. They can only exist in theories. We are, therefore, led to consider theories with non-Abelian gaugings. Some interesting non-Abelian monopole solutions are known in gauged supergravity (namely, the Chamseddine-Volkov monopole [15]) and similar solutions must exist in theories with non-Abelian gaugings, of which there is a much wider variety.
In this paper we are going to study the classification of the supersymmetric solutions of supergravity coupled to vector multiplets with non-Abelian gaugins of the special-Kähler manifold (that we will call super-Einstein-Yang-Mills (SEYM) theories for short) with the aim of finding non-Abelian generalizations of the known supersymmetric extreme black holes of the ungauged theories [3] and supersymmetric embeddings of YM monopoles in supergravity. We will present the full classification of the general solutions and the explicit construction of several examples of the kind of solutions we were searching for: non-Abelian monopoles and black-holes555 A few examples have been published in Refs. [16, 17]. . We will actually give a recipe (see Section 3.5) which allows the (not always regular) embedding into supergravity of virtually any solution to the Bogomol’nyi equation [18].
While the existence of the monopoles was expected due to the existence of the globally regular Chamseddine-Volkov [15] and Harvey-Liu [19] monopole solutions, the existence of regular extreme black-holes with non-trivial non-Abelian hair is a bit more surprising given the existence of a non-Abelian baldness theorem [20]666 See the review paper [21] for further on this subject.
that states that all the regular black-hole solutions of the Einstein-Yang-Mills theory with colour charges are actually embeddings of solutions with Abelian charges. The truly non-Abelian solutions of the EYM theory (the Bartnik-McKinnon particle [22] and its black hole generalizations [23]), which are known only numerically, do not have any asymptotic gauge charges. By contrast, some of our solutions, which are fully analytical, do have genuinely non-Abelian charges at infinity. Some of our solution also have non-Abelian hair that does not result into any gauge charges at infinity. It is evident that the non-Abelian baldness theorems do not apply to SEYM theories, which have a different matter content, one in which the scalars play a prominent rôle.
One of the most interesting aspects of the supersymmetric black holes of ungauged supergravity is the existence of the attractor mechanism for the values of the scalars [2, 24]: independently of their asymptotic values, the values of the scalars on the event horizon are fully determined by the conserved charges. As a result, the Bekenstein-Hawking entropy only depends on conserved charges which is, by itself, a strong indication that it admits a microscopic interpretation. It is of utmost interest, then, to study if and how the attractor mechanism works for the supersymmetric non-Abelian black holes in these theories. Our answer will be positive in a properly generalized way.
The plan of this article is as follows: in Section 1 we will review gauged supergravity without hypermultiplets (to which we shall refer as SEYM), leaving information about isometries in Special Geometry and their implementation in supergravity for the Appendix A. In Section 2 we shall discuss the generic characteristics of the supersymmetric solutions, such as the Killing Spinor Equations and their implications for the equations of motion. In Section 3, we shall characterize the solutions in the timelike case obtaining the minimal set of equations that need to be solved in order to have supersymmetric solutions to SEYM. Section 3.5 contains the step-by-step procedure to construct supersymmetric solutions of the theory starting with a solution of the Bogomol’nyi equations on and which we will use in Section 4 to construct and study different examples of solutions belonging to this class. These solutions split up into globally regular monopoles and black holes. Appendices B and C contain some complementary information needed for Sec. (4). In Section 5 we solve the null case. A discussion of our results and our conclusions are contained in Section 6.
1 Gauged supergravity coupled to vector supermultiplets
In this section we shall describe the action, equations of motion and supersymmetry transformation rules of gauged supergravity coupled to vector multiplets. In order to make this description brief, we only discuss the differences with the ungauged case, which is described in detail in Ref. [11]. Some definitions and formulae related to the gauging of holomorphic isometries of special Kähler manifolds are contained in Appendix A. We also refer the reader to Ref. [25], the review Ref. [26], and the original works Refs. [27, 28] for more information.
The action restricted to the bosonic fields of these theories is
[TABLE]
where the potential , is given by
[TABLE]
where
[TABLE]
In these expressions is the gauge coupling constant, the are holomorphic Killing vectors of and the gauge covariant derivative (also Kähler-covariant when acting on fields of non-trivial Kähler weight) and is defined in Appendix A.
This is not the most general gauged supergravity: if the matrices that provide a representation of the Lie algebra of the gauge group , see Eq. (A.26), are written in the form
[TABLE]
we are then considering only the cases in which , so that only symmetries of the action are gauged, and . This last restriction is only made for the sake of simplicity as theories in which symmetries with are gauged have complicated Chern-Simons terms.
Within this restricted class of theories, then, we can use Eqs. (A.45) and (A.47) to rewrite the potential as
[TABLE]
Then, since is negative definite and the momentum map is real, the potential is positive semi-definite . For constant values of the scalars behaves as a non-negative cosmological constant which leads to Minkowski () or () vacua. The latter cannot be maximally supersymmetric, however.
For convenience, we denote the bosonic equations of motion by
[TABLE]
and the Bianchi identities for the vector field strengths by
[TABLE]
Then, using the action Eq. (1.1), we find
[TABLE]
In differential-form notation, the Maxwell equation takes the form
[TABLE]
For vanishing fermions, the supersymmetry transformation rules of the fermions are
[TABLE]
is given in Eq. (A.39).
The supersymmetry transformations of the bosons are the same as in the ungauged case
[TABLE]
2 Supersymmetric configurations: general setup
Our first goal is to find all the bosonic field configurations for which the Killing spinor equations (KSEs):
[TABLE]
admit at least one solution.
Our second goal will be to identify among all the supersymmetric field configurations those that satisfy all the equations of motion (including the Bianchi identities).
Let us initiate the analysis of the KSEs by studying their integrability conditions.
2.1 Killing Spinor Identities (KSIs)
The off-shell equations of motion of the bosonic fields of bosonic supersymmetric configurations satisfy certain relations known as (Killing spinor identities, KSIs) [29, 30]. If we assume that the Bianchi identities are always identically satisfied everywhere, the KSIs only depend on the supersymmetry transformation rules of the bosonic fields. These are identical for the gauged and ungauged theories, implying that their KSIs are also identical. If we do not assume that the Bianchi identities are identically satisfied everywhere, then they also occur in the KSIs, which now have to be found via the integrability conditions of the KSEs. In the ungauged case they occur in symplectic-invariant combinations, as one would expect, and take the form [9]
[TABLE]
where
[TABLE]
We have checked through explicit computation that these relations remain valid in the non-Abelian gauged case at hand.
Taking products of these expressions with Killing spinors and gamma matrices, one can derive KSIs involving the bosonic equations and tensors constructed as bilinears of the commuting Killing spinors.777See the appendix in Ref. [12] for the definitions and properties of these bilinears. In the case in which the bilinear is a timelike vector (referred to as the timelike case), one obtains [31] the following identities (w.r.t. an orthonormal frame with )
[TABLE]
where and is non-zero in the timelike case.
As discussed in Ref. [31], these identities contain a great deal of physical information. In this paper we shall exploit only one fact, namely the fact that if the Maxwell equation and the Bianchi identity are satisfied for a supersymmetric configuration, then so are the rest of the equations of motion. The strategy to be followed is, therefore, to first identify the supersymmetric configurations and impose the Maxwell equations and the Bianchi identities. This will lead to some differential equations that need be solved in order to construct a supersymmetric solution.
In the case in which is a null vector (the null case), renaming it as for reasons of clarity, one gets
[TABLE]
where is a null tetrad constructed with the Killing spinor and an auxiliary spinor as explained in Ref. [9].
These identities imply that the only independent equations of motion that one has to check on supersymmetric configurations are and . As before, these are the equations that need to be imposed in order for a supersymmetric configuration to be a supersymmetric solution.
2.2 Killing equations for the bilinears
In order to find the most general background admitting a solution to the KSEs, Eqs. (2.1) and (2.2), we shall assume that the background admits one Killing spinor. Using this assumption we will derive consistency conditions that the background must satisfy, after which we will prove that these necessary conditions are also sufficient.
It is convenient to work with spinor bilinears, and consequently we start by deriving equations for these bilinears by contracting the KSEs with gamma matrices and Killing spinors.
From the gravitino supersymmetry transformation rule Eq. (1.12) we get the independent equations
[TABLE]
which have the same functional form as their equivalents in the ungauged case. Hence, as in the ungauged case, is a Killing vector and the 1-form satisfies the equation
[TABLE]
The remaining 3 independent 1-forms ( and the are the Pauli matrices) are exact, i.e.
[TABLE]
From the gauginos’ supersymmetry transformation rules, Eqs. (1.13), we obtain
[TABLE]
The trace of the first equation gives
[TABLE]
while the antisymmetric part of the second equation gives
[TABLE]
The well-known special geometry completeness relation implies that
[TABLE]
which allows us to combine Eqs. (2.14) and (2.22), as to obtain
[TABLE]
Multiplying this equation by and using Eq. (2.21), we find
[TABLE]
At this point in the investigation, it is convenient to take into account the norm of the Killing vector : we shall investigate the timelike case in Section 3 and the null case in Section 5.
3 The timelike case
3.1 The vector field strengths
As is well-known, the contraction of the (anti-) self-dual part of a 2-form with a non-null vector, such as in the current timelike case, completely determines the 2-form, i.e.
[TABLE]
As is given by Eq. (2.24), the vector field strengths are written in terms of the scalars , and the vector . Observe that the component of proportional to is projected out in this formula: this implies that the field strengths have the same functional form as in the ungauged case. The covariant derivatives that appear in the r.h.s., however, contain explicitly the vector potentials.
The next item on the list is the determination of the spacetime metric:
3.2 The metric
As in the ungauged case we define a time coordinate by
[TABLE]
Unlike the ungauged case, however, the scalars in a supersymmetric configuration need not automatically be time-independent: with respect to the chosen -coordinate Eq. (2.21) takes the form
[TABLE]
It is convenient to choose a gauge in which the complex fields are time-independent, and one accomplishing just that is
[TABLE]
This gauge choice reduces Eq. (3.3) to
[TABLE]
on account of Eq. (A.46). It should be pointed out that this gauge choice is identical to the expression for obtained in ungauged case in Refs. [9, 11]. Further, using the above -independence and gauge choice in Eq. (2.25), we can derive
[TABLE]
where we made use of Eq. (A.45) and the reality of . Thus, with the standard coordinate choice and the gauge choice (3.4) the scalars and are time-independent.
Using the exactness of the 1-forms to define spacelike coordinates by
[TABLE]
the metric takes on the form
[TABLE]
where is a time-independent 1-form. This 1-form is determined by the following condition
[TABLE]
Observe that this equation has, apart from a different definition of the covariant derivative, the same functional form as in the ungauged case; before we start rewriting the above result in order to get to the desired result, however, we would like to point out that due to the stationary character of the metric, the resulting covariant derivatives on the transverse contain a piece proportional to . The end-effect of this pull-back is that we introduce a new connection on , denoted by , which is formally the same as but for a redefinition of the gauge field, i.e.
[TABLE]
In order to compare the results in this article with the ones found in [9], we introduce the real symplectic sections and defined by
[TABLE]
is the symplectic section defining special geometry and thence satisfies
[TABLE]
This then implies that our gauge choice can be expressed in the form
[TABLE]
and that the metric function can be written as
[TABLE]
Similar to the ungauged case, we can then rewrite Eq. (3.9) as
[TABLE]
whose integrability condition reads
[TABLE]
and we shall see that, apart from possible singularities [32, 31], the integrability condition is identically satisfied for supersymmetric solutions.
3.3 Solving the Killing spinor equations
In the previous sections we have found that timelike supersymmetric configurations have a metric and vector field strengths given by Eqs. (3.8,2.24) and (3.1) in terms of the scalars . It is easy to see that all configurations of this form admit spinors that satisfy the Killing spinor equations (2.1,2.2). The Killing spinors have exactly the same form as in the ungauged case [9]
[TABLE]
We conclude that we have identified all the supersymmetric configurations of the theory.
3.4 Equations of motion
The results of Section 2.1 imply that in order to have a classical solution, we only need to impose the Maxwell equations and Bianchi identities on the supersymmetric configurations. In this section, then, we will discuss the differential equations arrising from the applying the Maxwell and Bianchi equations on the supersymmetric configurations obtained thus far.
As we mentioned in Section 3.1 the field strengths of supersymmetric configurations take the same form as in the ungauged case [9] with the Kähler-covariant derivatives replaced by Kähler- and -covariant derivatives. Therefore, the symplectic vector of field strengths and dual field strengths takes the form
[TABLE]
Operating in the first term we can rewrite it in the form
[TABLE]
and using the equation of 1-form , Eq. (3.9), which is also identical to that of the ungauged case with the same substitution of covariant derivatives, we arrive at
[TABLE]
In what follows we shall use the following Vierbein () and the corresponding directional derivatives (), normalized as , that are given by
[TABLE]
With respect to this basis we
[TABLE]
and the gauge fixing (3.4) and the constraint (3.3) read
[TABLE]
The equation that the spacelike components of the field strengths satisfy can be rewritten in the form
[TABLE]
where the tilde indicates that the gauge field that appears in this equation is the combination defined in Eq. (3.10).
This equation is easily recognized as the well-known Bogomol’nyi equation [18] for the connection and the real “Higgs” field on . Its integrability condition uses the Bianchi identity for the 3-dimensional gauge connection and, as it turns out, is equivalent to the complete Bianchi identity for the 4-dimensional gauge connection . It takes the form
[TABLE]
Taking the Maxwell equation in form notation Eq. (1.11) and using heavily the formulae in Appendix A we find that all the components are satisfied (as implied by the KSIs) except for one which leads to the equation
[TABLE]
Plugging the above equation and the Bianchi identity (3.25) into the integrability condition for , Eq. (3.16), leads to
[TABLE]
which is, ignoring possible singularities, therefore identically satisfied.
3.5 Construction of supersymmetric solutions of SEYM
According to the KSIs, the supersymmetric configurations that satisfy the pair of Eqs. (3.25) and (3.26), or, equivalently, the pair of Eqs. (3.24) and (3.26) solve all the equations of motion of the theory. This implies that one can give a step-by-step prescription to construct supersymmetric solutions of any SEYM starting from any solution of the YM-Higgs Bogomol’nyi equations on :
Take a solution , to the equations
[TABLE]
As we have stressed repeatedly, these equations are nothing but YM-Higgs Bogomol’nyi equations on and there are plenty of solutions available in the literature. However, since in most cases the authors’ goal is to obtain regular monopole solutions on , there are many solutions to the same equations that have been discarded because they present singularities. We know, however, that in the Abelian case, the singularities might be hidden by an event horizon888More precisely they turn out to be coordinate singularities in the full spacetime and correspond, not to a singular point, but to an event horizon.. Therefore, we will not require the solutions to the Bogomol’nyi equations to be globally regular on . 2. 2.
Given the solution , , Eq. (3.26), which we write here again for the sake of clarity (as we will do with other relevant equations):
[TABLE]
becomes a linear equation for the s alone which has to be solved. For compact gauge groups a possible solution is
[TABLE]
for an arbitrary real constant (the r.h.s. of Eq. (3.26) vanishes for this Ansatz). 3. 3.
The first two steps provide . The next step, then, is to obtain as functions of by solving the model-dependent stabilization equations. The stabilization equations depend only on the specific model one is considering and does not depend on whether the model is gauged or not. 4. 4.
Given and , one can compute the metric function using Eq. (3.14)
[TABLE]
the physical complex scalars by
[TABLE]
and the metric 1-form using Eq. (3.15)
[TABLE]
This last equation can always be solved locally, as according to Eq. (3.27) its integrability equation is solved automatically, at least locally: Since the solutions to the covariant Laplace equations are usually local (they generically have singularities), the integrability condition may fail to be satisfied everywhere, as discussed for example in Refs. [32, 33, 31], leading to singularities in the metric. The solution Eq. (3.28), however, always leads to exactly vanishing , whence to static solutions.
and completely determine the metric of the supersymmetric solutions, given in Eq. (3.8)
[TABLE] 5. 5.
Once and have been determined, the 4-dimensional gauge potential can be found from Eq. (3.13)
[TABLE]
and from the definition of Eq. (3.10)
[TABLE]
The procedure we have followed ensures that this is the gauge potential whose field strength is given in Eq. (3.20).
In the next section we are going to construct, following this procedure, several solutions.
4 Monopoles and hairy black holes
As we have seen, the starting point in the construction of SEYM supersymmetric solutions is the Bogomol’nyi equation on . Of course, the most interesting solutions to the Bogomol’nyi equations are the monopoles that can be characterised by saying that they are finite energy solutions that are everywhere regular. The fact that the gauge fields are regular does, however, not imply that the full supergravity solution is regular. Indeed, the metric and the physical scalar fields are built out of the “Higgs field”, i.e. , and the precise relations are model dependent and requires knowing the solutions to the stabilization equation.
As the Higgs field in a monopole asymptotes to a non-trivial constant configuration, it asymptotically breaks the gauge group through the Higgs effect. In fact, as we are dealing with supergravity and supersymmetry preserving solutions, monopoles in our setting would have to implement the super-Higgs effect as for example discussed in Refs. [34]. If we were to insist on an asymptotic supersymmetric effective action, we would be forced to introduce hypermultiplets in order to fill out massive supermultiplets, but this point will not be pursued in this article.
The Bogomol’nyi equations admit more than just regular solutions, and we shall give families of solutions, labelled by a continuous parameter , having the same asymptotic behaviour as the monopole solutions. As they are singular on , however, we will use them to construct metrics describing the regions outside regular black holes: as will be shown, the members of a given family lead to black holes that are not distinguished by their asymptotic data, such as the moduli or the asymptotic mass, nor by their entropy and as such illustrate the non-applicability of the no-hair theorem to supersymmetric EYM theories. Furthermore, in all examples considered, the attractor mechanisms is at work, meaning that the physical scalars at the horizon and the entropy depend only on the asymptotic charges and not on the moduli nor on the parameter .
The plan of this section is as follows: in section (4.1) we shall repeat briefly the embedding of the spherically symmetric solutions to the Bogomol’nyi equations in the models. In all but one of these solutions, the asymptotic gauge symmetry breaking is maximal, i.e. the gauge symmetry is broken down to . In section (4.2), we will investigate the embedding of solutions that manifest a non-maximal asymptotic symmetry breaking: for this we take E. Weinberg’s spherically symmetric -monopole [35] embedded into . This monopole breaks the down to and has the added characteristic that, unlike the ’t Hooft-Polyakov monopole, the Higgs field does not vanish at the origin.
An interesting question is whether one can embed monopoles also into more complicated models. This question will be investigated in Section 4.3, where we consider gauged “Magic” supergravities.
4.1 Spherically symmetric solutions in gauged
Before discussing the solutions we need to make some comments on the model: the model we shall consider in this and the next section is the so-called model.999 The solutions in this and the next section can also be embedded into the -models, with similar conclusions. Contrary to Ref. [16], however, we have chosen not to deal with this model explicitly, and refer the reader to Appendix B for more details. In this model the metric on the scalar manifold is that of the symmetric space and the prepotential is given by
[TABLE]
which is manifestly invariant.
The Kähler potential is straightforwardly derived by fixing and introducing the notation ; this results in
[TABLE]
Observe that this expression for the Kähler potential implies that the ’s are constrained by .
As the model is quadratic, the stabilization equations are easily solved and leads to
[TABLE]
With this solution to the stabilization equation, we can express the metrical factor, Eq. (3.14), in terms of the as
[TABLE]
where in that last step we used the fact that in this article we shall consider only purely magnetic solutions, so that . The fact that we choose to consider magnetic embeddings only, implies be means of Eq. (3.15) that we will be dealing with static solutions.
In order to finish the discussion of the model, we must discuss the possible gauge groups that can occur in the -models: as we saw at the beginning of this section, these models have a manifest symmetry, under which the ’s transform as a vector. Furthermore, as we are mostly interested in monopole-like solutions, we shall restrict our attention to compact simple groups, which, as implied by Eq. (A.51), must be subgroups of . In fact, Eq. (A.51) and Eq. (A.42) make the stronger statement that given a gauge algebra , the action of on the ’s must be such that only singlets and the adjoint representation appear. For the -models there is no problem whatsoever as we can choose to be large enough as to accomodate any Lie algebra. Indeed, as is well-known any compact simple Lie algebra is a subalgebra of and the branching of the latter’s vector representation is exactly the adjoint representation of .
The simplest possibility, namely the -gauged model on , will be used in the remainder of this section, and the -gauged model will be used in section (4.2). The - and the -gauged models will not be treated, but solutions to these models can be created with great ease using the information in this section and Appendix C.
As we are restricting ourselves to purely magnetic solutions, which are automatically static, the construction of explicit supergravity solutions goes through the explicit solutions to the Bogomol’nyi equation (3.24). Having applications to the attractor mechanism in mind, and being fully aware of the fact that this class consists of only the tip of the iceberg of solutions, we shall restrict ourselves to spherically symmetric solutions to the Bogomol’nyi equations.
Working in gauge theories opens up the possibility of compensating the spacetime rotations with gauge transformations, and in the case of an gauge group this means that the gauge connection and the Higgs field, , after a suitable gauge fixing, takes on the form (See e.g. [36])
[TABLE]
Substituting this Ansatz into the Bogomol’nyi equation we find that and must satisfy
[TABLE]
All the solutions to the above equations were found in Ref. [37] and all but one of them contain singularities. Furthermore, not all of them have the correct asymptotics to lead to asymptotic flat spaces and only part of the ones that do can be used to construct regular supergravity solutions [16, 17]. Here, by regular supergravity solutions we mean that the solutions is either free of singularities, which is what is meant by a globally regular solution, or has a singularity but, like the black hole solutions in the Abelian theories, has the interpretation of describing the physics outside the event horizon of a regular black hole. The criterion for this last to occur is that the geometry near the singularity is that of a Robinson-Bertotti/ spacetime, implying that the black hole has a non-vanishing horizon area, whence also entropy.
The suitable solutions, then, break up into 3 classes:
(I) ’t Hooft-Polyakov monopole
This is the most famous solution and reads
[TABLE]
where is a positive constant. The renowned regularity of the ’t Hooft-Polyakov monopole opens up the possibility of creating a globally regular solution to the supergravity equations which is in fact trivial to achieve: for the moment we have been ignoring , which, since it is uncharged under the gauge group, is just a real, spherically symmetric harmonic function we can parametrize as
[TABLE]
It is clear, however, that if we want to avoid singularities, we must take , so that the only free parameter is .
Let us then discuss the regularity conditions imposed by the metric: as was said before, the solutions are automatically static, so that if singularities in the metric are to appear, they arise from the metrical factor . Plugging the solution for the Higgs field into the expression (4.4), we find
[TABLE]
As one can infer from its definition in Eq. (4.8), the function is a monotonic, positive semi-definite function on and vanishes only at , where it behaves as ; its behaviour for large is given by , which means that we should choose large enough in order to ensure the positivity of the metrical factor. A convenient choice for is given by imposing that asymptotically we recover the standard Minkowskian metric in spherical coordinates: this condition gives from which we find the final metrical factor and can then also calculate the asymptotic mass, i.e.
[TABLE]
Written in this form, it is paramount that the metric is globally regular and interpolates between two Minkowksi spaces, one at and one at .
In order to show that the solution is a globally regular supergravity solution, we should show that the physical scalars are regular. In the -models the scalars are given by (introducing the outward-pointing unit vector )
[TABLE]
so that the regularity is obvious. The scalars also respect the bound as can be seen from the fact that the bound corresponds to the positivity of the metrical factor. This regularity of the scalars and that of the spacetime metric are related [31].
(II) Hairy black holes
A generic class of singular solutions is indexed by a free parameter , called the Protogenov hair, and can be seen as a deformation of the ’t Hooft-Polyakov monopole, i.e.
[TABLE]
The effect of introducing the parameter is to shift the singularity of the cotangent from to , i.e. outside the domain of , but leaving unchanged its asymptotic behaviour.101010 One can consider the limiting solution for , the result of which was called a black hedgehog in Ref. [16]. This solution has, apart from not containing hyperbolic functions, no special properties and will not be considered seperately. This not only means that the function vanishes at some , but also that it becomes singular at , so that in order to build a regular solution we must have . Using then the general Ansatz for , Eq. (4.9), in order to calculate the metrical factor, we find in stead of Eq. (4.10)
[TABLE]
As the asymptotic behaviour of is the same as the one for the ’t Hooft-Polyakov monopole, the condition imposed by asymptotic flatness still is . Given this normalization, the asymptotic mass is
[TABLE]
which should be positive for a physical solution. In this respect, we would like to point out that the product should be positive as otherwise the metrical factor would become negative or zero, should it coincide with the zero of , at a finite distance, ruining our interpretation of the metric as describing the outside of a regular black hole. This then implies that the mass is automatically positive. Finally, let us point out that neither the mass nor the modulus depend on the Protogenov hair parameter .
The metrical factor is clearly singular at , but given the interpretation of the metric this is not a problem as long as the geometry near , which corresponds to the near horizon geometry, is that of an space. This is the case if
[TABLE]
is positive and can thence be identified with the entropy of the black hole.
The scalars for this solution are given by
[TABLE]
whose asymptotic behaviour is the same as for the ’t Hooft-Polyakov monopole. Its behaviour near the horizon, i.e. near , is easily calculated to be
[TABLE]
and does not depend on the moduli nor on the Protogenov hair, but only on the asymptotic charges. Observe, however, that since at some finite , there is a 2-sphere outside the horizon at which the scalars vanish, which is not a singularity for the scalars of this model.
(III) Coloured black holes
There is another particular solution to the Bogomol’nyi equation that has all the necessary properties, and this solution is given by
[TABLE]
This solution has the same behaviour as the hairy solutions, but is such that in the asymptotic regime it has no Higgs v.e.v. nor colour charge. Given the foregoing discussion, it is clear that this solution can be used to build a regular black hole solution, and we can and will be brief.
The regularity of the metric goes once again through the judicious election of and : the normalization condition implies that which then also implies that the asymptotic mass of the solution is . It may seem strange that the YM-configuration does not contribute to the mass, but it does so, at least for a regular black hole solution, in an indirect fashion: the condition for a regular horizon is clearly given by Eq. (4.16), which implies that . With these choices then, the scalars are regular for and at the horizon they behave as in Eq. (4.18).
4.2 Non-maximal symmetry breaking in gauged
In Ref. [35], E. Weinberg presented an explicit solution for a spherically symmetric monopole solution that breaks the parent gauge group down to ; in this section we will discuss the embedding of this solution into supergravity and also generalize it to a family of hairy black holes by introducing Protogenov hair111111 In Ref. [38] the general equations for a spherically symmetric solution to the Bogomol’nyi equations were derived. This opens up the possibility of analysing the system along the lines of Ref. [37], but for the moment this has not lead to anything new..
The starting point of the derivation of Weinberg’s monopole is the explicit embedding of an ’t Hooft-Polyakov monopole into an subalgebra of . In order to make this embedding paramount we take the generators of to be , () and (). These generators satisfy the following commutation relations
[TABLE]
where we have introduced the ’t Hooft symbols and . The (resp. ) are self-dual (resp. anti-selfdual) 2-forms on and satisfy the following relations
[TABLE]
We would like to stress that is not the complex nor the Hermitean conjugate of .
Following Weinberg we make the following Ansatz for the -valued connection and Higgs field, taking () to be the generators of ,
[TABLE]
where , , and are functions of only. and are determined by the criterion that we have an ’t Hooft-Polyakov monopole in some -subalgebra, which we take to be generated by the . One way of satisfying this criterion is by choosing
[TABLE]
which implies that the Bogomol’nyi equation in the sector reduce to Eqs. (4.6) and (4.7).
The analysis of the Bogomol’nyi equations in the remaining sectors impose the constraint that and the differential equations121212 In order to go from Weinberg’s notation [35] to ours one needs to change , , , , and also .
[TABLE]
The final ingredient, needed for the calculation of the metrical factor, consists of finding an expression for the -invariant quantity : this is
[TABLE]
In conclusion, given a solution to Eqs. (4.6,4.7,4.25) and (4.26) we can discuss their embedding into the -gauged -model by means of Eq. (4.27).
Weinberg’s monopole in supergravity
The explicit form of Weinberg’s monopole is given by the solution in Eq. (4.8) and
[TABLE]
where the profile function , given by
[TABLE]
depends on a positive parameter called the cloud parameter. The cloud parameter is a measure for the extention of the region in which the Higgs field in the - and the -directions are active: in fact when the profile functions vanishes identically and we are dealing with an embedding of the ’t Hooft-Polyakov monopole. The maximal extention is for which then means that .
As one can see from the definitions, and are positive semi-definite functions that asymptote exponentially to zero. This not only means that the gauge symmetry is asymptotically broken to , but also that and will not contribute to the asymptotic mass, nor to the normalization condition. Unlike the ’t Hooft-Polyakov monopole or the degenerate Wilkinson-Bais -monopole (C.11), however, the regularity of the solution does not imply that the Higgs field vanishes at ! In fact, near one finds that
[TABLE]
It is this behaviour that may pose a problem for creating a globally regular solution and is the reason for including it in this article.
Using Eqs. (4.4) and (4.27) and choosing as in Sec. (4.1) , we can write the metrical factor as
[TABLE]
where we already used the normalization condition . As mentioned above, and asymptote exponentially to zero and cannot contribute to the mass, which is the one for the ’t Hooft-Polyakov monopole, i.e. .
Let us then investigate the behaviour of (4.32) at : a simple substitution shows that
[TABLE]
which is always positive so that the non-zero value of the Higgs field at the origin is no obstruction to the construction of a globally regular supergravity solution. The remaining question as far as the global regularity of the solution is concerned, is whether there are values of for which the metrical factor (4.32) becomes negative. This however never happens as one can see from Fig. (1) which shows a plot of for the values of and .
Another hairy black hole
The introduction of Protogenov hair, i.e. a real and positive parameter , in Weinberg’s monopole solution is trivial and leads to the following solution
[TABLE]
supplemented by the expression for and given in Eq. (4.13). As far as the limiting cases of this family is concerned, it is clear that Weinberg’s monopole is obtained in the limit ; in the limit we find that and the solution splits up into the direct sum of an black hedgehog, i.e. an limit of (4.13), and an coloured black hole, Eq. (4.19).
As in the case of the hairy black holes, the introduction of the hair parameter preserves the asymptotic behaviour of Weinberg’s monopole and the solution is regular for . This immediately implies that the normalization condition for once again reads and that the asymptotic mass of this solution is given by Eq. (4.15), which is positive with the usual proviso that .
As in the case of the hairy black holes in the -gauged -models, the regularity of the metric imposes the constraint that the entropy
[TABLE]
be positive. This positivity of the entropy also ensures that the physical scalars stay in their domain of definition at . Indeed, the physical scalars can be compactly written as
[TABLE]
which are therefore regular for . Their value at is
[TABLE]
which, as in the case of the solution, depend only on the asymptotic charges.
4.3 Non-Abelian solutions in Magic models
In this section we would like to discuss the embeddings of monopole solutions into the gauged Magic supergravity theories. We want to show that it is not always possible to construct, given a prepotential for a theory, a globally regular solution based on a given monopole solution. We would like to stress that this holds for a given prepotential, as the choice of symplectic section for a given gauged model is physical due to the breakdown of symplectic invariance.
To start looking for ways to embed monopoles into gauged magic supergravities, we must discuss first the possible gaugings of the magic models, which boils down to a group theory problem whose outcome is given in Table 1, which we are going to explain now.
The scalar manifolds of the magic models are based on symmetric coset spaces , which are given in the second and the third column in the table. As the isometry-group of the scalar manifold, which for the magic models is isomorphic to , acts on the symplectic section defining the model (see Appendix A), we should specify under what representation of it transforms; this representation is given in the column denoted as . The following 2 columns determine how the isotropy subgroup acts on the complex scalars ; the reason why this is important will be discussed presently.
As we are interested in monopoles, we shall restrict ourselves to compact gauge groups , which implies that . Moreover, as we restricted ourselves to a specific class of gaugings, i.e. gaugings that satisfy Eq. (A.42), we should use a prepotential that is -invariant. Manifestly -invariant prepotentials for the magic models were given in Ref. [40]. These prepotentials are of the -type and have the form
[TABLE]
where is a cubic -invariant131313 By we mean minus the -factors. , whose value for the specific magic model can be found in the seventh column of Table 1.
Another implication of our choice of possible gauge groups is that we can only consider for which the branching of the -representation of the to -representations contains only the adjoint representation and singlets. This is a very restrictive property and the maximal possibilities we found are listed in the last column of Table 1.
Having discussed the possible models, we must then start discussing the actual embedding of the magnetic monopoles. The first thing is to solve the stabilization equation to find in terms of . This is a complicated question but luckily a general solution exists and was found by Bates and Denef [33]; this solution uses the fact that the generic entropy functions for these models are known. For our purposes, however, the full machinery is not needed. Instead, we shall consider the simpler setting of embedding a purely magnetic monopole in the matter sector and only turn on an electric component for the graviphoton. This means that we should solve the stabilization equations,
[TABLE]
where we absorbed the function into the ’s. This system admits a solution
[TABLE]
With this solution to the stabilization equation, it is then straightforward to use Eq. (3.14) to determine
[TABLE]
4.3.1 The -magic model
Let us then consider the -magic model, which allows an gauging. The reason why this is the case is easy to understand: as one can see from Table 1 the ’s transform under as a representation. Choosing to gauge the diagonal means identifying the left and the right actions so that w.r.t. the diagonal action the ’s transform as , which is just what we wanted.
The spherically symmetric monopole solution to the Bogomol’nyi equations were found by Wilkinson and Bais in Ref. [41], and a discussion of these solutions is given in Appendix C. In order to discuss the embedding of the WB-monopole, we gather the components of the symplectic vector into a matrix, , and as this matrix behaves as the sum of a singlet and the adjoint under the diagonal , we must take it to be
[TABLE]
where is defined in Eq. (C.2) and
[TABLE]
is a real and spherically symmetric harmonic function. If we then also conveniently redefine , where
[TABLE]
is another real harmonic function, we can express Eq. (4.43) as
[TABLE]
Given the asymptotic behaviour of the WB solution, let us for clarity discuss the non-degenerate solution whose asymptotic behaviour is given in Eq. (C.10), we can normalize the solution to be asymptotically Minkowski by demanding that
[TABLE]
Using this normalization, we can then extract the asymptotic mass which turns out to be
[TABLE]
and must be ensured to be positive.
Let us then look for a globally regular embedding of the WB-monopole by tuning the free parameters: as before, we shall take in order to avoid the Coulomb singularities in the Abelian field strengths. The first obvious remark is that is already fixed in terms of and the due to Eq. (4.48), so that we need to discuss the possible values for : a first constraint for comes from the positivity of the mass. Using the facts that and , which follow from the constraint and the chosen ordering, in the mass formula (4.49) we see that this implies
[TABLE]
As we are interested in finding globally regular embeddings, we should discuss the regularity of the metric at : as the ’s vanish at the origin we see that regularity implies that
[TABLE]
It is not hard to see that the above holds for the 2 bounds on derived in Eq. (4.50). At this point then, the real question is whether, given the constraints on and derived above, there are values for other than or for which the metrical factor in Eq. (4.47) vanishes; from the monotonicity of and it is clear that if this is to happen, then this is because the factor vanishes. Seeing, then, that the combination takes values between and , we see that Eq. (4.47) never vanishes if
[TABLE]
In order to finish the discussion of the regularity, we must have a look at the physical scalars: for the above embedding they are schematically given by , where is given in Eq. (4.42). The regularity then follows straightforwardly from the regularity of monopole solution and the metric.
4.3.2 The -magic model
All the embeddings of YM monopoles discussed till now, share a common ingredient, namely the occurrence of additional Abelian fields, whose associated harmonic functions can be used to compensate for the vanishing of the Higgs field at . In the above example, this rôle is played by and and in the and -models by the graviphoton. In fact, a model in which no such a compensator exists is the -magic model.
As displayed in Table 1, the in the matter sector lie in the of , which corresponds to holomorphic 2-forms. As admits an as a singular subgroup for which the relevant branching is , we can try to embed an WB monopole [41]. This monopole is given, as in the case, by 3 functions () and their embedding into the -model has . The asymptotic behaviour can of course be compensated for by choosing judiciously, but the real problem lies at . At the origin the vanish as , and