ECM-UB-97/32 UG-9/97 hep-th/9711043 M-brane intersections from worldvolume superalgebras 8 9 9 1 Eric Bergshoeff1, Joaquim Gomis 2 and Paul K. Townsend 2 ∗ n a J 1Institute for Theoretical Physics, Nijenborgh 4, 9 9747 AG Groningen, The Netherlands 3 v 2 Departament ECM, Facultat de F´ısica, Universitat de Barcelona and Institut de Fisica d’Altes 3 4 Energies, 0 Diagonal 647, E-08028 Barcelona, Spain 1 1 7 9 Abstract / h t - It is known that the M-branes of M-theory correspond to p-form charges in the D=11 spacetime su- p e persymmetry algebra. Here we show that their intersections are encoded in the p-form charges of their h worldvolume supersymmetry algebras. Triple intersections are encoded in double intersection worldvol- : v ume algebras with eight supercharges. i X r a ∗OnleavefromDAMTP,UniversityofCambridge,UnitedKingdom. 1 1 Introduction The 11-dimensional (D=11) M-theory is currently the leading candidate for a unified theory of particle interactions, despite the fact that a complete understanding of its microscopic degrees of freedom (es- pecially those relevant to compactifications to four spacetime dimensions) remains elusive. Fortunately, many of the implications of M-theory can be deduced from semi-classical considerations. At this level, M-theory appears to be a theory of one ‘basic’, or 1/2-supersymmetric, ‘object’ which can appear in various dual forms. These can be found as solutions of the D=11 supergravity field equations. Alterna- tively, the possibilities may be investigated more abstractly via the ‘M-theory superalgebra’ [1]; this is a modification of the standard D=11 supersymmetry algebra to incorporate 2-form and 5-form central charges [2]. The space components of these charges are associated [3, 4] with the M-2-brane and M-5- brane. The time componentsof the 2-formchargemaybe associated[4,5] with aboundary ofthe D=11 spacetime, as in the Hoˇrava-Witten (HW) construction of the heterotic string [6]. For terminological convenience we refer to this as the M-9-brane; it can in fact be viewed as a decompactification limit of eight D-8-branes coincident with an orientifold 8-plane in Type IA superstring theory [7]. The time component of the 5-form charge is associated with the Kaluza-Klein (KK) monopole solution of D=11 supergravity, alias the D-6-brane of IIA supergravity [8]. We shall refer to this as the M-KK-monopole. Finally, the space components of the 11-momentum are associated with the massless quanta of D=11 supergravity or, at the classical level, with the ‘M-wave’ solution of D=11 supergravity [9]. This is the only 1/2 supersymmetric configuration permitted by the standard D=11 supersymmetry algebra, hence the necessity of considering the full M-theory superalgebra. To summarize, there are five ‘basic’, 1/2-supersymmetric, constituents of M-theory. These are M−Wave, M−2−brane, M−5−brane, M−KK−monopole, M−9−brane. (1) The M-2-brane and M-5-brane are often referred to collectively as ‘M-branes’; they are in fact the only bone fide branes. All the same, we shall find it convenient to refer collectively to all five ‘basic’ con- stituents as ‘M-branes’. M-theory configurations with less than 1/2 supersymmetry may now be found as ‘intersecting’ M-branes [10]. The simplest cases are those in which two M-branes intersect to yield a configuration with 1/4 supersymmetry (of the M-theory vacuum). From the point of view of each of the two constituent branes, the intersection appears as a 1/2 supersymmetric object in its worldvolume. Thus, one way to study 1/4-supersymmetricintersections is by lookingfor 1/2 supersymmetric solutions oftheworldvolumefieldequationsofthevariousM-branes. Forexample,itisknownthattwoM-5-branes may intersect on a 3-brane, preserving 1/4 supersymmetry [10, 11], and that an M-2-brane may end on anM-5-brane[12,13], againpreserving1/4supersymmetry (the boundarymaybe consideredasastring intersection for present purposes). From the perspective of the M-5-brane (or one of them if there is more than one) these intersections should appear as 1/2 supersymmetric 3-brane and string solutions of its effective worldvolume field equations. Both solutions have recently been found [14, 15], following the discovery of analogous solutions of the worldvolume field equations of D-branes [16, 17]. It is remark- able that these solutions, while ostensibly just configurations of a particular worldvolume field theory, actually provide their own spacetime interpretation as intersecting branes; this comes about because the worldvolume scalars determine the spacetime embedding. Here we take this a step further by showing that the spacetime interpretation of a worldvolume p- brane preserving 1/2 supersymmetry is encoded in the worldvolume supersymmetry algebra. Indeed, all 1/4supersymmetricM-braneintersectionsmaybededucedthisway. Thefirstpointtoappreciateisthat worldvolume p-branes are associatedwith p-forms in the worldvolume supersymmetry algebra. In itself, this is not surprising given our experience with the spacetime supersymmetry algebra. For example, it waspointedoutin[15]thatthemaximalcentralextensionoftheD=6(2,0)worldvolumesupertranslation 2 algebra of the M-5-brane has the anticommutator {QI,QJ}=ΩIJP +Y[IJ]+Z(IJ), (Ω Y[IJ] =0), (2) α β [αβ] [αβ] (αβ) IJ where α,β = 1,...,4 is an index of SU∗(4) ∼= Spin(5,1) and I = 1,...,4 is an index of Sp(2), with ΩIJ being its invariantantisymmetric tensor. TheY-chargeis a worldvolume1-formandthe Z-chargea worldvolume self-dual 3-form. The presence of such charges in the worldvolume supersymmetry algebra is implied by the existence of the 1/2 supersymmetric string and 3-brane solutions of the M-5-brane worldvolume field equations. There is still a puzzle to be resolved, however. There are a total of five 1-form charges and ten 3-form charges (in the irreducible 5 and 10 representations of Sp(2)), which exceeds the number of distinct worldvolume strings and 3-branes. In the spacetime supersymmetry algebra, every p-form charge corresponds to a distinct p-brane. This is clearly not so for worldvolume algebras. The resolution of this puzzle is simply that the Sp(2) representations provide the information needed to reconstruct the spacetime interpretation of the branes within the M-5-brane. For example, identifying sp(2)∼=Spin(5)as the double coverof the SO(5) rotationgrouptransverseto the M-5-brane worldvolumeintheD=11spacetime,weseethatthe1-formsY defineavectorinthis transverse5-space. This information serves to identify the 1-brane in the worldvolume as the boundary of an M-2-brane in theM-5-brane. Note,however,thatthe5representationofSp(2)couldequallywellhavebeeninterpreted asdefining a4-forminthe transverse5-space,sothe 1-branehasanotherspacetime interpretationasthe intersection of the M-5-brane with another M-5-brane,which is again an M-brane intersection known to preserve 1/4 supersymmetry [18]. So far we have implicitly considered only the space components of Y. As pointed out in [15], the time componentscanbe viewedasthe spacecomponentsofaworldvolumedual5-formandaretherefore chargesfor a space-filling 5-branein the M-5-brane. We interpret this to mean that the givenM-5-brane is actually inside another brane, so that its ‘intersection’ with it coincides with the M-5-brane. If the 5 representation is taken to define a 1-form in transverse 5-space we might then conclude that the M- 5-brane is inside a 6-brane. There is no M-theory 6-brane as such but there is the M-KK-monopole. Adopting this interpretation we conclude, correctly [19], that an M-5-brane ‘inside’ an M-KK-monopole preserves 1/4 supersymmetry. If instead we take the 5 representation to define a transverse 4-form then we conclude, again correctly [20], that an M-5-brane inside an M-9-brane preserves 1/4 supersymmetry; in fact, this is just the heterotic 5-brane in the HW-formulation of the heterotic string theory. There is still much to learn from the algebra (2). For example, we have not yet considered the Z chargesinthe10representationofSp(2). Sufficeittosaythatthecompleteanalysispreciselyduplicates the known classification of 1/4 supersymmetric intersections involving an M-5-brane. There remains a question of consistency. To take an example, we have just deduced from the M-5-brane worldvolume superalgebra that it can have a 1/4-supersymmetric intersection with an M-2-brane. We should be able toseethisfromtheM-2-brane’sworldvolumealgebra. Thisis(aftergaugefixing)aD=3fieldtheorywith N=8 supersymmetry. Allowing for all possible p-form central charges in the N=8 D=3 supertranslation algebra we arrive at the anticommutator {Qi,Qj}=δijP +Z(ij) +ε Z[ij], (δ Z(ij) =0), (3) α β (αβ) (αβ) αβ ij where Qi are the eight D=3 Majorana spinor supercharges (i = 1,...,8), and P is the 3-momentum1. ThissupersymmetryalgebrahasanSO(8)automorphismgroup,whichweinterpretastherotationgroup in the transverse 8-space. The supercharges do not transform as an 8-vector, however, but as a chiral SO(8) spinor. The 0-form and 1-form central charges are therefore Z(0−form):28, Z(1−form):35+, (4) 1This,atleast,istheobvious interpretationofthis‘charge’. Weshallseelaterthatithasanother interpretation. 3 where the 28 is either a 2-form or 6-form of the transverse 8-space and the 35+ is a self-dual transverse 4-form. Consider first the worldvolume 0-forms, corresponding to a 0-brane on the D=3 worldvolume. Choosingthe2-forminterpretationofthe28representationweseethatthe 0-braneistheresultofanin- tersectionwithanotherM-2-brane,aknown1/4supersymmetricM-braneintersection[10]. Ifweinstead choose the 6-form interpretation of the 28 then the worldvolume 0-brane acquires the interpretation as the intersection of the M-2-brane with a KK-monopole [21]. Consider now the worldvolume 1-forms, corresponding to a 1-brane on the D=3 worldvolume. This 1-brane is determined by a 4-plane in the transverse space and is therefore the intersection of an M-5-brane with the M-2-brane. We therefore recoverfromtheM-2-braneworldvolumesuperalgebratheintersectionwiththeM-5-branethatweprevi- ously deduced from the latter’s worldvolume superalgebra. Again, there is much more to be learnt from the algebra (3), but we postpone the complete analysis. TheremainingM-branesuperalgebrascanbedealtwithinthesameway. Thefinalresultisthateach 1/4-supersymmetricintersectionoftwoM-branesoccurspreciselyonceineachofthelistsofintersections derivedfromthe worldvolumesupersymmetryalgebrasofits constituents. Ifthesameanalysisisapplied to D-branes the results are not quite so nice. It remains true that every intersecting brane configuration preserving 1/4 supersymmetry corresponds to some spacetime interpretation of a p-form charge in the worldvolume superalgebra of either participating brane. Conversely, every p-form in a worldvolume superalgebrahasaspacetimeinterpretationasa1/4-supersymmetricconfigurationofintersectingbranes, but it is no longer always the case that every such interpretation corresponds to a 1/4-supersymmetric configuration; one has to impose additional ad hoc interpretational rules. The reason for this can be traced to some peculiarities of supersymmetry algebras with 16 supercharges. The complete list of such algebras is as follows D =1 N =16 so(9) 16 D−0 ∗ D =2 (16,0) so(9) 16 M −W D =2 (16,0) so(8) 8 ⊕8 IIB−W s s D =2 (8,8) so(8) 8 ⊕8 IIB−1 s c D =2 (8,8) so(8) 8 ⊕8 IIA−1 s s ∗ D =3 N =8 so(8) 8 M −2 s D =4 N =4 so(6) 4⊕4¯ D−3 D =5 N =2 so(5) 4 D−4 ∗ D =6 (2,0) so(5) 4 M −5 D =6 (1,1) so(4) 4 IIB−5 ∗ D =7 N =1 so(3) 2 D−6, M −KK D =8 N =1 so(2) 2 D−7 D =9 N =1 Z − D−8 2 ∗ D =10 (1,0) − − D−9, M −9 The value of N is given in the third column, with the convention that minimal supersymmetry is always N=1; for D=2,6,10, the number of chiral and antichiral spinors must be specified separately. The auto- morphism, or ‘R-symmetry’, algebra is given in the fourth column; the group is generally the spinorial coveringgroupand the representationto which the superchargesbelong is givenin the fifth column. For D=1,2 we have not given the largest possible automorphism algebra. For example, the (8,8) algebras actually have a Spin(8) groupacting independently on spinors of each chirality, but in the string-theory realization only the diagonal Spin(8) survives. Similarly, in D=1 the maximal automorphism group is SO(16) but only SO(9) is realized by 0-branes. This is because the supercharges must be spinors of the R-symmetry group as well as of the worldvolume Lorentz group, and SO(9) is the largest subgroup of SO(16)witha16-dimensionalspinorrepresentation. Inthelastcolumnwegiveabranefieldtheorythat realizes the supersymmetry algebra. Some branes do not appear in the list because they realize only a 4 proper subgroup of the automorphism algebra as a consequence of being the dimensional reduction of another brane with the same algebra. For example, the IIA-wave is essentially the same as the M-wave but with the SO(9) symmetry reduced to SO(8). If one associates the so(n) R-symmetry algebra of each of the above supersymmetry algebras to an SO(n) rotation group in a space transverse to a (p+1)-dimensional worldvolume then one arrives at an interpretationinwhichthep-braneisembeddedinalargerspacetimeoftotaldimensionD =(n+p+1). ThistotaldimensionisineachcaseeitherD=10orD=11. Thisfactmightwellbeconsideredremarkable to someone studying these algebras as a mathematical excercise, but it is of course expected from their interpretation as worldvolume superalgebras of D=10 or D=11 branes. Note that a single transverse dimension would be associated with a trivial SO(1) rotation group so that in those cases for which the R-symmetry algebrais trivialthere is anambiguity in the dimension of the transversespace: it couldbe zero or one. For example, if we suppose that the transverse space associated to the D=10 (1,0) algebra has dimension one we may interpret the algebra as that of the M-9-brane. If instead we choose the transverse space to have dimension zero we may interpret the algebra as that of the D-9-brane, which is effectively the D=10 spacetime of the Type I superstring. The D=9 N=1 supersymmetry algebra is another rather special case. Its Z automorphism group is the vestige of a rotation group in a 1- 2 dimensional transverse space, so the D-8-brane acquires a natural D=10 interpretation but, as usual, its 11-dimensional interpretation is obscure. An inspection of the above supersymmetry algebras reveals that most are dimensional reductions of some other one. The only exceptions are the five supersymmetry algebras corresponding to the world- volume field theories of the five M-branes. These have been marked with an asterisk in the first column. TheM-wave,M-5-braneandM-9-branealgebrasaremanifestlynotreductionsofanyotheralgebrainthe list. Dimensional reduction of the D-3-brane algebra yields an N=8 D=3 algebra but its automorphism algebra is so(7) rather than so(8). This is of course the algebra of the D-2-brane, which we did not include because it is just the reduction of the M-2-brane. Thus, the M-2-brane algebra is also not the reductionofanyotherone. ThisleavesthealgebraassociatedtotheM-KK-monopole. Strictlyspeaking, this is an exception to the rule because it is also the D-6-brane algebra, which is the reduction of the D-7-brane algebra. However, this algebra shares the nice features of the other M-brane algebras, and whenonetakesintoaccountthehiddendimensionimplicitintheKK-monopolemetric,thereisanatural D=11interpretation. Inanycase,itsinclusionamongthe‘exceptional’casesisforcedbythe intersecting brane interpretation of the p-form charges in the other M-brane algebras. We learn from all this that when determining the possible 1/4-supersymmetricbrane intersections from the worldvolume supersym- metry algebras of participating branes we need only consider the five M-branes. Thus, the worldvolume superalgebras of D=10 and D=11 branes proclaim the ‘supremacy’ of M-theory. Having determined all 1/4-supersymmetric intersections of M-branes from the supersymmetry alge- bras with 16 supercharges one is led to wonder whether intersections preserving less supersymmetry are encoded in the supersymetry algebras with fewer supersymmetry charges. The answer appears to be a qualified ‘yes’, but we shall not attempt a systematic analysis here. Rather, we shall present a few illustrative cases in which triple intersections with 1/8 supersymmetry can be deduced from the p-form charges in the superalgebra with 8 supercharges corresponding to 1/4-supersymmetric pairwise brane intersections. One case of interest is Type I brane intersections viewed as triple intersections in which the ‘third brane’is a D-9-brane. Another class of cases can be deduced from the worldvolumealgebra of the 3-brane in the M-theory 5-brane. Analysis of the p-forms in this algebra shows, in particular, that the 3-brane in the M-5-brane is a D-brane for the self-dual string. We shall provide both D=10 IIB and D=11 M-theory explanations of this fact. 5 2 M-brane intersections 2.1 M-wave The M-wave algebra is {Qi ,Qj }=δiP +Z(ij), (δ Z(ij) =0), (5) − − j = = ij = where the supercharges Qi (i = 1,···,16) are in the 16 spinor representation of SO(9). This has the − obvious reduction to the D=10 D-0-brane algebra {Qi,Qj}=δiH +Z(ij)(126⊕9). (6) j As indicated, the 0-formcentralchargesin this algebrafall into the 126 and 9 representationsofSO(9), as do the D=2 self-dual vector Z-charges in the M-wave algebra. Intersections of M-waves preserving 1/4 supersymmetry have been classified in [21], but since these are less familiar than those involving D-0-branes it is helpful to consider the latter first. The results for the M-wave are then obtained by re-interpreting the D-0-brane results in D=11. The 126 of SO(9) is either a 4-form or a 5-form in the transverse 9-space. This corresponds to a D-0-brane in a D-4-brane or a KK-monopole, respectively. The 9 of SO(9) is either a 1-form or an 8- form in the transverse9-space. This correspondsto a D-0-brane in a fundamental string or a D-8-brane, respectively. The D=11 interpretation of the same configurations is, in the notation of [21], (1|W,M5), (1|W,KK), (1|W,M2), (1|W,M9). (7) This is the complete list of 1/4-supersymmetric intersections of M-waves. 2.2 M-2-brane The M-2-brane algebra is given in (3). We considered earlier the spacetime interpretation of the space components of the 1-form and 3-form charges. This led to the following M2-brane intersections (0|M2,M2), (0|M2,KK), (1|M2,M5). (8) The time components of the 1-form charges are equivalent to space components of the dual 2-forms and hence correspond to a ‘worldvolume 2-brane’, i.e. worldvolume filling 1/2-supersymmetric 2-branes. These cases are naturally interpreted as ‘intersections’ in which the original M-2-brane is inside some larger brane. The latter is nominally a 6-brane since it is determined by an additional 4-plane in the transverse space but, as earlier, this is to be interpreted as a KK-monopole. We thus conclude that an M-2-brane inside a KK-monopole preserves 1/4 supersymmetry, as indeed it does. This yields the intersection (2|M2,KK). (9) To find the remaining intersections of M-2-braneswe must take into accountthat the algebra (3) admits 1/2supersymmetric chargeconfigurationsin whichonly the 3-vectorP is non-zero;itmust then be null. Since P is also a singlet of SO(8) we may consider it to be either a 0-form in the transverse 8-space or an 8-form. In the former case P has the obvious interpretation as the null 3-momentum of an M-wave ‘inside’ an M-2-brane, a known 1/4-supersymmetric ‘intersection’. In the latter case P represents the intersection of a 9-dimensional object with the M-2-brane; it must be interpreted as a charge associated totheboundaryofanM-2-braneonanM-9-brane. ThefactthatP isnullisanindicationthatthestring boundaryis chiral,consistentwithits HW-interpretationasthe heteroticstring. We havenowfoundthe additional M2-brane intersections (1|M2,W), (1|M2,M9), (10) and these complete the list of possible 1/4-supersymmetric intersections involving M-2-branes. 6 2.3 M-5-brane The M-5-brane worldvolume algebra is given in (2). We have already explained how the Y-charge leads to the intersections (1|M5,M2), (1|M5,M5), (5|M5,KK), (5|M5,M9). (11) Consider now the self-dual 3-form chargeZ carried by a 3-brane in the M-5-brane. This charge is in the 10 representation of Sp(2), which may be interpreted as either a 2-form or a 3-form in the transverse 5-space. In the 2-form interpretation, the 3-brane is the intersection with a second M-5-brane. In the 3-form interpretation, it is the intersection with an M-KK-monopole. We thus deduce the intersections (3|M5,M5), (3|M5,KK). (12) Finally, we must consider the SU∗(4) ∼= Spin(5) 5-vector P. If P is assumed to be a 0-form in the transverse space then P has the obvious interpretation as the null 5-momentum of an M-wave in the M-5-brane. However, we could also suppose P to be a transverse 5-form, which would indicate a 6- dimensional object. This must be interpreted as an M-KK-monopole intersecting the M-5-brane on a chiral string. We thus deduce the intersections (1|M5,W), (1|M5,KK). (13) Note that if the non-compact hyper-Kahler 4-metric of the KK-monopole were replaced by a compact hyper-Kahler metric on K then the ‘intersection (1|M5,KK) would become an M-5-brane wrapped 3 on K , alias the heterotic string [23]. This goes some way to explaining why the string intersection in 3 (1|M5,KK)ischiral. Inanycase,wehavenowfoundthecompletesetofpossibleM-5-braneintersections preserving 1/4 supersymmetry. 2.4 M-KK-monopole The M-KK-monopole algebra is {Qi,Qj} = εij(CΓ ) Pµ+C Z(ij) (14) α β µ αβ αβ +εij(CΓµν) Z +(CΓµνρ) Z(ij), αβ µν αβ µνρ where i = 1,2 is an Sp(1) ∼= SU(2) index. We therefore have the following p-brane charges with their SU(2) representions Z(p=0):3 Z(p=2):1 Z(p=3):3 Z(p=4):3 Z(p=5):1 (15) The p=4,5 cases come from the time components of the 3-form and 2-form charges. The algebra (14) is alsothealgebraoftheD-6-braneanditwillbeconvenienttobeginwithananalysisofitsimplicationsfor D-6-brane intersections. In this case the transverse space is three-dimensional, so an SU(2) singlet can be interpretedaseither a 0-formoras a3-forminthe transverse3-space. Similarly,anSU(2)tripletcan be interpretedas either a 1-formor as a 2-formin the transversespace. Eachpossibility yields a distinct configuration of intersecting branes. The details are much as before and the result is that one finds all 1/4-supersymmetric intersections of D-6-branes for which the intersection is a p-brane with p=0,2,3,4,5. There remain two p=1 intersections not yet accounted for. One is a IIA-wave in a D-6-brane which is obviously associated with P, interpreted as the null 7-momentum of the wave. The other is found by taking the space components of the null 7-vector P to be a 3-form in transverse 3-space. In this case 7 P can be interpreted as the charge of a chiral string formed by the intersection of a D-4-brane with the D-6-brane. This can be represented by the array D6: 1 2 3 4 5 6 − − − D4: 1 − − − − − 7 8 9 Note that there are no dimensions in which the D-4-brane could be separated from the D-6-brane, so that a chiral theory on the intersection is indeed possible (the 11th dimension does not invalidate this argumentbecausetheM-KK-monopoleisnotstrictlyabranewithalocalizedworldvolume). Infact,this configurationisdual(atleastformally)toaD-stringinaD-9-brane,aliasthe (chiral)typeI superstring. Let us now return to the M-KK-monopole. The transverse space is now 4-dimensional, despite the Spin(3) R-symmetry. This means that the transverse dual to, say, a transverse 1-form is now a 3-form, whereas it was previously a 2-form. However, what was previously a 1-form might now really be a two- form,inwhichcasethetransversedualwouldstillbea2-form. Inotherwords,inassociatingformsoftwo possible ranksto eachworldvolumep-form in the algebra,one of the two rankswill now increase by one, the other one being unaffected. There is therefore an additional choice that must now be made, which amounts to a choice of where to put the extra dimension when re-interpreting the previous D-6-brane resultsinD=11. Thechoiceisnotarbitrary,however,becauseitisdeterminedbythepossibleM-branes. Only one choice has a D=11 interpretation because there is only one way to interpret each D-6-brane intersectionasaKK-monopoleintersection. Inthe caseof0-braneintersections,the twoD-6-branecases lift to the intersections (0|M2,KK), (0|M2,KK), (16) where the underlining in the second case indicates that the compact isometry direction of the KK- monopole is also an isometry of the M2-brane; in other words, the M-2-brane is wrapped on the KK circle. The two D-6-brane cases with 2-brane intersections lift to (2|M2,KK), (0|KK,KK). (17) The two 3-brane intersections are (3|M5,KK), (3|M5,KK), (18) with the same meaning to underlining as before. Similarly, the 4-brane intersections are (4|KK,KK)a, (4|KK,KK)b, (19) where the notation is that of [21]: the ‘a’ case is the one in which the KK circle is common to both KK ‘monopoles’ (the arrays and supergravity solutions may be found in [21]). The 5-brane intersections are (5|M5,KK), (5|M9,KK). (20) We thus recover from the M-KK algebra the M-5-brane intersection with an M-KK-monopole deduced previously from the M-5 algebra. In addition we find the KK intersection with an M-9-brane [20]; this can be interpreted as a KK-monopole having an M-5-brane boundary on an M-9-brane. Finally, the wave/chiralstring intersections in D=11 are (1|W,KK), (1|M5,KK), (21) which we deduced earlier from the M-W and M-5 algebras. 8 2.5 M-9-brane The M-9-brane worldvolume algebra is {Q ,Q } = (CP+Γ ) Pµ+(CΓµνρσλ) Z+ , (22) α β µ αβ αβ µνρσλ where P+ is the chiral projection operator on spinors and Z+ is a self-dual 5-form. When this is interpreted as the D-9-brane algebra the 5-form charge is associated with a 5-brane in the D-9-brane, alias the Type I D-5-brane. We shall consider this interpretation in more detail in the following section. For its M-9-brane interpretation we must suppose that there is a one-dimensional transverse space. The 5-formZ+ could representeither a 0-formor a 1-formin this transverse‘1-space’. In the 0-formcase we have an M-5-brane inside an M-9-brane, alias the heterotic 5-brane. In the 1-form case we have what might appear to be a 5-brane intersection of the M-9-brane by a 6-brane, but this should be interpreted as an M-KK-monopole with a boundary on the M-9-brane. Thus, we confirm the cases (5|M9,M5), (5|M9,KK), (23) found earlier from the M-5 and M-KK algebras. To find the remaining M-9-brane intersections we note that the space components of the 10-momentum P could be either 0-forms or 1-forms in the transverse space. IntheformercasewehavetheobviousinterpretationasanM-waveinanM-9-brane. Inthelatter case the object intersecting the M-9-brane is two-dimensional; it must be an M-2-brane with boundary on the M-9-brane. The fact that P is null is again an indication of the chirality of the string boundary, which is in fact the heterotic string. Thus we deduce the intersections (1|M9,W), (1|M9,M2), (24) which we also found previously from the M-W and M-2 algebras. We have now found the complete set of M-9-brane intersections preserving 1/4-supersymmetry. 3 Triple intersections We now turn to triple orthogonal intersections of M-branes or Type II branes preserving 1/8 super- symmetry. The common-intersection worldvolume field theory realizes a supersymmetry algebra with 8 supercharges. The list of such algebras is as follows D =1 N =8 so(8) 8 s ∗ D =2 (8,0) so(8) 8 IB−1 s D =2 (4,4) so(4) 4⊕4 D =3 N =4 so(4) 4 m−2 D =4 N =2 u(2) 2⊕2¯ D =5 N =1 su(2) 2 ∗ D =6 (1,0) su(2) 2 1B−5 Thealgebrasmarkedwithasterisksarethosewhicharenotthedimensionalreductionofanyother. These aretheworldvolumealgebrasoftheType1stringand5-brane,whichwecalltheIB-stringandIB-5-brane (because of the IIB originof the Type 1 string). Double intersections of these 1B-branes,preserving 1/4 of the supersymmetry of the Type 1 vacuum, can be considered as special cases of 1/8-supersymmetric triple intersections of IIB branes in which the ‘third brane’ is the D-9-brane. We shall consider these cases first. The Type 1 spacetime supersymmetry algebra is the same as the M-9-brane worldvolume algebra considered above. As in that case, the null 10-vector P does ‘double duty’ as a wave momentum and 9 chiral string charge,while Z+ does ‘double duty’ as the charge of a 5-brane and a KK-monopole. In the past, the necessity for charges to do ‘double duty’ in the N=1 D=10 supersymmetry algebra has always seemedadrawbackoftheapproachbutwehavejustseenthatithasanaturalexplanationfromtheM-9- brane perspective (as a result of the 0-form and 1-forminterpretations in transverse space). In any case, we may now attempt to determine the intersection of Type 1 branes by considering the supersymmetry algebrasofthe 1B-string/1B-waveandthe1B-fivebrane/1B-KK-monopole. Forsimplicitywerestrictthe discussion to the 1B string and 5-brane. The 1B-string worldsheet supertranslation algebra is the D=2 (8,0) algebra (i=1,...,8) {Qi ,Qj }=δijP +Z(ij)(35+), (δ Z(ij) =0). (25) − − = = ij = As indicated, the 1-formZ-chargesarein the 35+ representationofSO(8), whichdefines a 4-forminthe transverse 8-space. We conclude, correctly, that the 1B-string inside a 1B-5-brane preserves 1/4 of the supersymmetry of the Type 1 D=10 Minkowskivacuum. We should be able to see this intersectionfrom the 5-brane’s worldvolume supersymmetry algebra. This is the D=6 (1,0) algebra (i=1,2;α=1,···,4) {Qi,Qj}=εijP +Z(ij) . (26) α β [αβ] (αβ) TheZ-chargeisanSU(2)-tripletofself-dualD=63-forms. Thisisthechargeofa3-braneinthe5-brane. The SU(2) ∼= Spin(3) R-symmetry of the algebra (26) suggests a spacetime interpretation in D=9, but the SU(2)doubletofsuperchargescanequally wellbe interpretedas a spinorof SO(4), inwhich casean SU(2) triplet is to be interpreted as a self-dual 2-form in a transverse 4-space. Thus, the 3-brane is the intersection of two 1B-5-branes. This possibility is to be expected from the IIB origin of the 1B string. We have still to find the 1B-string in the 5-brane. Since the string is chiral it should be associated with a null 6-vector. The only possibility is P, which therefore does ‘double duty’ as a string charge and the 6-momentum of a 1B-wave in the 1B-5-brane. Another class of 1/8-supersymmetric triple intersections is obtained by considering double inter- sections of branes confined to the core of a KK-monopole. For example, consider a IIB-3-brane in a IIB-KK-monopole. The IIB-3-brane is effectively confined to a D=6 spacetime. In fact, this is one real- ization of the iib-3-brane of the ‘little’ D=6 IIB string theory, alias iib-string theory2. The worldvolume supersymmetryalgebraoftheiib-3-braneistheN=2D=4algebra. The{Q,Q}anticommutator,inWeyl spinor notation, is {Qi,Q } = δiP +Wi (Wi =0), α β˙j j αβ˙ αβ˙j αβ˙i {Qi,Qj} = ε εijZ+Y(ij) . (27) α β αβ (αβ) One sees from this that the iib-3-brane admits 1/2-supersymmetric p-branes for p=0,1,2,3. The R- symmetry group is U(2). We interpret the U(1) ∼= Spin(2) factor as the rotation group of the 2-space transverse to the worldvolume of the D=6 iib-3-brane, and the SU(2) factor as the rotational isometry group of the KK-monopole implicit in the iib-theory’s D=10 spacetime realization. The iib-p-brane charges are in the following U(2) representations Z(p=0):1 W(p=1):3 Y(p=2):3 W(p=3):3 (28) +1 0 +1 0 where the subscript on the SU(2) representations indicates the U(1) charge (Z and Y are complex). The 0-branes must be interpreted as intersections, or endpoints, of iib-strings on iib-3-branes. Thus iib-3-branes are d-branes for iib-strings, a fact that is evident from their D=10 spacetime interpretation as IIB-branes in a KK-monopole. The 1-branes are intersections with a IIB-3-brane for which the two 2Wehavefoundthediscussionof‘little’theoriesin[22]usefulinorientingourideas. Werefertothatpaperforfurther references. 10