The Standard Model Fermion Spectrum from Complex Projective spaces

arXiv:hep-th/0207078 v2 14 Nov 2002 Preprint typeset in JHEP style - HYPER VERSION DIAS-STP-02-07 The Standard Model Fermion Spectrum From Complex Projective Spaces Brian P. Dolan and C. Nash Dept. of Mathematical Physics, NUI, Maynooth, Ireland and School of Theoretical Physics, Dublin Institute for Advanced Studies, 10 Burlington Rd., Dublin 8, Ireland [email protected], [email protected] Abstract: It is shown that the quarks and leptons of the standard model, including a right-handed neutrino, can be obtained by gauging the holonomy groups of complex projective spaces of complex dimensions two and three. The spectrum emerges as chiral zero modes of the Dirac operator coupled to gauge fields and the demonstration involves an index theorem analysis on a general complex projective space in the presence of topologically non-trivial SU(n)×U(1) gauge fields. The construction may have applications in type IIA string theory and non-commutative geometry. Keywords: Field Theories in Higher Dimensions, Differential and Algebraic Geometry, Non-Commutative Geometry. Contents 1. Introduction 1 2. The Particle Spectrum From CPn3 3. Conclusions 6 A. The Index Theorem For CPn8 1. Introduction It has been a long standing goal of theoretical particle physics to unify space-time symmetries with the internal SU(3)×SU(2)×U(1) gauge symmetry of the standard model. Most current research in this direction relies on string theory in higher dimensions, hoping to derive grand unified theories as low energy limits of the full theory, and a crucial aspect of this programme is the rˆole of compact internal spaces. Compact internal spaces were first introduced into physics in Kaluza-Klein models where a coset space G/H with isometry group Gand holonomy group Hcan give rise to a gauge group Gin 4-dimensional space-time. A pure Kaluza-Klein approach was largely abandoned in the 80s due in part to the realisation that it was difficult, if not impossible, to obtain chiral Fermions this way [1]. In this paper we take a different approach to internal coset spaces, focusing on the holonomy group Hrather than G. We shall show that a single generation of the standard model spectrum, including a right handed neutrino, can arise from the complex projective spaces CP2and CP3. The approach adopted here has the attractive feature that it is somewhat more in tune with the spirit of general relativity than standard Kaluza-Klein theory. In the standard approach one assumes that the coset space has a specific metric which has isometry group G, which is somewhat contrary to the philosophy general relativity where a particular metric is merely one solution of Einstein’s field equations and there may be many others with smaller isometries, indeed a generic solution has none. In the construction used here it is the holonomy group that is important and for a complex manifold with real dimension 2nthis group is generically U(n). For 4-dimensional space time equipped with a Lorentzian metric the holonomy group is generically SO(1,3) and general relativity can be viewed as a gauge theory of tangent space rotations using this group. Incorporating spinors into general relativity requires –1– gauging the double cover of SO(1,3), namely Spin(1,3) ∼ =Sl(2,C). If a compact complex internal space with real dimension 2nis added the extra holonomy group is generically U(n) and so it seems quite natural to gauge SU(n)×U(1). Taking two internal spaces with n= 2 and n= 3 would give SU(3)×SU(2)×U(1)×U(1) which is not quite what is required, but is close. To see how the standard model spectrum might arise in this way consider first a Dirac spinor on a general compact complex manifold of real dimension 4. This has 4 components and decomposes as (2,1) + (1,2)−→ 20+ (11+1−1) (1.1) under Spin(4) ∼ =SU(2) ×SU(2) →SU(2) ×U(1). Shifting the U(1) charge by −1 and rescaling the result by 1/2 gives 2−1/2+ (10+1−1).(1.2) On the other hand a compact complex manifold of real dimension 6 has 8 component Dirac spinors which decompose as 4+4−→ (31+1−3) + (3−1+13) (1.3) under Spin(6) ∼ =SU(4) →SU(3) ×U(1). Taking a single chiral spinor in the 4of SU(4), shifting the U(1) charge by 3 and rescaling the result by 1/6, gives (32/3+10).(1.4) A single generation of the standard model spectrum, including a right-handed neutrino, can be obtained by tensoring (1.2) with (1.4) and simply adding the U(1) charges. These group theory decompositions are valid on any complex manifolds of real dimension 4 and 6 respectively and the shift in the U(1) charges may be achieved by coupling the Fermions to U(1) gauge fields with appropriate topological (monopole) charges. Mathematically this requires tensoring the spin bundle with an appropriate line bundle, which may or may not be possible on any given manifold depending on whether or not appropriate line bundles exist. It is shown in section 2 that for CP2and CP3the required line bundles do indeed exist (indeed for CP2spinors cannot even be defined unless such a line bundle is introduced). However obtaining the standard model spectrum requires introducing some natural higher rank bundles as well. Furthermore, an index theorem analysis shows that the representations (1.2) and (1.4) can be realised as zero modes of the Dirac operator with precisely the correct handedness for one generation of the standard model, including a right-handed neutrino. This is a topological statement, it is completely independent of the choice of metric on these spaces. Of course there is the question of generations—the construction presented here only produces a single generation of the standard model. This point is elaborated –2– on in the concluding section 3, along with some suggestions of how the construction might fit into viable models such as type IIA superstring theory and non-commutative geometry. Details of the index theorem on CPn, required for the analysis in the text, are contained in an appendix, where closed form expressions for the index of the Dirac operator coupled to topologically non-trivial SU(n)×U(1) gauge fields are derived. 2. The Particle Spectrum From CPn In this section it will be argued that the spectrum of a single generation of the standard model can be obtained from CP2×CP3by gauging the holonomy group, U(2)×U(3), with the two U(1) factors identified in one particular combination. This generalises the case of CP2which gives the electroweak sector [2]. The analysis is based on the index theorem for the Dirac operator on CPn∼ =SU(n+ 1) U(n)(2.1) derived in the appendix. We are interested in zero modes of the Dirac operator, for Fermions which are either singlets or transform under the fundamental nrepresentation of SU(n), when topologically non-trivial gauge fields are present (CPn analogues of the monopole field on S2and the SU(2) BPST instanton on S4). The gauge group is taken to be the holonomy group U(n) of CPn. The index is the difference between the number of positive chirality zero modes of the Dirac operator and the number of negative chirality zero modes and, for convenience, the important formulae from the appendix are reproduced here. For a singlet with U(1) charge Y(n)=q, where qis an integer, the index is νq=1 n!(q+ 1) ···(q+n), q ∈Z(2.2) corresponding to a line bundle over CPnwith gauge group U(1). A Fermion in the fundamental representation of SU(n), with a U(1) charge Y(n)=q+1 n, has index νq,n=(q+ 1) ···(q+ (n−1))(q+n+ 1) (n−1)! .(2.3) We are free to rescale the U(1) charges differently for different values of n, but for a given nthe ratio of the singlet charge qto the fundamental charge q+1 nis fixed. The simplest case is CP1∼ =S2, where the holonomy group is U(1). Introducing U(1) gauge fields and Fermions with charge qthere are νq=q+ 1 zero modes of the Dirac operator. The case q=−1 gives no unpaired zero modes—generically there are none at all but even when zero modes exist left and right-handed particles with charge −1 occur in pairs, like an electron in QED. For q6=−1 the theory is necessary chiral. For example choosing the convention that positive νcorresponds to –3– right-handed spinors in 4-dimensions, q=−2 would be interpreted as a left-handed particle with hypercharge −2. However there is no place in this construction for a right-handed particle with the same charge as this would require q=−2 and ν > 0. This is necessarily a chiral theory for q6=−1, and would generate a lethal gauge anomaly in 4-dimensions. To introduce quarks we turn to CP3—since the holonomy group of CP3is U(3) quarks can be introduced as triplets. For example an SU(3) triplet with q=−3 gives ν−3,3= 1 and has U(1) charge Y(3) =−8/3. Since the overall normalisation of the charge is at our disposal, this could represent either: a right handed d-quark (rescale charge by 1/8), dR=3−1/3,(2.4) or a right-handed u-quark (rescale charge by −1/4), uR=32/3.(2.5) Either possibility forces us to interpret positive chirality as corresponding to righthanded Fermions in 4-dimensions in order to match the standard model spectrum. It is instructive to examine the complex conjugate representations in order to understand the CPT conjugate of the spectra. In the appendix it is shown that under complex conjugation of the bundles over CPnthe U(1) charges transform as Y(n)→Y(n)=−Y(n)−(n+ 1) and the index as ν→ν= (−1)nν. So for the examples (2.4) and (2.5) above, with just a single SU(3) triplet on CP3with Y(3) =−8/3, Y(3) =−4/3 and complex conjugation flips the chirality so it maps dR=3−1/3−→ 3−1/6(2.6) and uR=32/3−→ 31/3= (d)L.(2.7) If we start with a right-handed u-quark complex conjugation forces us to introduce the left-handed anti-dwhile starting with a right-handed d-quark complex conjugation forces the introduction of a state that has no place in the standard model. This nicely illustrates the kind of constraints that CPT can place on the possible choices. For weak interactions we bring in CP2with holonomy group U(2). An index theorem analysis then allows us to obtain a single generation of the electroweak sector, including a right-handed neutrino. This is achieved by taking the following three representations: •An SU(2) singlet with q= 0 giving zero charge and index ν0= +1; •A second singlet with q=−3 giving charge Y(2) =−3 and index ν−3= +1; •An SU(2) doublet with q=−2 giving charge Y(2) =−3/2 and ν−2,2=−1. –4– Interpreting positive νas giving right-handed spinors, and rescaling the charge by 1/3, this results in a single generation of particles of the electroweak sector of the standard model, including a right-handed neutrino: 10= (V)R1−1=eR2−1/2=VL eL(2.8) (the normalisation is such that the electric charge is Q=I3+Y). Now for CP2 complex conjugation preserves the chirality and sends Y(2) →Y(2) =−Y(2) −3, so complex conjugation maps (2.8) to 1−1=eR10= (V)R2−1/2=eL VL(2.9) (for the 2the electric charge is Q=−I3+Yand of course the 2is equivalent to the 2under rotation by the Pauli matrix iσ2). The curious conclusion is that, contrary to what one might na¨ıvely expect, complex conjugation does not change the sign of the hypercharge but instead interchanges the electron and neutrino, leaving the electroweak multiplet in (2.8) invariant. One complete generation of the quark sector of the standard model can be obtained by combining (2.8) with the SU(3) triplet in (2.5), provided the total hypercharge is defined as a particular linear combination of the CP2and the CP3charges, Y=−1 4Y(3) +1 3Y(2). Taking the total chirality to be the product of the two individual chiralities, and interpreting positive chirality as right-handed, gives the particle spectrum of the strong sector of the standard model: (3,1)2/3=uR(3,1)−1/3=dR(3,2)1/6=uL dL.(2.10) The electroweak sector can be included by combining an SU(3) singlet on CP3, with zero charge and ν0= 1, with (2.9): (1,1)0=VR(1,1)−1=eR(1,2)−1/2=VL eL.(2.11) Equations (2.10) and (2.11) constitute a single generation of the standard model. Now observe that the combination Y=1 4Y(3) −1 3Y(2) does simply change sign under complex conjugation, Y→Y=−Y. In addition the CP3chirality changes while that of CP2does not so the overall chirality, which is the product of the two, flips. The net result is that complex conjugation of (2.10) and (2.11) does indeed reproduce the anti-particles: (3,1)2/3=uR,(3,1)−1/3=dR,(3,2)1/6=uL dL(2.12) 7−→ c.c. (3,1)−2/3= (u)L,(3,1)1/3= (d)L,(3,2)−1/6=(u)R (d)R –5– and (1,1)0=VR,(1,1)−1=eR,(1,2)−1/2=VL eL(2.13) 7−→ c.c. (1,1)0= (V)L,(1,1)1= (e)L,(1,2)1/2=(V)R (e)R. It is interesting that the construction presented here necessarily requires the introduction of a right-handed neutrino, as recent experimental evidence for neutrino oscillations requires just such a state for the simplest explanation of the results [3], [4]. In summary a single complete generation of the standard model, (2.10) and (2.11), arises from a selection of five different bundles, three over CP2and two over CP3: •Two SU(2) singlets on CP2with q= 0 and q=−3 and an SU(2) doublet with q=−2; •An SU(3) singlet on CP3with q= 0 and an SU(3) triplet with q=−3. In terms of the Spincstructures described in the appendix these correspond to the two bundles on CP2∧0,∗TCP2⊗1⊕L3 (2) ⊕F(2) ⊗L2 (2) (2.14) on CP3∧0,∗TCP3⊗1⊕F(3) ⊗L3 (3) (2.15) where, on CP2,L(2) is the generating line bundle and F(2) the rank 2 bundle satisfying L(2) ⊕F(2) =I3, while, on CP3,L(3) is the generating line bundle and F(3) the rank 3 bundle satisfying L(3) ⊕F(3) =I4, with I3and I4denoting trivial bundles of rank 3 and 4 respectively. 3. Conclusions It has been shown that a single generation of the standard model Fermion spectrum, including a right-handed neutrino, can be obtained from the holonomy groups of CP2 and CP3by tensoring with appropriate bundles. Physically this means introducing background SU(3) ×SU(2) ×U(1) gauge fields which are topologically non-trivial, containing analogues of monopoles and instantons. A number of questions present themselves. Firstly there is no obvious sign of three generations. Of course one can obtain more generations by taking copies, but there seems no compelling reason to take three such copies and not some other number. This may be related to the question of what possible rˆole the isometry group may play. In the introduction a virtue –6– was made of the fact that the construction is not tied down to a specific metric on CPn, but if one introduces the Fubini-Study metric one then has isometry group SU(n+ 1). On CP2one has SU(3) and, using this as a horizontal generation group, the fundamental representation would give three generations. But then it is not clear what the rˆole of the SU(4) from CP3would be. Alternatively it is possible to manufacture three copies by including CP1with q= 2, giving ν2= 3, and then simply ignoring the U(1) charge on CP1. For the moment we have no compelling suggestion as to how the generations might appear and we leave this as an open question. Secondly there is the question of the extra U(1). The holonomy group of CP2× CP3is U(3)×U(2) ∼ =SU(3)×SU(2)×U(1)×U(1) and we have taken a U(1) which is only one particular combination of the two U(1) factors, ignoring the other one. In fact the standard model spectrum has true group S(U(3) ×U(2)), [5], and the very fact that it is reproduced here means that the group being used is S(U(3) ×U(2)) rather than the full holonomy group U(3) ×U(2). There may be a deeper reason for this, but for the moment we confine ourselves to observing that it works empirically. How might the construction fit into realistic models? It would certainly seem too na¨ıve to take CP2×CP3as an internal space—it has real dimension 10 which is too large for string theory and simply adding it on to 4-dimensional space-time produces a 14-dimensional space-time which would have anomalies. It may be that one could realise CP2and CP3as a brane within a brane in type IIA string theory, which has anti-symmetric tensor fields of rank 2 and 4 in its R-R sector as well as their hodge duals, though IIA string string theory is a non-chiral theory so the Weyl Fermion on CP3would have to be put in by hand. Any such interpretation would necessarily be rather different to the standard approach, as it would not involve grand unified theories directly. Alternatively a “fuzzy” Kaluza-Klein approach may be of interest where the continuum manifolds of CP2and CP3are replaced by non-commutative finite dimensional matrix approximations with a finite number of degrees of freedom [2]. It would then be more appropriate to think of multiple copies of 4-dimensional space-time rather than an internal continuous manifold, somewhat analogous to Connes’ non-commutative geometry approach to the standard model with two copies of space-time [6], [7]. Star-products on CPnwere studied in [8] and [9], and the spectrum of the Dirac operator was investigated in [10] and [11]. The smallest vector space that could be used for a non-trivial matrix representation is n+ 1. For CP2we get 3, which relates to the discussion of the generation problem above, while CP2×CP3would require 3 ×4 = 12 copies. While these are all interesting and important problems we leave them open for further work. –7– A. The Index Theorem For CPn For a general complex manifold Xof dimension nthe total Chern class is the sum of the individual Chern classes c(X) = 1 + c1(X) + c2(X) + ···+cn(X).(A.1) In line with common usage we write ck(X) for ck(TX)—the kth Chern class of the tangent bundle. In particular cn(X) is the top form and the Gauss–Bonnet theorem states that evaluating cn(X) on Xgives the Euler characteristic of X: i.e. cn(X)[X] = χ(X). For CPnthe Chern classes are all generated by a single 2 dimensional class x [12] c(CPn) = (1 + x)n+1 = 1 + (n+ 1)x+n(n+ 1) 2x2+···+ (n+ 1)xn.(A.2) Note that xn+1 is a (2n+2) dimensional class and so xn+1 = 0—when xis represented by a 2 form ω, say, this corresponds to the fact that ωn+1 = 0 on a 2n-dimensional manifold. The normalisation is such that ZCPn ωn= 1 (A.3) and so, since cn(X) = (n+ 1)xn≡(n+ 1)ωn, we have χ(CPn) = n+ 1,(A.4) which is the Euler characteristic of CPnas required by the Gauss-Bonnet theorem. The form ωis the curvature of a line bundle Lwhose complex conjugate Lwe shall refer to as the generating line bundle over CPnwith Chern class c(L) = 1 −x. (A.5) Another line bundle that will be important is the canonical line bundle Kwhich is the maximum exterior power of the cotangent bundle T∗X: i.e. K=∧nT∗X. (A.6) Khas Chern class given by c(K) = 1 −(n+ 1)x. (A.7) The minus sign appears because c1(K) = c1(T∗X) = −c1(TX).(A.8) –8–