Schwinger-Dyson approach and its application to generate a light composite scalar A. Doff1 and A. A. Natale2 1Universidade Tecnol´ogica Federal do Paran´a - UTFPR - DAFIS Av Monteiro Lobato Km 04, 6 1 84016-210, Ponta Grossa, PR, Brazil 0 2 [email protected] b e 2Instituto de F´ısica Te´orica, F UNESP Rua Dr. Bento T. Ferraz, 4 2 271, Bloco II, 01140-070, ] h S˜ao Paulo - SP, Brazil p - p Centro de Ciˆencias Naturais e Humanas, e h Universidade Federal do ABC, [ 09210-170, Santo Andr´e - SP, Brazil 2 v 4 [email protected] 2 0 Abstract 4 0 We discuss the possibility of generating a light composite scalar boson, in a scenario that we . 1 0 may generically call Technicolor, or in any variation of a strongly interacting theory, where by 6 1 light we mean a scalar composite mass about one order of magnitude below the characteristic scale : v i of the strong theory. Instead of most of the studies about a composite Higgs boson, which are X r based on effective Lagrangians, we consider this problem in the framework of non-perturbative a solutions of the fermionic Schwinger-Dyson and Bethe-Salpeter equations. We study a range of mechanisms proposed during the recent years to form such light composite boson, and verify that such possibility seems to be necessarily associated to a fermionic self-energy that decreases slowly with the momentum. 1 I. INTRODUCTION The 125 GeV new resonance discovered at the LHC [1, 2] has many of the characteristics expected for the Standard Model (SM) Higgs boson. If this particle is a composite or an elementary scalar boson is still an open question that probably will be answered at LHC13 run. ThepossibilitythatalightcompositestatecouldbediscoveredattheLHC,inascenario that we may generically call Technicolor, or in any variation of a strongly interacting theory has been extensively discussed in the literature [3]. The chiral and gauge symmetry breaking in quantum field theories can be promoted by fundamental scalar bosons through the Higgs boson mechanism. However the main ideas about symmetry breaking andspontaneous generation of fermion andgauge boson masses in field theory were based on the superconductivity theory. Nambu and Jona-Lasinio proposed one of the first field theoretical models based on the ideas of superconductivity, where all the most important aspects of symmetry breaking and mass generation, as known nowadays, wereexploredatlength[4]. ThemodelofRef.[4]containsonlyfermionspossessing invariance under chiral symmetry, althoughthis invarianceisnot respected bythevacuum ofthetheory and the fermions acquire a dynamically generated mass (m ). As a consequence of the chiral f symmetry breaking by the vacuum the analysis of the Bethe-Salpeter equation (BSE) shows the presence of Goldstone bosons. These bosons, when the theory is assumed to be the effective theory of strongly interacting hadrons, are associated to the pions, which are not true Goldstone bosons when the constituent fermions have a small bare mass. Besides these aspects Nambu and Jona-Lasinio also verified that the theory presents a scalar bound state (the sigma meson). InQuantum Chromodynamics (QCD) the same mechanism is observed, where the quarks acquire a dynamically generated mass (m ). This dynamical mass is usually expected to dyn appear as a solution of the Schwinger-Dyson equation (SDE) for the fermion propagator when the coupling constant is above a certain critical value. The condition that implies a dynamical mass for quarks breaking the chiral symmetry is the same one that generates a bound-state massless pion. This happens because the quark propagator SDE is formally the same equation binding a quark and antiquark into the massless pseudoscalar state at zero momentum transfer (the pion). As shown by Delbourgo and Scadron [5], the same similarity of equations happens for the scalar p-wave state of the BSE, indicating the presence of a 2 bound state with mass (m = 2m ) and this scalar meson is the elusive sigma meson [6–8], σ dyn that is assumed to be the Higgs boson of QCD. Inspired in QCD, Technicolor was a theory invented to provide a natural and consis- tent quantum-field theoretic description of electroweak (EW) symmetry breaking, without elementary scalar fields. A possibility raised a few years ago is to have a Higgs arising as a composite pseudo-Goldstone boson (PGB) from the strongly interacting sector, where in this case the Higgs mass is protected by an approximate global symmetry and is only generated via quantum effects, models based on this approach are usually called composite Higgs models[9]. As pointed out in Refs.[9, 10], composite Higgs models require a degree of fine-tuning of parameters and most of the studies about a composite Higgs boson are based on effective Lagrangians [9]. The approach that we will discuss here is not based on effective Lagrangians or operators, but on the dynamics of the theory, i.e. on the non-perturbative solutions of the Schwinger-Dyson and Bethe-Salpeter equations. The freedom appearing on the coefficients of the many possible operators that can be introduced into an effective La- grangian in order to describe the SM scalar boson sector, is traded now by the self-energies of the new fermions that form the composite scalar boson. Assuming that the underlying theory is a non-Abelian SU(N) gauge theory, we will verify that the restriction imposed on the theory parameters by the existence of a light scalar composite are quite tight, and the construction of a realistic “Technicolor” model may indeed be a very precise engineering problem. In a very recent paper, ATLAS and CMS Collaborations [11], based on their combined data samples presented an improved precision value for m = 125.09 0.24GeV. The H ± improved knowledge of m yields more precise predictions for the Higgs couplings and until H now the coupling strengths to SM particles are consistent within the uncertainties expected for the SM Higgs boson. Probably in a realistic composite scalar model the fermionic couplings will involve an ETC group(or GUT) with a delicate vacuum alignment between ′ two type of composite scalar bosons H and H , where one of the composites resemble a fundamental scalar and the fermionic masses are not generated as usual, by different ETC mass scales, but by a different mechanism. In the Ref.[12] we considered a TC model with a horizontal symmetry where the top quark (or the third fermionic generation) obtain its mass ′ associated to a large ETC scale, and in this case only the composite boson H is coupled to the third fermionic generation, with a coupling resembling the one of a fundamental scalar 3 boson, although the aspects of generating the SM fermionic mass spectrum will not be touched here. We just commented this in order to remember that the generation of a light scalar composite is only part of the problem if we desire the full SM dynamical symmetry breaking, where the generation of the fermionic mass spectra is another enormous step in this direction, and, probably, the larger part of the problem. Our intent in this work is not to advocate that the 125 GeV scalar boson is a composite one, but to discuss how a scalar composite boson can be generated with a mass lighter than the characteristic scale of the strongly interacting theory. Of course, we will resort to our knowledge about QCD, where the lightest scalar boson has a mass about twice the characteristic scale Λ , as well as we will make analogy to the SM where the scalar QCD responsible for the gauge symmetry breaking has a mass about one order of magnitude below that of the Fermi scale ( (1)TeV). O The advantage of the approach that we shall propose here is that it allows to discover what type of dynamics can lead to a “light” composite, and also indicates what are the types of effective Lagrangians that favor such composite particle. Indeed we shall discuss that working with the theory dynamics, i.e. self-energies or bound state solutions, we can restrict the existence of certain terms in the many possible effective Lagrangians to describe the composite Higgs boson potential. Inthisworkweconsider theproblemofgeneratingonelightcompositescalar inastrongly interacting non-Abelian gauge theory assuming a range of mechanisms developed during the recent years. In Section 2 of this work we study the problem of generating at least one light composite scalar in a theory with a unique characteristic scale Λ. This analysis will be performed with the use of the Bethe-Salpeter equation. The same result will also be checked with the use of the effective potential for composite operators in Section 3, and we make a few remarks about the possible influence of mixing between different scalars formed within the same theory in Section 4. In Section 5 we discuss how the mass of a composite Higgs boson is modified in the presence of other interactions, i.e. any interaction that is not the one responsible to form the composite scalar state. In Section 6 we consider the possibility of generating a light composite scalar happening in the case where we have at least two composite bosons, related to two different scales and there is a strong mixing between the scalars [13], i.e. we may have a see-saw mechanism where one of the scalar composites may turn out to be quite light. In Section 7 we present a brief discussion about how the mixing 4 mechanism, discussed in Section 6, can be extended to models with more than one TC group. In Section 8 we make a brief remark on the mass of vector composites, because this is one of the main signals of a new strongly interacting theory and the Section 9 contains our conclusions. II. SCALAR COMPOSITE BOSON MASS IN ISOLATION: BETHE-SALPETER EQUATION APPROACH In this section we shall consider the problem of generating at least one light composite scalar in a strongly interacting non-Abelian gauge theory with a unique characteristic scale Λ. Most of our discussion will be guided by QCD, which is the only strongly interacting non-Abelian theory that we have to compare with. The scale Λ has a similar role as the QCD scale (Λ ), where it is known that quarks generate a condensate QCD q¯q 1/3 Λ µ, (1) QCD | h i | ≈ ≈ where µ is the dynamical quark mass. At the same time that the QCD chiral symmetry is broken Goldstone bosons are formed (the pions) and a set of scalars are also generated [6–8, 14]. Considering a minimum number of quarks we shall have at least one light scalar boson (the sigma meson), whose mass is m = 2µ . (2) σ Eq.(2) comes out from the following relation [5, 15] Σ(p2) ≈ ΦPBS(p,q)|q→0 ≈ ΦSBS(p,q)|q2=4µ2 , (3) where the solution (Σ(p2)) of the fermionic Schwinger-Dyson equation (SDE), that indicates thegenerationofadynamicalquarkmassandchiralsymmetrybreakingofQCD,isasolution of the homogeneous Bethe-Salpeter equation (BSE) for a massless pseudoscalar bound state (ΦPBS(p,q)|q→0), and is also a solution of the homogeneous BSE of a scalar p-wave bound state (ΦSBS(p,q)|q2=4µ2), which implies the existence of the scalar (sigma) boson with the mass described above. Eq.(2) has a central point here, and in all subsequent discussion we will verify how this equation can be modified in the general case of a strongly interacting theory. In order to do 5 this we will work with the most general fermionic self-energy or bound state solution, i.e. a solution that may describe any possible dynamics of a SU(N) non-Abelian gauge theory. Many models have considered the possibility of a light composite Higgs based on effective Higgs potentials as reviewed in Ref.[9]. The reason for the existence of the different models (or different potentials) for a composite Higgs, is a consequence of our poor knowledge of the strongly interacting theories; that is reflected in the many possible choices of parameters in the effective potentials. On the other hand the composite scalar boson mass can be calculated based on the dynamics of the theory [16], and this approach, although more complex, is more restrictive than the analysis of potential coefficients in several specific limits. Our starting point is the most general asymptotic fermionic self-energy expression for a non-Abelian gauge theory [12, 17]: Σ(p2) µ µ2 α 1+bg2(µ2)ln p2/Λ2 −γcos(απ) . (4) ∼ p2 (cid:18) (cid:19) (cid:2) (cid:0) (cid:1)(cid:3) In the above expression Λ is the characteristic mass scale of the strongly interacting theory forming the composite Higgs boson, and for simplicity we assume Λ µ. Note that dynam- ≈ ical mass µ is not an observable. In principle, it must have a simple relation with Λ, and from the QCD experience we may expect that they are of the same order. g is the strongly interacting running coupling constant, b is the coefficient of g3 term in the renormalization group β function, γ = 3c/16π2b, (5) and c is the quadratic Casimir operator given by 1 c = [C (R )+C (R ) C (R )], (6) 2 1 2 2 2 3 2 − where C (R ) are the Casimir operators for fermions in the representations R and R that 2 i 1 2 form a composite boson in the representation R . 3 The parameter α in Eq.(4) varies between 0 and 1. When α = 1 Eq.(4) gives the known asymptotic self-energy behavior predicted by the operator product expansion (OPE) [18] µ3 Σ(1)(p2 ) . (7) → ∞ ∼ p2 When α = 0 we obtain Σ(0)(p2) µ 1+bg2(µ2)ln p2/µ2 −γ . (8) ∼ (cid:2) (cid:0) (cid:1)(cid:3) 6 The asymptotic expression shown in Eq.(8) was determined in the appendix of Ref.[19] and it satisfies the Callan-Symanzik equation. It has been argued that Eq.(8) may be a realistic solution in a scenario where the chiral symmetry breaking is associated to confinement and the gluons have a dynamically generated mass [20–22]. This solution also appears when using an improved renormalization group approach in QCD, associated with a finite quark condensate [23–25], and it minimizes the vacuum energy as long as n > 5 [26]. Moreover, f this specific solution is the only one consistent with Regge-pole like solutions [27]. Finally, the apparent explicit breaking of the chiral symmetry described by Eq.(8) seems also to be induced by the critical Wilson term in the QCD action [28], which is a quite intriguing result compatible with arguments presented in Ref.[20] about the importance of confinement for chiral symmetry breaking. The important fact is that this is the hardest (in momentum space) asymptotic behavior allowed for a bound state (or self-energy) solution in a non- Abelian gauge theory, and it is exactly for this reason that a constraint on γ arises from the BSE normalization condition [29–31] implying γ > 1/2. (9) This condition has also been re-obtained recently associated with the positivity of the scalar composite mass in Ref.[32]. In the infrared region the self-energy is approximately constant and equal to µ up to a momentum p 3µ [22]. Therefore Eq.(7) and Eq.(8) reflect the extreme limits, that can ≈ be obtained in a strongly interacting non-Abelian gauge theory, of how a self-energy can decrease with the momentum. Any possible self-energy solution of an asymptotically free SU(N) gauge theory can be described by Eq.(4) with an appropriate α value. Eq.(7) is the soft behavior dictated by OPE, while Eq.(8) is the one generated in the case where the chiral symmetry breaking is totally dominated by four-fermion interactions [33, 34], in a limit that can also be termed as extreme walking. In the QCD case many authors claim that the asymptotic self-energy is given by Eq.(7), although with 6 flavors we are quite near the conformal window [35]. Nowadays it is known that we may have solutions with a momentum behavior varying between Eq.(7) and Eq.(8) depending on the theory dynamics [33, 34]. The existence of “effective” four-fermion interactions may change the asymptotic behavior into a hard one [33, 34], and lattice simulations are beginning to study the self-energy behavior of SU(N) theories in the limit γ 1/2. ≫ 7 We can now discuss how the scalar masses can be computed with the help of Eq.(4), and how different scalar mass values will be obtained when we vary the parameter α of that equation in the range 0 to 1. The scalar mass (m ) at leading order comes out from Eq.(2), S i.e. m = 2µ. (10) S As we said µ is not an observable, and it should be written in terms of measurable quantities and by group theoretical factors of the strong interaction responsible for forming the com- posite scalar boson. In order to do so we will assume that the scalar composite responsible for the SM symmetry breaking give masses (M ) to the electroweak bosons in the same W way proposed in the traditional Technicolor (TC) models, where the dynamical mass µ is related to the technipion decay constant (F ) and to the SM vacuum expectation value (v) Π by [36] g2n F2 g2v2 M2 = w d Π = w , (11) W 4 4 where n is the number of technifermion doublets, v 246GeV, and F is given by the d Π ∼ Pagels and Stokar expression [37] N dp2p2 p2dΣ(p2) F2 = TC Σ2(p2) Σ(p2) . (12) Π 4π2 (p2 +Σ2(p2))2 − 2 dp2 Z (cid:20) (cid:21) Using Eq.(11) and Eq.(12) it is now possible to write the values of the scalar boson mass generated in a strongly interacting gauge theory in terms of the SM vacuum expectation value and the quantities that characterize a SU(N) non-Abelian theory with n fermions f forming the scalar boson. The masses, calculated with the two extreme self-energy solutions giving by Eq.(7) and Eq.(8) (associated to α = 0 or 1), are: 8π2bg2(2γ 1) 1/2 (0) m 2 v − (13) S ≈ Nn " (cid:18) f (cid:19) # 4 8π2 1/2 (1) m 2 v . (14) S ≈ 3 Nn "r (cid:18) f(cid:19) # These equations involve only known quantities if we know the strongly interacting theory. Note that the scalar boson mass, or the composite Higgs mass, scales differently with the SU(N) parameters depending on the dynamics of the theory. The factor bg2(2γ 1) may − certainly modify mass predictions when comparing Eq.(13) and Eq.(14), what cannot be obtained varying naively only N and n . However, as we shall discuss in the next paragraph, f 8 the difference between the extreme values is even more complicated. Note that positivity of Eq.(13) requires the constraint of Eq.(9) to be obeyed. The result of Eq.(10) comes out from the comparison of the homogeneous BSE with the associated SDE[36], but the full bound state properties are subjected to the non-homogeneous BSE,whichincludesitsnormalizationcondition, asclearlydiscussedbyLlewellynSmith[31]. The BSE normalization condition is given by [29] 2 F ∂ 2ı Π q = ı2 d4pTr (p) F(p,q) (p) m µ P ∂qµ P (cid:18) dyn(cid:19) Z (cid:26) (cid:20) (cid:21) (cid:27) ∂ + d4pd4kTr (k) K(p,k,q) (p) (15) P ∂qµ P Z (cid:26) (cid:20) (cid:21) (cid:27) where 1 (p) S(p)G(p)γ S(p), P ≡ (2π)2 5 F(p,q) = S−1(p+q)S−1(p), S(p) is the fermion propagator, Σ(p)/µ = G(p) and K(p,k,q) is the BSE kernel. The BSE normalization constrain the self-energy solution if this solution is a hard one. Eq.(15) can be divided into two parts: 2ı(F /µ)2q = I0 + IK . (16) Π ν ν ν Contracting the above equation with qν and computing it at q2 = m2, after some algebra S we verify that the final equation can be put in the form m2 = 4µ2 (I0 +IK), (17) S × where I0 and IK are the integrals of Eq.(15) contracted with the momentum qν and IK is a complicated expression but of (g2(p2)) when compared to I0. If we neglect the higher O order term (IK) we verify that the normalization condition changes the mass value given by Eq.(10) by a factor I0. More importantly, I0 starts being different from 1 as we go towards the limit α 0! In this limit we obtain [16, 17] → bg2(2γ 1) m(0)2 4µ2 3bg2(2γ 1)/4(1+ − ) . (18) S ≈ − 2 (cid:18) (cid:19) First, in the limit that α 0 in order to have a positive mass we recover the constraint → given by Eq.(9), and note that we have not yet written the factor µ in terms of observables 9 quantities as performed to obtain Eq.(13). The limit of Eq.(18) can be called extreme walking, or we may also say that in this limit the chiral symmetry breaking is dominated by an effective four-fermion interaction [33, 34]. Secondly, the normalization effect lowers considerably the composite scalar masses [16, 17, 22]. Eq.(18) has been analyzed in the case of different groups and number of fermions [16, 17], andit was verified that it is quite difficult to obtaina light composite scalar boson associated to the SM gauge symmetry breaking, or, for instance, a scalar boson with a expected mass of 125 GeV when the strongly interacting theory has a characteristic mass scale of O(1)TeV. Actually we cannot say that it is impossible to obtain such light composite scalar mass, however it is quite probable that in order to obtain a light composite scalar generating the SM gauge symmetry breaking we must have a large number of fermions or fermions in a higher dimensional representation. Refs.[16, 22] contain tables and figures of composite scalar masses for different groups and number of fermions. This means that any composite scalar boson candidate to be the SM Higgs boson, when considered in isolation, will be associated to a theory with a large chiral symmetry breaking, generating a large number of Goldstone bosons. This comment will be particularly important when discussing the possibility of a light composite Higgs boson in the presence of other interactions. Of course, the result that we discussed here does not need to be linked to the SM gauge symmetry breaking, and Eq.(18) tell us that a composite scalar boson, in a non-Abelian gauge theory, much lighter than its characteristic scale can be generated only for very hard asymptotic self-energies. The condition γ > 1/2 seems to be crucial to obtain a small composite scalar mass as shown in Eq.(18), however this condition was not explored at depth in Ref.[16, 17], and it is interesting to confront this condition with lattice simulations that study the conformal region of SU(N) theories. In the case where fermions are in the fundamental representation of the SU(N) group Eq.(9) implies in the following inequality 9(N2 1) − > 1. (19) N(11N 2n ) f − In the SU(3) and SU(2) cases this constraint indicates that we must have n larger than 5, f and according Eq.(19) the minimum n value increases slowly as we increase N. The most f recent lattice data about the conformal window in SU(3), which may lead to small scalar masses, indicates that its lower edge may be between n = 6 and n = 8 [35], whereas in f f 10