ITEP-LAT/2009-10 The fluctuational region on the phase diagram of lattice Weinberg - Salam model M.A. Zubkova,b 0 1 0 a ITEP, B.Cheremushkinskaya 25, Moscow, 117259, 2 n Russia a J 5 b Moscow Institute of Physics and Technology, 1 141700, Dolgoprudnyi, Moscow Region, Russia ] t a Abstract l - p The lattice Weinberg - Salam model without fermions is investi- e gated numerically for the realistic choice of bare coupling constants h [ correspondent to the value of the Weinberg angle θ ∼ 30o, and the W finestructureconstantα ∼ 1 . Onthephasediagramthereexiststhe 3 100 v vicinity of the phase transition between the physical Higgs phase and 6 the unphysical symmetric phase, where the fluctuations of the scalar 0 1 fieldbecomestrong. TheclassicalNambumonopolecanbeconsidered 4 as an embryo of the unphysical symmetric phase within the physi- . 9 cal phase. In the fluctuational region quantum Nambu monopoles 0 are dense and, therefore, the perturbation expansion around trivial 9 vacuum cannot be applied. The maximal value of the cutoff at the 0 : given values of coupling constants calculated usingthe lattices of sizes v i 83 ×16, 123 ×16, and 164 is Λc ∼ 1.4±0.2 Tev. As the lattice sizes X used are rather small we consider this result as preliminary. r a 1 Introduction In some phenomenological models that describe condensed matter systems1 there exists the vicinity of the finite temperature phase transition that is called fluctuational region. In this region the fluctuations of the order pa- rameter become strong. The contribution of these fluctuations to certain 1One ofthe examples of suchmodels is the Ginzburg - Landautheory of superconduc- tivity. 1 physical observables becomes larger than the tree level estimate. Thus the perturbation theory in these models fails down within the fluctuational re- gion. Our main supposition is that the lattice Weinberg - Salam model (at T = 0) looks similar to the mentioned models. Namely, we expect that there exists the vicinity of the phase transition between the Higgs phase and the symmetric phase in the Weinberg - Salam model, where the fluctuations of the scalar field become strong and the perturbation expansion around trivial vacuumcannotbeapplied. Accordingtothenumerical resultsthecontinuum theory is to be approached within the vicinity of the phase transition, i.e. the cutoff is increased along the line of constant physics when one approaches the point of the transition. That’s why we expect that the conventional prediction on the value of the cutoff admitted in the Standard Model based on the perturbation theory may be incorrect. According to the conventional point of view the upper bound Λ on the cutoff in the Electroweak theory (without fermions) depends on the Higgs mass. It is decreased when the Higgs mass is increased. And at the Higgs mass around 1 Tev Λ becomes of the order of M . At the same time for H M ∼ 200 Gev the value of Λ can be made almost infinite2. This conclusion H is made basing on the perturbation expansion around trivial vacuum. Inthepresent paperwereporttheresults ofthenumerical investigationof themodelatthevalueofthescalarselfcouplingλ = 0.009,thebareWeinberg angle θ = 30o, and the renormalized fine structure constant around 1/100. W The bare value of the Higgs boson mass is around 270 Gev in the vicinity of the phase transition. We calculate the constraint effective potential V(|Φ|) for the Higgs field Φ. In the physical Higgs phase this potential has a minimum at a certain nonzero value φ of |Φ|. This shows that the spontaneous breakdown of the m Electroweak symmetry takes place as it should. However, there exists the vicinityofthephasetransition, wherethefluctuationsoftheHiggsfieldareof the order of φ while the hight of the ”potential barrier”3 H = V(0)−V(φ ) m m 2Here we do not consider vacuum stability bound on the Higgs mass related to the fermion loops. 3The meaning of the words ”potential barrier” here is different from that of the one - dimensionalquantummechanicsasheredifferentminimaofthe potentialformthe three- dimensionalspherewhile inusual1D quantummechanicswiththe similarpotentialthere are two separated minima with the potential barrier between them. Nevertheless we feel it appropriate to use the chosen terminology as the value of the ”potential barrier hight” 2 is of the order of V(φ +δφ)−V(φ ), where δφ is the fluctuation of |Φ|. We m m expect that in this region the perturbation expansion around trivial vacuum Φ = (φ ,0)T cannot be applied. We call this region of the phase diagram m the fluctuational region (FR) in analogy to the condensed matter systems. The mentioned supposition is confirmed by the investigation of the topo- logical defects composed of the lattice gauge fields that are to be identified with quantum Nambu monopoles [1, 2, 3]. We show that their lattice den- sity increases when the phase transition point is approached. Within the FR these objects are so dense that it is not possible at all to speak of them as of single monopoles 4. Namely, within this region the average distance between the Nambu monopoles is of the order of their size. Such complicated config- urations obviously have nothing to do with the conventional vacuum used of the continuum perturbation theory. We have estimated the maximal value of the cutoff in the vicinity of the transition point. The obtained value of the cutoff appears to be around 1.4 Tev. 2 The lattice model under investigation The lattice Weinberg - Salam Model without fermions contains gauge field U = (U,θ) (where U ∈ SU(2), eiθ ∈ U(1) are realized as link variables), and the scalar doublet Φ , (α = 1,2) defined on sites. α The action is taken in the form 1 S = β ((1− 1 TrU )+ (1−cosθ ))+ 2 p tg2θ p plaqXuettes W −γ Re(Φ+U eiθxyΦ)+ (|Φ |2 +λ(|Φ |2 −1)2), (1) xy x x Xxy Xx where the plaquette variables are defined as U = U U U∗ U∗ , and θ = p xy yz wz xw p θ + θ − θ − θ for the plaquette composed of the vertices x,y,z,w. xy yz wz xw measures the difference between the potentials with and without spontaneous symmetry breaking. 4It has been shown in [4] that at the infinite value of the scalar self coupling λ = ∞ moving along the line of constantphysics we reach the point on the phase diagramwhere the monopole worldlines begin to percolate. This point was found to coincide roughly with the position of the transition between the physical Higgs phase and the unphysical symmetric phase of the lattice model. This transition is a crossover and the ultraviolet cutoff achieves its maximal value around 1.4 Tev at the transition point. 3 Here λ is the scalar self coupling, and γ = 2κ, where κ corresponds to the constantusedintheinvestigationsoftheSU(2)gaugeHiggsmodel. θ isthe W Weinberg angle. Bare fine structure constant α is expressed through β and θ as α = tg2θW . In our investigation we fix bare Weinberg angle equal W πβ(1+tg2θW) to 30o. The renormalized fine structure constant can be extracted through the potential for the infinitely heavy external charged particles. All simulations were performed on lattices of sizes 83 ×16 and 123 ×16. Several points were checked using the larger lattice (164). In order to simulate the system we used Metropolis algorithm. The ac- ceptance rate is kept around 0.5 via the automatical self - tuning of the suggested distribution of the fields. At each step of the suggestion the ran- dom value is added to the old value of the scalar field while the old value of Gauge field is multiplied by random SU(2)⊗U(1) matrix. We use Gaus- sian distribution both for the random value added to the scalar field and the parameters of the random matrix multiplied by the lattice Gauge field. We use two independent parameters for these distributions: one for the Gauge fields and another for the scalar field. The program code has been tested for the case of frozen scalar field. And the results of the papers [2] are repeated. We also have tested our code for the U(1) field frozen and repeat the results of [6]. For the values of couplings used on the lattice 164 the autocorrela- g tion time for the gauge fields is estimated as about N ∼ 2500 Metropolis auto steps. (The correlation between the values of the gauge field is less than 3% for the configurations separated by Ng Metropolis steps. Each metropolis auto step consists of the renewing the fields over all the lattice.) The autocorre- lation time for the scalar field is much less Nφ ∼ 20. The estimated time auto for the preparing the equilibrium starting from the cold start is about 18000 Metropolis steps. 3 Phase diagram In the three - dimensional (β,γ,λ) phase diagram the transition surfaces are two - dimensional. The lines of constant physics onthe tree level arethe lines ( λ = 1 MH2 = const; β = 1 = const). We suppose that in the vicinity γ2 8βM2 4πα W of the transition the deviation of the lines of constant physics from the tree level estimate may be significant. However, qualitatively their behavior is the same. Namely, the cutoff is increased along the line of constant physics when γ is decreased and the maximal value of the cutoff is achieved at the 4 Figure 1: The phase diagram of the model in the (γ,λ)-plane at β = 12. The dashed line is the tree - level estimate for the line of constant physics correspondenttobareM0 = 270Gev. Thecontinuouslineisthelineofphase H transition between the physical Higgs phase and the unphysical symmetric phase. transition point. Nambu monopole density in lattice units is also increased when the ultraviolet cutoff is increased. At β = 12 the phase diagram is represented on Figure 1. The physical Higgs phase is situated right to the transition line. The position of the transitionis localizedat thepoint where thesusceptibility extracted fromthe Higgs field creation operator achieves its maximum. The following variable is considered as creating the Z boson: Z = Zµ = sin[Arg(Φ+U eiθxyΦ )]. (2) xy x x xy y We use the susceptibility χ = hH2i−hHi2 extracted from H = Z2 . We y xy observe no difference between the values of the susceptibility calcuPlated using the lattices of the sizes 83 ×16, 123 ×16, and 164. This indicates that the transition may be a crossover. It is worth mentioning that the value of the renormalized Higgs boson mass does not deviate significantly from its bare value. Namely, for λ around 0.009 and γ in the vicinity of the phase transition bare value of the Higgs mass is around 270 Gev while the observed renormalized value is 300 ± 70 Gev (see the next section for the details). 5 4 Masses and the lattice spacing In order to evaluate the masses of the Z-boson and the Higgs boson we use the correlators: 1 h ZµZµi ∼ e−MZ|x0−y0| +e−MZ(L−|x0−y0|) (3) N6 x y Xx¯,y¯ Xµ and 1 (hH H i−hHi2) ∼ e−MH|x0−y0| +e−MH(L−|x0−y0|), (4) x y N6 Xx¯,y¯ Here the summation is over the three “space” components of the four - x¯,y¯ vectors x and y whilePx ,y denote their “time” components. N is the lattice 0 0 length in ”space” direction. L is the lattice length in the ”time” direction. In lattice calculations we used two different operators that create Higgs bosons: H = |Φ| and H = Z2 . In both cases H is defined at the site x x y xy x x, the sum is over its neigPhboring sites y. y After fixPing the unitary gauge (Φ = 0; Φ ∈ R; Φ ≥ 0), lattice Elec- 2 1 1 troweak theory becomes a lattice U(1) gauge theory with the U(1) gauge field A = Aµ = [−Z′ +2θ ]mod2π, (5) xy x xy where the new lattice Z - boson field (different from (2)) is defined as Z′ = Arg(Φ+U eiθxyΦ ). (6) x xy y The usual Electromagnetic field is related to A as A = A + Z′ − EM 2sin2θ Z′. W The physical scale is given in our lattice theory by the value of the Z- boson mass Mphys ∼ 91 GeV. Therefore the lattice spacing is evaluated to Z be a ∼ [91GeV]−1M , where M is the Z boson mass in lattice units. The Z Z similar calculations have been performed in [4] for λ = ∞. It has been shown that the W - boson mass unlike M depends strongly on the lattice size due Z the photon cloud. Therefore the Z - boson mass was used in [4] in order to fix the scale. That’s why in the present paper we do not consider the W - boson mass. Our data obtained on the lattice 83 × 16 shows that Λ = π = (π × a 91 GeV)/M is increased slowly with the decrease of γ at any fixed λ. We Z investigated carefully the vicinity of the transition point at fixed λ = 0.009 and β = 12. It has been found that at the transition point γ = 0.273±0.002 c 6 the value of Λ is equal to 1.4±0.2 Tev. The check of a larger lattice (of size 123×16)doesnotshowanessential increaseofthisvalue. WealsocalculateΛ onthelattice164 at thetwo points(one isat thetransitionpoint andanother is within the physical phase). Again we do not observe the increase of Λ. However, at the present moment we do not exclude that such an increase can be observed on the larger lattices. That’s why careful investigation of the dependence of Λ on the lattice size (as well as on λ) must be performed in order to draw the final conclusion. On Fig. 3 the dependence of M in Z lattice units on γ is represented at λ = 0.009 and β = 12. In the Higgs channel the situation is much more difficult. First, due to the lack of statistics we cannot estimate the masses in this channel using the correlators (4) at all considered values of γ. At the present moment we can represent the data at the two points on the lattice 83 ×16: (γ = 0.274, λ = 0.009, β = 12) and (γ = 0.290, λ = 0.009, β = 12). The first point roughlycorresponds to thepositionofthe transitionwhile thesecond point is situateddeep withintheHiggsphase. Thesets ofcoupling chosen correspond to bare Higgs mass around 270 Gev. That’s why in this channel, in principle, bound states of the gauge bosons may appear. This situation was already considered in earlier studies of SU(2) Gauge - Higgs model (see, for example, [7] and references therein). Following these studies we interpret the mass found in this channel at small ”time” separations as the Higgs mass. We suppose that the bound states of gauge bosons may appear in correlator (4) at larger ”time” separations. At the point (γ = 0.274, λ = 0.009, β = 12) we have collected enough statistics to calculate correlator (4) up to the ”time” separation |x −y | = 0 0 4. The value γ = 0.274 corresponds roughly to the position of the phase transition. The mass found in this channel in lattice units is ML = 0.75±0.1 H while bare value of M is M0 ∼ 270 Gev. At the same time ML = 0.23± H H Z 0.007. Thus we estimate at this point M = 300±40 Gev. H At the point (γ = 0.29, λ = 0.009, β = 12) we calculate the correlator with reasonable accuracy up to |x − y | = 3. At this point bare value of 0 0 M is M0 ∼ 260 Gev while the renormalized Higgs mass in lattice units is H H ML = 1.2±0.3. At the same time ML = 0.41±0.01. Thus we estimate at H Z this point M = 265±70 Gev. H It is worth mentioning that in order to calculate Z - boson mass we fit the correlator (3) for 8 ≥ |x −y | ≥ 1. In order to calculate the Higgs boson 0 0 mass at γ = 0.274 we use the data for the correlator (4) at 4 ≥ |x −y | ≥ 0. 0 0 In order to calculate the Higgs boson mass at γ = 0.29 we use the correlator 7 Figure 2: Z - boson mass in lattice units at λ = 0.009 and β = 12. Circles correspondtolattice83×16. Trianglescorrespondtolattice123×16. Squares correspond to lattice 164 (the error bars are about of the same size as the symbols used). for 3 ≥ |x −y | ≥ 0. 0 0 5 Effective constraint potential We have calculated the constraint effective potential for |Φ| using the his- togram method. The calculations have been performed on the lattice 83×16. The probability h(φ) to find the value of |Φ| within the interval [φ−0.05;φ+ 0.05) has been calculated for φ = 0.05+N ∗0.1, N = 0,1,2,... This proba- bility is related to the effective potential as h(φ) = φ3e−V(φ). That’s why we extract the potential from h(φ) as V(φ) = −logh(φ)+3logφ (7) It is worth mentioning that h(0.05) is calculated as the probability to find the value of |Φ| within the interval [0;0.1]. Within this interval logφ is ill defined. That’s why we exclude the point φ = 0.05 from our data. Instead we calculate V(0) using the extrapolation of the data at 0.15 ≤ φ ≤ 2.0. The extrapolation is performed using the polynomial fit with the powers of φ up to the third (average deviation of the fit from the data is around 1 per 8 Figure 3: The effective constraint potential at λ = 0.009 and β = 12. Black squares correspond to γ = 0.273. Empty squares correspond to γ = 0.29. c Triangles correspond to γ = 0.279. The error bars are about of the same size as the symbols used. cent). Next, we introduce the useful quantity H = V(0)−V(φ ), which is m called the potential barrier hight (here φ is the point, where V achieves its m minimum). In Table 1 we represent the values of φ and H for λ = 0.009, β = 12. m One can see that the values of φ and H increase when γ is increased. At m γ = 0.273 the minimum of the potential is at φ = 0. This point corresponds to the maximum of the susceptibility constructed of the Higgs field creation operator. Atγ = 0.274wealsoobserve theonlyminimum forthepotentialat φ = 0. At γ = 0.275 minimum of the potential is observed at φ = 0.85±0.1 m with the very small barrier hight. That’s why we localize the position of the transition point at γ = 0.273±0.002. It is important to understand which value of barrier hight can be consid- ered as small and which value can be considered as large. Our suggestion is to compare H = V(0)−V(φ ) with H = V(φ +δφ)−V(φ ), where δφ m fluct m m is the fluctuation of |Φ|. From Table 1 it is clear that there exists the value of γ (we denote it γ ) c2 such that at γ < γ < γ the barrier hight H is of the order of H while c c2 fluct for γ << γ the barrier hight is essentially larger than H . The rough c2 fluct estimate for this pseudocritical value is γ ∼ 0.278. c2 9 Table 1: The values of φ , H, H , and Nambu monopole density ρ at m fluct selected values of γ for λ = 0.009, β = 12 (Lattice 83 ×16.) γ φ H H ρ m fluct 0.273 0 0 0.1±0.1 0.098±0.001 0.274 0 0 0.04±0.1 0.081±0.001 0.275 0.85±0.1 0.01±0.06 0.15±0.05 0.067±0.001 0.276 1.05±0.1 0.05±0.06 0.16±0.01 0.054±0.001 0.277 1.25±0.05 0.19±0.05 0.25±0.05 0.044±0.001 0.278 1.35±0.1 0.28±0.07 0.25±0.06 0.035±0.001 0.279 1.45±0.05 0.5±0.06 0.25±0.06 0.028±0.001 0.282 1.75±0.05 1.04±0.07 0.31±0.07 0.014±0.001 0.284 1.95±0.05 1.41±0.08 0.38±0.08 0.0082±0.0005 0.286 2.05±0.05 1.86±0.08 0.35±0.08 0.0049±0.0002 0.288 2.15±0.05 2.33±0.08 0.32±0.07 0.0029±0.0002 0.29 2.25±0.05 2.82±0.08 0.44±0.08 0.0017±0.0001 We estimate the fluctuations of |Φ| to be around δφ ∼ 0.6 for all con- sidered values of γ at λ = 0.009, β = 12. It follows from our data that φ >> δφ at γ << γ while φ ∼ δφ at γ > γ. m c2 m c2 Basing on these observations we expect that in the region γ << γ the c2 usual perturbation expansion aroundtrivial vacuum of spontaneously broken theory can be applied to the lattice Weinberg - Salam model while in the FR γ < γ < γ it cannot be applied. c c2 At the value of γ equal to γ ∼ 0.278 the calculated value of the cutoff c2 is 1.0±0.1 Tev. 6 The renormalized coupling In order to calculate the renormalized fine structure constant α = e2/4π R (where e is the electric charge) we use the potential for infinitely heavy exter- nalfermions. Weconsider Wilsonloopsfortheright-handedexternal leptons: WR (l) = hReΠ e2iθxyi. Here l denotes a closed contour on the lattice. lept (xy)∈l We consider the following quantity constructed from the rectangular Wilson loop of size r×t: V(r) = lim log W(r×t) . At large enough distances we t→∞ W(r×(t+1)) 10