A common origin of two accelerating universes: inflation and dark energy Kunio Kaneta, Kin-ya Oda, Motohiko Yoshimura Abstract We develop a quantum theory of inflaton and its decay product of various gauge boson pairs to investigate the preheating towards thermalized universe. The inflaton decay into gauge-boson pairs is shown to be inevitably accompanied by tachyon-mass-like correction to inflation potential that ultimately leads to an inflaton escape out of trapped local potential minimum towards the field infinity. This gives rise to a conversion mechanism of early inflationary acceleration to a quintessence dark energy acceleration at late stages of cosmic evolution. The success of the escape depends on how standard particles are incorporated into a scheme of extended Jordan-Brans-Dicke gravity. New types of super-radiance mechanism that enhance the ending of preheating are identified and compared with the Dicke model. I Introduction The prime objective of cosmology is to clarify how hot big bang universe started and how the initial universe developed to the present state we observe now. Issues related to this objective include inflation, creation of matter and radiation, baryon asymmetry, nucleo-synthesis, dark matter, and dark energy. We report in the present paper our recent work related to some of these problems. The inflationary scenario solved the outstanding problems of flatness and horizon that face the early universe by introducing a scalar field with its approximate cosmological constant behavior driving the exponential expansion inflation , inflation 2 , cosmology textbooks . Inflation however leaves an empty space devoid of radiation and matter. It is believed that field oscillation at the end of inflation leads to copious particle production. The popular mechanism of particle production is based on the parametric amplification such as described by the Mathieu (a special class of Floquet-type) equation of periodic coefficient due to field oscillation. Despite of many interesting achievements along this direction dk – preheating review , the outcome still does not ensure that hot big bang universe emerges necessarily. This is one of the most important problems we face in early cosmology. Another important problem in cosmology is relation of inflation at an early epoch and dark energy at late epochs. We explore possibilities of these two accelerating universes arising from the same origin. In koy 24-1 we presented one of these possibilities using two-component scalar fields. In the present work we give a more attractive possibility based on a single inflaton. It is inevitable that inflation and the late-time acceleration is caused by a scalar field or scalar fields. It is desirable to introduce the scalar inflaton from some definite theoretical principle. We pay special attention to a class of extended Jordan-Brans-Dicke (eJBD) jbd scalar-tensor gravity. It is found that our extension koy 23 of the original JBD and massive JBD theory ejbd brings about new changes to the original JBD gravity, and the problem of how standard or grand unified particle physics is incorporated becomes crucial in these scalar-tensor gravity. We have further developed, with regard to the preheating stage of particle production, a new idea koy 24-2 of super-radiant particle production of inflaton decay. Super-radiant decay arises in a phase coherent region, speeding up the transfer of inflaton energy to those of produced particles. In the present work we provide a quantum theory treating both inflaton and produced particle on an equal footing, thereby unifying the two stages of parametrically amplified particle production and super-radiant ending of preheating. This approach bypasses a complicated mode summation method often necessary in parametric amplification schemes of the Mathieu system developed in kls , my 95 and fkyy . The most important result of the present quantum formulation is that a smooth inflaton exit out of a trapped local potential minimum to the field infinity is realized, and a quintessence type quintessence , cosmology textbooks of accelerating universe emerges. This is due to our new discovery of tachyon instability associated with particle production, which makes the exit out of the trapped region successful for a single inflaton field. We thus solve the problem of relating two accelerating universes. We avoid in this work to use the terminology of reheating often used in the literature, since it is neither clear nor necessary that there has been a stage of heated universe prior to inflation. Throughout this paper we use the unit of \hbar=c=k_{B}=1 except that we recover \hbar in a part when we discuss the WKB approximation. Contents of the rest are organized as follows. II Extended Jordan-Brans-Dicke theory: How particle physics is incorporated A.   Coupling to ordinary matter in lagrangian B.   Trace anomaly and inflaton decay to gauge boson pairs C.  Comparison with other popular schemes of matter coupling III Quantum theory of preheating A.   Quantum formalism B.   Perturbation theory to the second orders C.   Tachyon instability associated with particle-pair production D.   Results of higher order perturbation IV Super-radiant stage of preheating A.   Dicke model and its identification to our quantum theory B.  Macro-coherent amplification V Transition from inflationary to dark energy stages of acceleration A.   Inflaton escape B. Epochs of thermalization and potential restoration VI Summary and outlook VII Appendix: Relevance of trace anomaly II Extended Jordan-Brans-Dicke theory: How particle physics is incorporated II.1 Coupling to ordinary matter in lagrangian The lagrangian density of our extended Jordan-Brans-Dicke (eJBD) theory is derived by a Weyl transformation from the Jordan metric frame to the Einstein metric frame. The cosmological constant problem present in the Jordan frame is absent in the Einstein frame if eJBD field decreases exponentially towards the field infinity. Besides the eJBD gravity part, it is important to specify how our eJBD field denoted by \chi couples to standard particle fields (gauge fields, quarks, leptons, and Higgs bosons). Our scheme uses the lagrangian density in the Einstein frame, \displaystyle{\cal L}=-\frac{M_{\rm P}^{2}}{2}R+\frac{1}{2}(\partial\chi)^{2}-V_{\chi}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu} \displaystyle\hskip-14.22636pt-\bar{\psi}\left(\gamma(i\partial-e^{-\gamma_{g}\chi/M_{\rm P}}gA)-e^{-\gamma_{y}\chi/M_{\rm P}}y_{\psi}\,H\right)\psi+\cdots\,. (1) Conformal factors e^{-\gamma_{i}\chi/M_{\rm P}} appear in the terms for which neither gauge invariance nor conformal symmetry is at work. Two dimensionless parameters \gamma_{g},\gamma_{y} in this formula characterize the strength of inflaton coupling to matter, with M_{\rm P}=1/\sqrt{8\pi G}\sim 2.4\times 10^{18}\, GeV. We shall compare later this matter coupling to more popular ones akin to the original JBD gravity. The \chi inflaton potential V_{\chi} is taken as \displaystyle V_{\chi}=V_{0}\,(\frac{\chi}{M_{\rm P}})^{2}\,e^{-\gamma_{\chi}\chi/M_{\rm P}}\,, (2) \displaystyle V_{0}>0\,,\hskip 8.5359pt\gamma_{\chi}>0\,. This potential has two extrema of \chi=(0,2/\gamma_{\chi})M_{\rm P} : the minimum is at \chi=0 , with its potential value 0 . The potential around the local minimum is approximated by a harmonic oscillator (HO); \displaystyle V_{\rm HO}=\frac{V_{0}}{M_{\rm P}^{2}}\,\chi^{2}\,. (3) The inflaton mass m_{\chi} at the local minimum is \sqrt{2V_{0}/M_{\rm P}^{2}} , estimated to be around 2\times 10^{12}\, GeV from our previous studies koy 24-1 . The potential maximum is at \chi=2M_{\rm P}/\gamma_{\chi} , with its maximum value, \displaystyle V_{\rm max}=\frac{4e^{-2}\,V_{0}}{\gamma_{\chi}^{2}}=0.541\frac{V_{0}}{\gamma_{\chi}^{2}}\,. (4) The potential shape V_{\chi}/V_{0} is illustrated in Fig( 1 ) which describes harmonic motion around the field origin for inflation, and the quintessential motion towards the field infinity \chi\rightarrow+\infty . It was shown in koy 24-1 that a variant of similar \chi potential leads to a successful inflation consistent with the observed spectral index and limit of tensor to scalar ratio. A further simplified model here also does the successful job. What is important in the potential choice is existence of harmonic oscillator part surrounded by a barrier decreasing towards the field infinity, and detailed potential form is irrelevant to simultaneously realize inflation and the dark energy. Still, it is amusing to note that the potential we propose here is a simple product of popular chaotic and quintessence potentials, inflation , cosmology textbooks , quintessence . Cosmological constant present in the Jordan metric frame becomes exponentially suppressed in the Einstein frame, thus easing the cosmological constant problem. Figure 1: Time evolution of global potential shapes: inflaton oscillating epoch at the left, de-stabilized epoch in the middle, and late-time epoch after the potential recovery. Figure 2: Late-time evolution since nucleo-synthesis till the present time. Assumed parameters are \Omega_{\chi}(a_{0})=0.687,w_{\chi}(a_{0})=-0.998 . II.2 Trace anomaly and inflaton decay to gauge boson pairs Our eJBD field evolves according to the wave equation under the flat Friedman-Robertson-Walker metric, \displaystyle\partial^{2}\chi+\partial_{\chi}V_{\chi}=\bar{T}\,, (5) \displaystyle\bar{T}=g_{\mu\nu}\,2\frac{\delta(\sqrt{g}{\cal L})}{\delta g_{\mu\nu}}=\bar{T}_{m}+\left({\rm graviton\;and\;}\chi\;{\rm parts}\right)\,, (6) with \bar{T} the trace of energy-momentum tensor including the small gravitational coupling of order 1/M_{\rm P} in the expansion, as verified by the expansion, e^{-\gamma_{g}\chi/M_{\rm P}}\sim 1-\gamma_{g}\chi/M_{\rm P} . The first term \bar{T}_{m} in ( 6 ) in the right hand side is the contribution from standard matter fields. Right after inflation and when inflaton starts oscillation, the gauge symmetry is spontaneously broken since there is no temperature effect restoring the symmetry, but time dependent inflaton mass is of the order, M_{\rm P} , hence all standard particle masses may be regarded as nearly massless. This makes \bar{T}_{m} nearly vanishing. There is however an important quantum correction of the trace anomaly ps . The trace-anomaly related contribution gives an effective decay amplitude of \chi\rightarrow A_{1}A_{2} with A_{i}\,,i=1,2 two massless gauge fields, as fully explained in Appendix below. The amplitude is \displaystyle\gamma_{J}\,\frac{\chi}{M_{\rm P}}\sum_{A}\,\frac{b_{A}\alpha_{A}^{2}}{8\pi}\,(F^{A\,(1)})^{\rho\sigma}\,F^{A\,(2)}_{\rho\sigma}\,. (7) \alpha_{A}=g_{A}^{2}/4\pi are various gauge couplings, and b_{A} is the coefficient of renormalization group. There is essentially no observational constraint on the new conformal factor \gamma_{J} here, and this decay gives a promising candidate for the instability we work out. In grand unified theories there are other important contributions whose implications shall be discussed elsewhere. In analogy to QED (Quantum Electro Dynamics) the squared field strength may be expressed in terms of electric and magnetic type gauge fields of A types; \displaystyle\frac{1}{2}(F^{A\,(1)}_{\rho\sigma})(F^{A\,(2)})^{\rho\sigma}=\left(\vec{E}^{A\,(1)}\cdot\vec{E}^{A\,(2)}-\vec{B}^{A\,(1)}\cdot\vec{B}^{A\,(2)}\right) (8) \displaystyle+({\rm cubic\;terms\;for\;non-Abelian\;gauge\;fields})\,. Estimate in SO(10) grand unification gives \sum_{A}\frac{b_{A}\alpha_{A}}{16\pi}\sim-0.014 koy 24-2 . We shall often omit gauge field species i and its summation \sum_{A} for notational simplicity. Inflaton decay rate \Gamma_{B} in perturbation theory is readily calculable, assuming no time variation of various constants, to give \displaystyle\Gamma_{B}=\gamma_{J}^{2}\,(\sum_{A}\frac{b_{A}\alpha_{A}}{8\pi})^{2}\,\frac{m_{\chi}^{3}}{16\pi^{2}M_{\rm P}^{2}}\,. (9) We shall show later that this rate is dramatically enhanced leading to almost instantaneous decay at the super-radiant phase of preheating. Since inflatons are at rest in all stages of cosmic evolution, this decay is characterized by back-to-back pair production of momentum configuration, \vec{k}\,,-\vec{k} with its magnitude fixed by |\vec{k}|\equiv k=m_{\chi}/2 . This is important to create macro-coherence at super-radiance stage in which the phase coherence region is not limited by the wavelength of light unlike in the original Dicke model dicke – sr textbook . This is an enormous advantage to realize a rapid termination of preheating, since the use of decay rate of the Born approximation leads to slow decay accompanied by the difficulty of thermalization. II.3 Comparison with other popular schemes of matter coupling It would be instructive to compare our eJBD matter-coupling to other more popular ones based on the original JBD idea jbd . A popular scheme introduced in def uses the matter coupling in the Einstein metric frame; taking \alpha_{0},\beta_{0} as constants, \displaystyle{\cal L}_{m}={\cal L}_{st}(\psi_{m};A^{2}g_{\mu\nu})\,, (10) \displaystyle\hskip-14.22636ptA^{2}=A^{2}(\chi)=\exp[-\alpha_{0}\frac{\chi}{\sqrt{8\pi}\,M_{\rm P}}+\beta_{0}(\frac{\chi}{\sqrt{8\pi}\,M_{\rm P}})^{2}]\,, (11) with \psi_{m} representing standard fields generically. (We have slightly changed the reference point of \chi in this formula from def , whose dimensionless \chi is not well defined, but we believe the formula above agreeing with their results.) The original JBD uses the linear term alone with \beta_{0}=0 . Despite of apparent difference of this parametrization from ours, these are closely related. It is sufficient to inspect the contribution of anomaly term in the field equation ( 6 ). A^{2} parameterization gives, in the eJBD massless limit, \displaystyle\partial^{2}\chi=-\frac{\chi}{\sqrt{8\pi}\,M_{\rm P}}\,\alpha_{0}\bar{T}_{\rm anomaly}+O\left((\frac{\chi}{M_{\rm P}})^{2}\right) (12) while ours gives to the right hand side \displaystyle-\frac{\chi}{M_{\rm P}}\,\gamma_{J}\sum_{A}\bar{T}_{\rm anomaly}+O\left((\frac{\chi}{M_{\rm P}})^{2}\right)\,. (13) Thus, to this order, \displaystyle\alpha_{0}=\sqrt{8\pi}\,\gamma_{J}\,,\hskip 8.5359pt\beta_{0}=0\,, (14) for a single gauge field. Correspondence between two coupling schemes becomes more complicated if one includes contributions from other sources of quark mass term and Higgs contribution. But one thing is clear: the popular scheme emphasizes the contribution from \propto\beta_{0} term in the exponent of A^{2}(\chi) , while ours emphasizes multiple contributions of \alpha_{0} types, although one can readily incorporate in our formalism \beta_{0} type exponents. Our coupling scheme ( 13 ) is an extension of the original JBD gravity jbd which was invented at a time when only QED is firmly established. Our present scheme fully respects the gauge invariance of established particle physics in the present days, but abandon the uniform and simplest form of Weyl conformal transformation. as stressed in koy 24-1 . The standard model of particle physics is based on the spontaneously broken gauge symmetry, and the theory is SU(3) \times SU(2) \times U(1) gauge invariant at the lagrangian level. There are five pieces of gauge invariant lagrangian densities in the action principle; squared gauge field-strength, fermion kinetic term, fermion Yukawa coupling, Higgs kinetic term, and Higgs potential. Our principle is that these five pieces are independently Weyl rescaled. By redefining field operators according to the canonical formalism, one can have a plethora of conformal A^{2} factors. One of the simplest forms is the one given here. This form emerges when higher loop corrections are incorporated in super-string theory damour-polyakov . Since it is difficult to convincingly derive the low energy limit of the super-string theory, we shall follow a bottom-up approach using our parametrization. III Quantum theory of preheating III.1 Quantum formalism Events after inflation occur independently at different space points within the Hubble horizon, with its size given by H^{-1}=a/\dot{a} , a(t) being the cosmic scale factor. Hence we may introduce independent inflaton fields, b_{s}(t),b_{s}^{\dagger}(t) , at different sites s . All inflatons are assumed at rest with zero momentum. We thus introduce their hamiltonian, \displaystyle\chi(t)=\frac{1}{\sqrt{2m_{\chi}V}}\,\sum_{s}\,(b_{s}(t)\,e^{-im_{\chi}t}+b_{s}^{\dagger}(t)\,e^{im_{\chi}t})\,, (15) \displaystyle[b_{s}(t)\,,b_{s^{\prime}}^{\dagger}(t)]=\delta_{s,s^{\prime}}\,,\hskip 8.5359pt[b_{s}(t)\,,b_{s^{\prime}}(t)]=0\,, (16) \displaystyle H_{0}(\chi)=\int d^{3}x_{s}\,\frac{1}{2}(\,\pi_{\chi}^{2}+m_{\chi}^{2}\chi^{2}\,)=m_{\chi}\sum_{s}\left(b_{s}^{\dagger}b_{s}+\frac{1}{2}\right) \displaystyle+({\rm oscillating\;terms})\,. (17) Site variable s in operators b_{s}(t)\,,b_{s}^{\dagger}(t) are merely labels for distinction, and there is no actual coordinate \vec{x}_{s} dependence due to the space translational invariance of inflaton time evolution, at the initial time of time evolution. We may write interaction hamiltonian that holds in the interaction picture; \displaystyle H_{\rm int}=-c_{A}\int d^{3}x_{s}\,\chi(t)(\vec{E}^{2}-\vec{B}^{2}) (18) \displaystyle=-c_{A}\int\frac{Vd^{3}k}{(2\pi)^{3}}\,\sum_{s}\,\frac{2\omega_{k}}{\sqrt{m_{\chi}V}} \displaystyle\sum_{\rm pol}\,\left(\vec{e}(k)\cdot\vec{e}(-k)\,b_{s}^{\dagger}(t)\,a_{k}a_{-k}\,e^{i(m_{\chi}-2\omega_{k})t}\right. \displaystyle+\vec{e}(k)^{*}\cdot\vec{e}(-k)^{*}\,b_{s}(t)\,a_{k}^{\dagger}a_{-k}^{\dagger}\,e^{i(-m_{\chi}+2\omega_{k})t} (19) \displaystyle+\vec{e}(k)\cdot\vec{e}(-k)b_{s}(t)\,a_{k}a_{-k}\,e^{-i(m_{\chi}+2\omega_{k})t} \displaystyle\left.+\vec{e}(k)^{*}\cdot\vec{e}(-k)^{*}b_{s}^{\dagger}(t)\,a_{k}^{\dagger}a_{-k}^{\dagger}\,e^{i(m_{\chi}+2\omega_{k})t}\right)\,, (20) \displaystyle c_{A}=\frac{\gamma_{J}\,b_{A}\alpha_{A}^{2}}{8\pi M_{\rm P}}\,. (21) Operators a_{k},a_{k}^{\dagger} are annihilation and creation operators of various gauge vector potentials using the radiation gauge. The first two terms in ( 19 ) give contributions that survive on the mass shell at m_{\chi}=2\omega_{k} , while the second two contributions in ( 21 ) are those from off the mass shell. Operators, a_{\pm k},a_{\pm k}^{\dagger},b_{s} , are dimensionless. We shall often consider, for simplicity, the maximum polarization sum \sum_{\rm pol}\vec{e}(k)\cdot\vec{e}(-k)=2 (happens to coincide with a squeezed state), which occurs when the back-to-back gauge field pair has either parallel or anti-parallel linear polarizations. So far we discussed gauge boson pair coupling to eJBD \chi field. If the equally possible Higgs boson pair is adopted, the coupling is replaced by c_{A}=y_{\psi}^{2}\gamma_{y}/(64\pi^{2}M_{\rm P}) with weaker Yukawa coupling y_{\psi}^{2}/4\pi than \alpha . We shall not purse this possibility any further in the present work. We separate the on-shell contribution and denote it by the prime ′ , \displaystyle H_{\rm int}^{\prime}(t)=-4\sqrt{2}c_{A}\,\int\frac{Vd^{3}k}{(2\pi)^{3}}\,\frac{\omega_{k}}{\sqrt{2m_{\chi}V}}\,\sum_{s}\, \displaystyle\left(b_{s}^{\dagger}\,a_{k}a_{-k}\,e^{i(m_{\chi}-2\omega_{k})t}+b_{s}\,a_{k}^{\dagger}a_{-k}^{\dagger}\,e^{i(-m_{\chi}+2\omega_{k})t}\right)\,. (22) The volume dependence shown here is natural, since the positive and negative frequency components of quantum inflaton field are \displaystyle\chi^{(+)}(t)=\frac{1}{\sqrt{2m_{\chi}V}}\sum_{s}b_{s}(t)\,e^{-im_{\chi}t}\,, (23) \displaystyle\chi^{(-)}(t)=\frac{1}{\sqrt{2m_{\chi}V}}\sum_{s}b_{s}^{\dagger}(t)\,e^{im_{\chi}t}\,, (24) with the conventional volume dependent \sqrt{V} factor included. This on-shell contribution H_{\rm int}^{\prime}(t) is akin to the short time average previously considered in koy 24-2 . The volume V in cosmology may effectively be taken as the Hubble volume H^{-3} . The time dependence of inflaton in the harmonic potential simplifies to \chi^{(\pm)}(t)=\chi^{(\pm)}(0)\,e^{\mp im_{\chi}t} , while it has extra time dependence under more general \chi potentials, V_{\chi} . As initial states we take states without radiation pair. It is necessary to distinguish states of quantum \chi and radiation fields separately; at a single site we may take excited states as well, hence use the notation |n_{s}\rangle_{\chi}\,,n_{s}=0,1,2,\cdots . The general class of initial states at {\cal N}_{s} sites is a direct product; \displaystyle|\Psi[\{n_{s}\},0]\rangle=\Pi_{s}\,(|n_{s}\rangle_{\chi}\,|0\rangle_{\rm rad}\,)\,. (25) Radiation quanta are dispersed in space, and there is no need to attach them to particular sites. Spatially homogeneous states are however adequate for consideration right after the end of inflation of the kinds, hence we only consider the type of initial states, \displaystyle|\Psi(n;0)\rangle=\Pi_{s}\,(|n\rangle_{\chi,s}|0\rangle_{\rm rad}\,)\,. (26) Dicke-type initial state we later consider is |\Psi(1;0)\rangle . We also need to consider excited states such as |\Psi[{0_{s},2_{s^{\prime}\,}},2\rangle , meaning that \chi is in the ground state only at site s , and is in the second excited states at all other sites. Time evolved states are not spatially homogeneous and the distribution of \chi state at different site is automatically described in our formalism. States evolve by a unitary operator, and for instance the Dicke-type state in the interaction picture follows according to \displaystyle|\Psi(1;0)\rangle_{i}\rightarrow|\Psi(t)\rangle=U(t,t_{i})|\Psi(1;0)\rangle_{i}\,, (27) \displaystyle U(t,t_{i})={\cal T}\,\exp[-i\int_{t_{i}}^{t}dt^{\prime}\,H^{\prime}_{\rm int}(t^{\prime}-t_{i})]]\,, (28) with {\cal T} the time ordering. This is not a purely quantum evolution, since we dropped the off-shell parts of interaction hamiltonian. We may call this approximation semi-quantum theory. It is convenient to introduce a conjugate pair of combined field operators, \displaystyle\hskip-5.69046ptQ^{(+)}(t)=\chi^{(+)}(t)a_{k}^{\dagger}a_{-k}^{\dagger}\,,\hskip 8.5359ptQ^{(-)}(t)=\chi^{(-)}(t)a_{k}a_{-k}\,. Using conjugate dynamical variables Q^{(\pm)}(t) , the interaction hamiltonian becomes \displaystyle H_{\rm int}^{\prime}(t)=-4\sqrt{2}c_{A}\int\frac{Vd^{3}k}{(2\pi)^{3}}\,\omega_{k}\,e^{2i\omega_{k}t}Q^{(+)}(t)+({\rm h.c.})\,, (30) The relevant time integration is \displaystyle-i\int_{t_{i}}^{t}dt^{\prime}\,H^{\prime}_{\rm int}(t^{\prime}-t_{i}) \displaystyle=-4\sqrt{2}c_{A}\int\frac{Vd^{3}k}{(2\pi)^{3}}\,\frac{\omega_{k}}{2\omega_{k}-m_{\chi}}\times \displaystyle\left[\left(\cos(m_{\chi}-2\omega_{k})(t-t_{i})-1\right)\right.\left(Q_{\rm HO}^{(+)}(0)-Q_{\rm HO}^{(-)}(0)\right) \displaystyle\hskip-8.5359pt\left.+i\sin(m_{\chi}-2\omega_{k})(t-t_{i})\left(Q_{\rm HO}^{(+)}(0)+Q_{\rm HO}^{(-)}(0)\right)\right]\,. (31) for the pure HO case. For convenience we introduce the phase space factor relevant in cosmology, \displaystyle\Omega_{k}\equiv\int\frac{Vd^{3}k}{(2\pi)^{3}}=\frac{H^{-3}}{8\pi^{2}}m_{\chi}^{2}|\Delta\omega_{k}|\,, (32) \displaystyle\hskip 28.45274pt\Delta\omega_{k}=\omega_{k}-\frac{m_{\chi}}{2}\,, (33) with 1/H the Hubble length. The quantity \Delta\omega_{k} is taken as a fraction of m_{\chi}/2 . The real part of operator coefficients in time-integrated hamiltonian behaves like \propto 2\Delta\omega_{k}\,(t-t_{i})^{2} , as t-t_{i}\rightarrow 0 , while its imaginary part behaves like \propto i\omega_{k}(t-t_{i}) . These behaviors are typical of quantum fluctuation and dissipation. III.2 Perturbation theory to the second order Let us systematically work out more details of gauge boson pair production amplitude and expectation values in initial states. To the first order perturbation, there is only type of decay amplitude, \displaystyle-i\,\langle\Psi(0;1)|\int_{0}^{t}dt^{\prime}H_{\rm int}^{\prime}(t^{\prime})|\Psi(1;0)\rangle\,, (34) whose square gives decay rate after a phase space integration. To the second order of interaction, there are two types of decay amplitudes, the difference due to the involved number of sites, one or two. They are \displaystyle(1)\;{\cal A}_{2}= \displaystyle\hskip-14.22636pt-\frac{1}{2}\langle\Psi[{0_{s1},0_{s2},2_{s^{\prime\prime}}},2]|{\cal T}\,\left(\int_{0}^{t}dt^{\prime}H_{\rm int}^{\prime}(t^{\prime})\right)^{2}|\Psi(1;0)\rangle\,, (35) \displaystyle(2)\;{\cal A}_{2}^{\prime}= \displaystyle-\frac{1}{2}\langle\Psi[{0_{s},2_{s^{\prime}}},2]|{\cal T}\,\left(\int_{0}^{t}dt^{\prime}H_{\rm int}^{\prime}(t^{\prime})\right)^{2}|\Psi(2;0)\rangle\,. (36) Site occupancy is different in the two cases: there are two sites in the ground state in (1), while there is only one in (2). This is due to that the initial states are different. Decay amplitudes and decay rates are different in the two cases, but what might be confusing is that to the leading (\Delta\omega_{k}\,t)^{2} expansion results are the same; \displaystyle\hskip-14.22636pt(1)\;{\cal A}_{2}=\frac{\tilde{\chi}^{2}}{2}\,(1-e^{2i\Delta\omega_{k}\,t})^{2}=-2\tilde{\chi}^{2}(\Delta\omega_{k}\,t)^{2}+\cdots\,, (37) \displaystyle\hskip-14.22636pt\tilde{\chi}=2\sqrt{2}c_{A}\chi^{(+)}\,\frac{m_{\chi}\Omega_{k}}{2\Delta\omega_{k}}=\frac{\sqrt{2}}{16\pi^{2}}(\frac{3}{2})^{3}c_{A}\chi^{(+)}\,(m_{\chi}t)^{3}\,, (38) \displaystyle\hskip-17.07182pt(2)\;{\cal A}_{2}^{\prime}=-\frac{\tilde{\chi}^{2}}{2}\,|1-e^{2i\Delta\omega_{k}\,t})|^{2}=-2\tilde{\chi}^{2}(\Delta\omega_{k}\,t)^{2}+\cdots\,. (39) The difference of {\cal A}_{2} and {\cal A}_{2}^{\prime} appears in the neglected dotted terms. The decay amplitude {\cal A}_{2} in case (1) corresponds to the mode-summed probability amplitude in the Mathieu chart in past approaches kls , my 95 and fkyy , and exponentially grows in time. The decay amplitude is however dramatically changed when higher order effects are included, and grows much faster beyond critical time, as is made clear below. It becomes important to clarify time dependence of field and mass parameter m_{\chi} that appear in the basic coupling \tilde{\chi} in ( 38 ). Time variation of field \chi^{(\pm)}=\chi/2 follows the law in the matter-dominated universe, \rho_{\chi}(t)\propto a^{-3}\propto t^{-2} , from which one derives the behavior of product, m_{\chi}\chi\propto 1/t . This relation is collaborated with HO motion governed by the equation including the Hubble friction, \displaystyle\ddot{\chi}+3H\dot{\chi}+m_{\chi}^{2}\chi=0\,,\hskip 8.5359ptH=\frac{2}{3t}\,. (40) The differential equation is solved in terms of Bessel type functions, and its long time limit is given by \displaystyle\rho_{\chi}=m_{\chi}^{2}\chi^{2}\rightarrow(m_{\chi}\chi_{0})^{2}(\frac{t_{0}}{t})^{2}\propto a^{-3}(t)\,. (41) Assuming a constant m_{\chi} , we arrive at \chi(t)\propto 1/t . t_{0} is the onset time of inflaton oscillation. The basic coupling is then expressed by \displaystyle\tilde{\chi}=\frac{27\sqrt{2}}{(16\pi)^{2}}c_{A}\chi(m_{\chi}t)^{3}\frac{\chi}{M_{\rm P}} \displaystyle=\frac{3\sqrt{6}}{32\pi^{2}}c_{A}M_{\rm P}\,(\frac{z_{0}}{z})^{9/2}\frac{\chi}{M_{\rm P}}\,, (42) with 3\sqrt{6}/(32\pi^{2})=0.0233 . z_{0} here is the redshift factor at t=t_{0} . We have used as a time variable the redshift factor z . Time variation in shorter time scales are given by oscillating functions in terms of \Delta\omega_{k}t=m_{\chi}t\times(\Delta\omega_{k}/m_{\chi}) where we take \Delta\omega_{k}/m_{\chi}=O(0.1\sim 0.01) . We should have expected uncontrollable divergence caused by the interaction hamiltonian \chi F_{\mu\nu}F^{\mu\nu} which is non-renormalizable operator, but the gauge invariance makes the self-energy controllable. III.3 Tachyon instability associated with particle-pair production What is important equally to the decay amplitude is the expectation value in the initial state. This describes quantum fluctuation, and its interpretation gives \chi potential correction related to particle production. The relation of the two constitutes an important part of fluctuation-dissipation theorem. The expectation value in the Dicke initial state is given by \displaystyle-\frac{1}{2}\,\langle\Psi(1;0)|{\cal T}\left(\int_{0}^{t}dt^{\prime}H_{\rm int}^{\prime}(t^{\prime})\right)^{2}|\Psi(1;0)\rangle={\cal E}_{2}\,, (43) \displaystyle\Re{\cal E}_{2}=-\frac{\tilde{\chi}^{2}}{2}\left(1-2\cos(2\Delta\omega_{k}\,t)+\cos(4\Delta\omega_{k}\,t)\right)\,, (44) \displaystyle\Im{\cal E}_{2}=-\frac{\tilde{\chi}^{2}}{2}\left(-2\sin(2\Delta\omega_{k}\,t)+\sin(4\Delta\omega_{k}\,t)\right)\,, (45) to the second order of interaction. We plot the real and the imaginary parts in Fig( 3 ). The positive imaginary part is a kind of anti-dissipation, and may appear in the parametric amplification. The negative real part that starts at 2\Delta\omega_{k}t=3\pi/2 is of particular interest: it is correlated with decreasing positive imaginary part. The positive imaginary part indicates the growing pair production. In Section V C we shall show that the escape probability of eJBD field out of trapped potential region is greatly enhanced. Figure 3: Ideal form of potential correction plotted against 2\Delta\omega_{k}\,t : the real part given by ( 44 ) in solid black and the imaginary part given by ( 45 ) in dashed orange. For simplicity \tilde{\chi}^{2}/2=1 is assumed. When the imaginary part in orange is positive and the real part in black is negative, the strong parametric resonance effect is at work in instability bands. The quantum fluctuation given by the real part is \chi field dependent; \propto\chi^{2} . It is thus natural to interpret this as a \chi mass correction, hence we equate the t\rightarrow 0 limit of real part to \displaystyle\int d^{4}x\,\delta{\cal V}_{\chi}=\frac{\delta m_{\chi}^{2}}{2}\chi^{2}t\,(H^{-1})^{3}\,. (46) This interpretation leads to an identification of mass correction, \displaystyle\delta m_{\chi}^{2}=-c\,(c_{A}M_{\rm P})^{2}\,M_{\rm P}^{2}(\frac{z_{0}}{z})^{3}\,, (47) \displaystyle c=\frac{160}{3}(\frac{2}{3})^{3}\frac{(27\sqrt{2})^{2}}{(16\pi)^{4}}=3.61\times 10^{-3}\,. (48) We have proved that a tachyon mass with \delta m_{\chi}^{2}<0 emerges to the lowest second order of interaction. The result of negative squared mass correction implies that the tachyon instability near the field origin occurs. Let us study whether this instability occurs with a reasonable time range of t_{0} to the termalization time t_{R} , t_{0}\sim t_{R}=O(4\times 10^{3})t_{0} . The tachyon instability occurs at \delta m_{\chi}^{2}+m_{\chi}^{2}=0 with m_{\chi} the HO mass. This condition is equivalent to \displaystyle-c\,(c_{A}M_{\rm P})^{2}\,\frac{M_{\rm P}^{2}}{m_{\chi}^{2}}(\frac{z_{0}}{z})^{3}+1=0\,, (49) One may thus derive the red shift factor of instability time t_{\rm tac} , \displaystyle z_{\rm tac}=c^{1/3}(c_{A}M_{\rm P})^{2/3}(\frac{M_{\rm P}}{m_{\chi}})^{2/3}z_{0}\,, (50) with c^{1/3}=0.153 . For a choice of \gamma_{J}=0.01\sim 0.1 in the formula c_{A}M_{\rm P}=\gamma_{J}\alpha/16\pi , z_{\rm tac}/z_{0} falls in its acceptable range of 2\times 10^{-5}\sim 0.9\times 10^{-4} , assuming m_{\chi}=M_{\rm P} (smaller values than this are favored). The tachyon instability indicated by the negative quantity in ( 47 ) is one of our major findings in the present quantum theory of preheating, and opens an interesting possibility towards the transition from the early inflationary to the late dark energy acceleration in eJBD gravity of a single field. One can determine |\Delta\omega_{k}| (but not its sign) by making an alternative calculation in the second order perturbation; the Feynman diagram technique. The diagram in the second order contains a kind of polarization scalar \Pi_{s}(k_{1},k_{2}) made of one-loop gauge boson pair propagation; \displaystyle\Pi_{s}(k_{1},k_{2})=\int d^{4}x\,e^{i(k_{1}+k_{2})\cdot x} \displaystyle\hskip-8.5359pt\langle 0|T\left[\left(F^{\rho\sigma}(x)F_{\rho\sigma}(x)\right)\left(F^{\mu\nu}(0)F_{\mu\nu}(0)\right)\right]|0\rangle\,, (51) with the field strength F_{\mu\nu}(x)=\partial_{\mu}A_{\nu}-\partial_{\nu}A_{\mu} related to the vector potential A_{\mu}(x) . Using the Feynman gauge propagator of vector potential, we derive \displaystyle\Pi_{s}(k_{1},k_{2})=6\left((k_{1}+k_{2})^{2}\right)^{2}\,, (52) \displaystyle k_{1}=(\omega_{k},\vec{k})\,,\hskip 8.5359ptk_{2}=(\omega_{k},-\vec{k})\,. (53) Thus, \Pi_{s}(k_{1},k_{2})=6m_{\chi}^{4} , since \omega_{k}=m_{\chi}/2 due to the energy conservation. This simplification brings about a correspondence to the previous formula, \displaystyle\frac{6m_{\chi}^{4}}{q_{0}^{2}-\vec{q}^{2}-m_{\chi}^{2}+i0^{+}}\hskip 8.5359pt(q_{0}=2\omega_{k}\,,\vec{q}=0) (54) \displaystyle\Leftrightarrow\frac{\Omega_{k}2m_{\chi}}{V}\frac{1}{2\omega_{k}(2\omega_{k}-m_{\chi}+i0^{+})}\,F\,, (55) \displaystyle F=F(\Delta\omega_{k}\,t)=\left(1-e^{-2i\Delta\omega_{k}\,t}-2i\Delta\omega_{k}\,t\right)\,. (56) The absolute value of the function F here behaves as (\Delta\omega_{k}t)^{3/2} for large times, and we conclude from this \displaystyle|\Delta\omega_{k}|=c\,M_{\rm P}(\frac{m_{\chi}}{M_{\rm P}})^{2/5}(\frac{z}{z_{0}})^{9/10}\,, (57) \displaystyle c=(24\pi^{2})^{2/5}(\frac{\sqrt{3}}{2})^{3/5}\sim 8.17 (58) \displaystyle\Omega_{k}\sim 0.493\,(\frac{m_{\chi}}{M_{\rm P}})^{12/5}(\frac{z_{0}}{z})^{18/5}\,. (59) The value |\Delta\omega_{k}|/m_{\chi}=1\sim 0.1 corresponds to z/z_{0}=(0.097\sim 0.0075)\,(m_{\chi}/M_{\rm P})^{2/3} . In numerical computations that follow we shall take \Delta m_{k}=0.1\,m_{\chi} as a standard reference. III.4 Results of higher order perturbation It is both necessary in later discussions and also illuminating for a deeper understanding of what occurs in our system to work out higher orders beyond the second. The n- th order perturbation theory gives the expectation value in the Dicke type initial states, \displaystyle\hskip-14.22636pt{\cal E}_{n}=\frac{(-i)^{n}}{n!}\,\langle\Psi(1;0)|{\cal T}\left(\int_{0}^{t}dt^{\prime}\,H^{\prime}_{\rm int}(t^{\prime})\right)^{n}|\Psi(1;0)\rangle\,. (60) For this quantity to be non-vanishing, n must be even positive integer, since the final state \langle\,{\rm final}\,| goes back to the initial state \langle\,{\rm initial}\,| . To the second order, this happens to be identical to the decay amplitude, thus {\cal E}_{2}={\cal A}_{2} . The relation of expectation value to the potential correction is \displaystyle\langle\Psi(1;0)|U(t,0)|\Psi(1;0)\rangle=\exp[-itH^{-3}\langle\delta V_{\chi}\rangle]\,. (61) Decay and excitation at the same site such as |1_{s}\rangle_{\chi}\rightarrow|0_{s}\rangle_{\chi}\rightarrow|1_{s}\rangle_{\chi} is unlikely to occur due to the diffusion of produced pairs deep into the momentum phase space. We thus invoke an assumption that the time ordered product contains amplitudes of n pair production followed by n pair annihilation between sandwiched initial states, giving non-vanishing contributions in even orders alone. We thus have, in the HO potential region, \hskip-14.22636pt{}_{\rm rad}\!\langle n|_{\chi}\!\Pi_{s}\langle 0_{s}|\left(Q^{(+)}_{\rm HO}(0)\right)^{n}\Pi_{s}|1_{s}\rangle_{\chi}|0\rangle_{\rm rad}=n!\,(\chi^{(+)})^{n}\,. (62) Decays occur at n sites out of entire N_{s} sites. It is reasonable to take for \chi^{(+)} the half of field value, \displaystyle\chi^{(+)}=\frac{1}{\sqrt{2m_{\chi}V}}\,_{\chi}\!\langle 0_{s}|b_{s}|1_{s}\rangle_{\chi}=\frac{\chi}{2}\,. (63) Since n\,Q^{(+)}_{\rm HO} operators in ( 62 ) are independent at different sites, the relevant pair production is restricted to n=2 multiplied by the involved site number. This gives the factor to the expectation value, \displaystyle{\cal E}_{2n}=({\cal E}_{2})^{n}=\tilde{\chi}^{2n}\left(e^{-2i\Delta\omega_{k}t}-1+2i\Delta\omega_{k}t\right)^{n}\,. (64) The summation over all 2n orders give a simple geometric series, \displaystyle\sum_{n}{\cal E}_{2n}=\frac{1}{1+{\cal E}_{2}}\,. (65) We interpret this as the potential correction \langle\delta V_{\chi}\rangle using the relation, \displaystyle\exp[-i\langle\delta V_{\chi}\rangle tH^{-3}]=\frac{1}{1+{\cal E}_{2}}\,. (66) This gives \displaystyle\langle\delta V_{\chi}\rangle=-i\frac{1}{6}M_{\rm P}^{4}(\frac{z}{z_{0}})^{6}\ln(1+{\cal E}_{2})\,. (67) From the convergence of the series, one must impose a condition, \displaystyle|{\cal E}_{2}|=\tilde{\chi}^{2}\,\left[2(1-\cos 2\Delta\omega_{k}t)-4\Delta\omega_{k}t\sin 2\Delta\omega_{k}t\right. \displaystyle\hskip 28.45274pt\left.+4(\Delta\omega_{k}t)^{2}\right]^{1/2}<1\,. (68) It is readily proved that the potential correction \Re\langle\delta V_{\chi}\rangle is always negative, implying the tachyon-like instability at all higher orders. There are other types of perturbation. Consider a successive decay, \displaystyle|2_{s}\rangle_{\chi}\rightarrow|1_{s}\rangle_{\chi}\rightarrow|0_{s}\rangle_{\chi}\,, (69) that occurs the same site s starting from the initial state |\Psi(2,0)\rangle . If the successive decay always occur, one can sum up in the expectation value to derive \displaystyle\sum_{n}\overline{{\cal E}_{4n}}=\frac{1}{1-\overline{{\cal E}_{4}}}\,, (70) \displaystyle\overline{{\cal E}_{4}}=\tilde{\chi}^{4}\left(1-e^{-4i\Delta\omega_{k}t}-4i\Delta\omega_{k}t\right)\,. (71) If two decay patterns occur simultaneously, one derives \displaystyle\langle\Psi(2;0)|U(t,0)|\Psi(2;0)\rangle=\frac{1}{1-{\cal E}_{2}^{2}-\overline{{\cal E}_{4}}}\,, (72) \displaystyle{\cal E}_{2}^{2}+\overline{{\cal E}_{4}}=\tilde{\chi}^{4}\left((e^{-2i\Delta\omega_{k}t}-1+2i\Delta\omega_{k}t)^{2}\right. \displaystyle\hskip 28.45274pt\left.+1-e^{-4i\Delta\omega_{k}t}-4i\Delta\omega_{k}t\right)\,, (73) \displaystyle\tilde{\chi}^{4}=2.931\times 10^{-7}\,(c_{A}M_{\rm P})^{4}(\frac{z_{0}}{z})^{18}(\frac{\chi}{M_{\rm P}})^{4}\,. (74) Note the sign difference of two geometric series, and the instability of identified potential correction holds everywhere for the two cases. We call the type that contains 1/(1-{\cal E}_{2}^{2}-\overline{{\cal E}_{4}}) type II model, while the previous one type I model. In type II model tachyon instability starts from the term, \propto-\chi^{4} at the field origin. When we later discuss the inflaton exit problem, we shall compare results of these two types. A more general type of decay patterns are possible if one starts from the initial state |\Psi(n;0)\rangle\,,n\geq 3 . It is more reasonable to adopt higher type of models, since inflation is likely to end with highly excited \chi states. For simplicity we consider type II alone as a representative of these general decay types and compare results in two cases of I and II types. Our higher order calculations in perturbation theory is summation of a special class of diagrams, which we believe to be dominant over others. A more explicit demonstration of this assertion is difficult, but may be worthwhile to attempt. The tachyon instability is a pre-requisite to destabilize the effective field potential, defined by the potential correction \langle\delta V_{\chi}\rangle added to the original V_{\chi} . We shall calculate the time of destabilization denoted by t_{\rm D} defined by the time when \langle\delta V_{\chi}\rangle becomes equal to the maximum value of V_{\chi} . The condition gives \displaystyle\frac{4e^{-2}}{\gamma_{\chi}^{2}}V_{0}=\frac{M_{\rm P}^{4}}{6}(\frac{z}{z_{0}})^{6}\,\Im{\cal E}_{2}\,, (75) with 4e^{-2}=0.541 . The favored value of the potential height V_{0}=(10^{16}{\rm GeV})^{4} is suggested in koy 24-1 from observations. Thus, there exists a mismatch between the potential correction of order M_{\rm P}^{4} and the original potential V_{0} , unless the destabilization occurs close to the thermalization time. This suggests a possibility of inflaton exit out of trapped region immediately after emergence of tachyon instability. We add a comment to results so far derived. In order to somewhat reinforce this correlation, we next consider more general potential which has a harmonic part near the field origin, \chi\approx 0 , but its deviation from HO is made to appear far away from the field origin. The inflaton field in the interaction picture has time dependence of the form, \chi^{\pm}(t)e^{\mp im_{\chi}t} . One cannot explicitly time-integrate the interaction hamiltonian in this case. We shall focus on the potential region close to HO regime, but allow its small deviation from HO. The approximation of slow time variation may then be useful taking expansion to the first derivative term, \displaystyle\chi(t^{\prime})\approx\chi(t)+\dot{\chi}(t)\,(t^{\prime}-t)\,, (76) near t^{\prime}=t (taken as a reference point of time). This makes it possible to integrate both terms \propto\chi(t) and \propto\dot{\chi}(t) over t^{\prime} ; \displaystyle-i\int_{0}^{t}dt^{\prime}\,H^{\prime}_{\rm int}\sim-i\,2\sqrt{2}\frac{c_{A}\Omega_{k}m_{\chi}}{2\Delta\omega_{k}} \displaystyle\left(-(a_{k}^{\dagger}a_{-k}^{\dagger}+a_{k}a_{-k})\left(\chi(t)\,\sin 2\Delta\omega_{k}t\;\;\right.\right. \displaystyle\left.-\frac{\dot{\chi}(t)}{2\Delta\omega_{k}}\,(1-\cos 2\Delta\omega_{k}t+2\Delta\omega_{k}t\sin 2\Delta\omega_{k}t)\right) \displaystyle+\left(i(a_{k}^{\dagger}a_{-k}^{\dagger}-a_{k}a_{-k})\left(\chi(t)(1-\cos 2\Delta\omega_{k}t)\right.\right. \displaystyle\left.-\dot{\chi}(t)\,t\cos 2\Delta\omega_{k}t\right)\,. (77) The on-shell limit of this integral is \displaystyle-i\int_{0}^{t}dt^{\prime}\,H^{\prime}_{\rm int}(t^{\prime})\rightarrow i\,2\sqrt{2}\frac{c_{A}m_{\chi}\Omega_{k}}{2\Delta\omega_{k}} \displaystyle 2\Delta\omega_{k}t\,\left(a_{k}^{\dagger}a_{-k}^{\dagger}(\chi+\frac{\dot{\chi}}{2\Delta\omega_{k}}\,3\Delta\omega_{k}t)\right. \displaystyle\hskip-14.22636pt\left.-i\,a_{k}^{\dagger}a_{-k}^{\dagger}\left(\chi\Delta\omega_{k}t+\dot{\chi}t\right)+({\rm h.c.})\right)+O(\,(\Delta\Omega_{k}t)^{3}\,)\,. (78) New terms \propto\dot{\chi} added to the HO case enhances the unstable real part effect if the sign condition \chi\dot{\chi}\,\Delta\omega_{k}>0 holds. Throughout discussions so far we neglected effect of Hubble friction. This is justified when instability and particle production occurs much faster than the Hubble time, as is the case. When one considers a long time behavior, this should be corrected, as is made in the equation ( 40 ). IV Super-radiant stage of preheating We consider the preheating stage in which inflaton \chi field equation is described by HO (harmonic oscillator) potential. After recapitulating the Dicke model, we shall show that our quantum model provides extensions of the Dicke idea in two directions; extended coherence area and more general level structures. IV.1 Dicke algebraic model and its identification to our quantum theory There exists an important relation on correlated changes of the numbers of inflaton quanta and radiation pairs. Each time the inflaton decays, it produces a radiation pair, hence the total number of inflaton and radiation pair is conserved: \Delta(n_{\chi}+n_{\rm rad-pair})=0 . At a single site it is assumed that there are only two levels, the ground state |0\rangle and the excited state |1\rangle , and the excited state is allowed to pair emit along with the pair annihilation elevating the ground state to the excited state. The spin algebra of two-level system at a single site is defined by \displaystyle\hskip-8.5359ptj_{1}=\frac{1}{2}(|1\rangle\langle 0|+|0\rangle\langle 1|)\,,\hskip 8.5359ptj_{2}=\frac{i}{2}(-|1\rangle\langle 0|+|0\rangle\langle 1|)\,, (79) \displaystyle j_{3}=\frac{1}{2}(|1\rangle\langle 1|-|0\rangle\langle 0|)\,. (80) They satisfy the SU(2) commutation relation, \displaystyle[j_{k}\,,j_{l}]=i\epsilon_{klm}\,j_{m}\,. (81) Dicke introduces the ansatz that the correlated change proceeds by a hypothetical angular momentum change, thereby maintaining phase coherence and permutation symmetry of site exchange. The spin algebra is extended to the angular momentum algebra consisting of many spins at different sites s , \displaystyle\vec{J}=\sum_{s}\vec{j}^{(s)}\,,\hskip 8.5359pt[\vec{j}^{(s)}\,,\vec{j}^{(s^{\prime})}]=0\,;\;\;{\rm if\;}s\neq s^{\prime}\,. (82) The basic Dicke assumption is that photon emission occurs via J_{-} and photon absorption via J_{+} , and the rate of photon emission is in proportion to \displaystyle(J_{-})^{\dagger}J_{-}=J_{+}J_{-}=\sum_{ss^{\prime}}j_{+}^{(s^{\prime})}\,j_{-}^{(s)}\,. (83) The second Dicke’s assumption concerns the initial state, for which it is taken all sites in the excited state, hence \displaystyle|\Psi\rangle_{\chi}=\Pi_{s}\,|1\rangle_{s}\,. (84) The projected angular momentum M is defined in a convenient range, M=N_{s},N_{s}-1,N_{2}-2,\cdots 0 such that it is always non-negative. The relation M+n=N_{s} holds. We now come to the important correspondence relation. We may identify the angular momentum operator of Dicke type to our quantum operator according to \displaystyle\hskip 28.45274ptJ_{\mp}\;\Leftrightarrow\;Q^{(\pm)}(k)\,, (85) \displaystyle|M\rangle=\left(Q^{(+)}(k)\right)^{N_{s}-M}\,\Pi_{s}\,|1_{s}\rangle\!_{\chi}|0\rangle\!_{\rm rad}\,, (86) where |1_{s}\rangle\!_{\chi} stands for the \chi first excited state at site s . The site permutation symmetry in the Dicke model is automatically respected in our quantum model due to the site sum in Q^{(+)}(k) . Extensions beyond the Dicke model are readily taken into account in our model. If we assume that all sites are occupied in the first excited state of \chi quantum |1_{s}\rangle\!_{\chi} , the initial state is given by |\Psi\rangle_{i}^{(1)}=|\Psi(1;0)\rangle=\Pi_{s}\,|1_{s}\rangle\!_{\chi}|0\rangle\!_{\rm rad} . This choice is inevitable in the Dicke model that assumes two atomic levels. We have seen that another choice of |\Psi(2;0)\rangle adds a new interesting feature in our model. This choice provides a departure from the Dicke model. We call these Dicke-type I, Dicke-type II models in the present work. IV.2 Macro-coherent amplification Our model hamiltonian gives eJBD scalar decay into a pair of gauge bosons with momentum correlation, while the original Dicke model is based on a single photon decay of excited atoms. The phase coherence necessary for the enhanced decay is limited in the Dicke model to the wavelength \lambda region, since the phase factor e^{i2\pi x/\lambda} may deviate outside of |x|>\lambda/2\pi from e^{i2\pi/\lambda}\sim 1 . The situation dramatically changes for decay of correlated gauge boson pair decay, since this time the phase factor relation e^{i(\vec{k}-\vec{k})\cdot\vec{x}}=1 holds beyond the wavelength region. This is the essence of macro-coherent enhancement macro-coherence , which was experimentally verified in para-H 2 molecule into two photons, a weak QED process two-photon macro-coherence . The number of inflaton quanta (equal to the site number N_{s} ) that participate in macro-coherence is the \chi number density \rho_{\chi}/m_{\chi} times the Horizon volume, \displaystyle N_{s}=\frac{\rho_{\chi}}{m_{\chi}}H^{-3}=3^{3/2}\,\frac{\rho_{\chi}M_{\rm P}^{3}}{m_{\chi}^{4}\chi^{3}}\,(1+r)^{-3/2}\,, (87) with r the ratio of \chi energy density to the radiation energy density. At the Planck epoch all quantities here are of order, the Planck values. Hence this number is of order unity. As the preheating proceeds, N_{s} dramatically increases. At the end of preheating one may use the relation m_{\chi}^{2}\chi^{2}\sim V_{0}\approx(10^{16}{\rm GeV})^{4} , hence m_{\chi}\sim 10^{12}\, GeV, giving N_{s}=O(10^{12}) . In intermediate times one may use the adiabatic relation, a constant a(t)T(t) independent of time t , which gives the redshift ratio z_{R}/z_{0}\sim 10^{-2} at thermalization to at Planck epoch. This leads to N_{s} time evolution, \displaystyle N_{s}=O(1)(\frac{M_{\rm P}}{m_{\chi}})^{2}\frac{M_{\rm P}}{\chi}=O(1)(\frac{z_{0}}{z})^{6}\frac{M_{\rm P}}{\chi}\,. (88) There is thus a considerable ambiguity in estimation of the site number N_{s} . In view of this ambiguity we leave this number unspecified except when quantified numbers are given. Figure 4: Energy flux given by ( 94 ) in our macro-coherent Dicke model: three cases of different site numbers given by 2J+1 are plotted against perturbative decay rate \times time. The peak positions follow the delayed time formula given by \propto\ln(2J+1)/(2J+1) given in ( 93 ). Cosmic scale factor a(t) in different preheating stages is calculated using the energy sum of eJBD quanta and produced gauge boson pairs; \displaystyle(\frac{\dot{a}}{a})^{2}=\frac{1}{3M_{\rm P}^{2}}\left(\rho_{\chi}+\rho_{\rm rad}\right)\,, (89) \displaystyle\rho_{\chi}=\frac{1}{2}\dot{\chi}^{2}+\frac{1}{2}m_{\chi}^{2}\chi^{2}\propto a^{-3}\,,\hskip 8.5359pt\rho_{\rm rad}\propto a^{-4}\,. (90) Contribution from interaction energy has been neglected. In the initial stage of inflaton dominance one can ignore radiation term \propto\rho_{\rm rad} , and one derives the time dependence of matter-dominated universe, a(t)\propto t^{2/3} , while in later stage of radiation-dominance the scale factor varies a(t)\propto t^{1/2} . They give the power law of Hubble rate, H(t)=2/(3t) or 1/(2t) , respectively. Decay process of a collected body of excited states goes on in an entirely different way from stochastic exponential decay law. Coherence is first developed, along with slowly proceeding exponential decay, but with accumulated coherence a collected body of excited states suddenly disappear, leaving behind the ground state. The algebraic equation that describes the coherence evolution was proposed by Dicke: his equation of occupancy probability P_{M}(t) at the projected angular momentum M is \displaystyle\frac{dP_{M}}{\Gamma_{k}^{R}\,dt}=|\langle J,M+1|J^{+}|J,M\rangle|^{2}P_{M+1} \displaystyle-|\langle J,M-1|J^{-}|J,M\rangle|^{2}P_{M}\,.\hskip 8.5359pt\Gamma_{k}^{R}=\Gamma_{B}N_{s} (91) supplemented by well-known relations, |\langle J,M+1|J^{+}|J,M\rangle|^{2}=(N_{s}-M)(N_{s}+M+1) etc. The quantity \Gamma_{k}^{R} is decay constant enhanced by the total number of sites within the horizon. One can prove the law of probability conservation from ( 91 ), \displaystyle\frac{d}{dt}\,\sum_{M=-J}^{J}P_{M}=0\,. (92) Delayed decay time denoted by t_{d} is the important concept to characterize the sudden collapse of excited states sr textbook , and is given by \displaystyle t_{d}=\frac{\ln N_{s}}{N_{s}}\,\frac{1}{\Gamma_{k}^{R}}\,. (93) Note that the combination of denominator N_{s}\Gamma_{k}^{R}=\Gamma_{B}N_{s}^{2} is doubly enhanced by the site number N_{s} . The depletion of excited states leaves behind \chi field corresponding to zero-point oscillation. We may call the factor N_{s}^{2}/\ln N_{s} the speed up factor of macro-coherent super-radiance. In Fig( 4 ) we illustrate time evolution of energy flux I(t) given by \displaystyle\frac{I(t)}{J^{2}}=-\frac{1}{J^{2}}\sum_{M=-J\sim J}\,M\frac{dP_{M}(t)}{dt}\,, (94) in our macro-coherent Dicke model. As the number of participating sites increases, the peak of the flux shifts to shorter times approximately according to the formula of delayed time \propto\ln(2J+1)/(2J+1) given in ( 93 ), at the same time with narrower widths of duration time, as expected. Perturbative decays occur only for 1-e^{-0.06}\sim 0.058 (at the time of 0.06\times the usual lifetime) of participating inflatons, while the decay is completed (almost completed in the J=100 case) in three J cases shown. V Transition from inflationary to dark energy stages of acceleration V.1 Inflaton escape probability We first explain our method of solving the exit problem. In our quantum formulation we can set up the Schroedinger equation of homogeneous inflaton field \chi ; \displaystyle i\hbar\frac{\partial\Psi}{\partial t}=-\frac{\hbar^{2}}{2}\frac{\partial^{2}\Psi}{\partial\chi^{2}}+\tilde{V}(\chi)\Psi\,, (95) \displaystyle\tilde{V}(\chi)=\left(V_{\chi}+\Re\langle\delta V_{\chi}\rangle\right)\,H^{-3}\,. (96) This equation is solved under the WKB ansatz in \hbar\rightarrow 0 limit; \displaystyle\Psi_{E}(\chi,t)=\exp[-\frac{i\,Et}{\hbar}+\frac{i}{\hbar}W(\chi;E)]\,. (97) To leading orders, its solution is given by \displaystyle\Psi_{E}(\chi,t)=N\,(E-\tilde{V}(\chi)\,)^{-1/4}\,\exp[-iEt] \displaystyle\hskip 14.22636pt\sin\left(\int_{a}^{\chi}d\chi\sqrt{E-\tilde{V}(\chi)}-\frac{\pi}{4}\right)\,, (98) with N the normalization and with \hbar dependence deleted. a is the turning point at which E-\tilde{V}(\chi=a)=0 . The reflected component is obtained by the replacement, \displaystyle\sin\left(\int_{a}^{\chi}d\chi\sqrt{E-\tilde{V}(\chi)}-\frac{\pi}{4}\right)\rightarrow \displaystyle\frac{\exp[i\left(\int_{a}^{\chi}d\chi\sqrt{E-\tilde{V}(\chi)}-\frac{\pi}{4}\right)]}{2i}\,. (99) The potential correction given to the second order of interaction describes its behavior near the field origin, establishing the tachyon mass. But it is not sufficient to describe the potential region around the original potential maximum. The potential correction should be applied to larger field values, and we adopt type II model giving the potential correction \propto\ln(1-{\cal E}_{2}-\overline{{\cal E}}_{4}) , as derived in ( 61 ). The inflaton distribution probability, using the reflected component, is given by \displaystyle|\Psi_{E}(\chi,t)|^{2}=\frac{|N|^{2}}{4\sqrt{|E-H^{-3}\langle\delta V_{\chi}\rangle|}}\,,\hskip 8.5359pt|E-H^{-3}\langle\delta V_{\chi}\rangle| \displaystyle\hskip 28.45274pt=|\,E-\frac{\sqrt{3}}{2}M_{\rm P}(\frac{z}{z_{0}})^{6}\Im\ln(1-{\cal E}_{2}^{2}-\overline{{\cal E}}_{4})\,|\,, (100) \displaystyle\hskip 28.45274pt|{\cal E}_{2}^{2}+\overline{{\cal E}}_{4}|<1\,. (101) The constraint ( 101 ) gives a time dependent upper limit on \chi field; \displaystyle\hskip 28.45274pt\frac{\chi}{M_{\rm P}}<\frac{32\pi^{2}}{3\sqrt{6}}\frac{1}{c_{A}M_{\rm P}}(\frac{z}{z_{0}})^{9/2}\,\times \displaystyle\hskip-14.22636pt\left[|(1-e^{-2i\Delta\omega_{k}t}-2i\Delta\omega_{k}t)^{2}+1-e^{-4i\Delta\omega_{k}t}-4i\Delta\omega_{k}t|\right]^{-1/4}\,. Figure 5: Escape probability, calculated using WKB probability formula ( 100 ) of type II and I models with the constraint ( 101 ), of eJBD inflaton out of trapped region plotted against a small dimensionless time \Delta\omega_{k}t (in terms of m_{\chi} the time axis should be multipled by m_{\chi}/\Delta m_{\chi}=10\sim 100 ). The trapped region is defined by the interior to the potential maximum of the original potential V_{\chi}=1.8\times 10^{-13} in the Planck unit at \chi_{\rm max} . The inflaton energy is taken as E=0 and to avoid the singularity the region inside the trapped region is defined by \chi_{\rm min}=\chi_{\rm max}/100 . Used parameters are a common \gamma_{g}=1 and z/z_{0}=0.1 of II type in solid and 0.01 dotted black, 0.1 of I type in dashed orange and 0.01 dash-dotted blue. One must be careful about the artifact of WKB approximation which gives a divergence at the classical turning point at E-H^{-3}\langle\delta V_{\chi}\rangle=0 . We avoid the problem by an appropriate choice of the energy E . The escape probability differs in different models, for instance, type I or type II model. We illustrate four parameter cases of escape probability |\Psi_{E}(\chi,t)|^{2} in Fig( 5 ) for E=0 . The normalization N has been calculated including the interior trapped region. Type I model gives a larger escape probability than type II model. The termination time of escape is estimated by the time when \chi energy is drained into radiation. One method of estimation is to calculate the total decay rate and use the constraint relation, {\cal E}_{2}=1 or {\cal E}_{2}^{2}+\overline{{\cal E}}_{4}=1 , depending on the model type, I or II. Total decay rate in type I model is given by \displaystyle\hskip 28.45274pt\Gamma_{B}\sum_{n}|{\cal E}_{2}|^{2n}=\frac{\Gamma_{B}}{1-|{\cal E}_{2}|^{2}}\,, (103) \displaystyle|{\cal E}_{2}|^{2}=4\tilde{\chi}^{2}|\,1-2\cos 2\Delta\omega_{k}t+2\Delta\omega_{k}(\Delta\omega_{k}-\sin 2\Delta\omega_{k}t\,|\,. The function 1/(1-|{\cal E}_{2}|^{2}) that appears in ( 103 ) is illustrated in Fig( 6 ), which shows that the escape should be completed within a few to several times \Delta\omega_{k}t at relevant redshifts z_{\rm escape} . Combined with redshift values used in Fig( 6 ), we observe that the escape occurs early during the preheating stage. The potential correction is forced to disappear when all inflaton excitation energy is drained into radiation pairs. The drained time denoted by t_{D} or its redshift z_{D} is estimated by equating HO \chi energy density \rho_{\chi} (its non-decay part) \propto z^{3} to the maximum potential energy density 2V_{0}/\gamma_{\chi}^{2} . Using our favorite values of V_{0}\,,\gamma_{\chi} , we derive \displaystyle\frac{z_{D}}{z_{0}}=\left(\frac{2V_{0}}{\gamma_{\chi}^{2}\,M_{\rm P}^{4}}\right)^{1/3}\sim 2.1\times 10^{-3}\,. (105) This is roughly of the same order of magnitude as the thermalization time, and is later than the escape time of order z_{\rm escape}/z_{0}=O(0.1\sim 0.01) . Figure 6: Total decay rate given by 1/(1-|{\cal E}_{2}|^{2}) : Four cases of (c_{A}M_{\rm P}),z,\chi/M_{\rm P} in type I are plotted: (0.01,0.1,0.1) in solid black, (0.01,0.05,0.1) in dashed orange (0.01,0.05, 0.05) in dash-dotted blue, and (0.01,0.1,0.05) in dotted black. V.2 Epochs of thermalization and potential restoration We have so far considered transverse gauge boson pair production at the preheating stage. This process creates a universe dominated by gauge boson pairs far from chemical and thermal equilibrium. The minimum requirement for thermalization is to produce energetically different particles and utilize scattering between particles of different energies, thereby exchanging their energies. The simplest mechanism is to take into account massive gauge boson decay into lighter fermion anti-fermion pair. Recall that prior to thermalization the gauge symmetry is spontaneously broken and various gauge bosons are massive and they can decay into lighter fermion pairs. In koy 24-2 we estimated the time at which thermalization takes place, and the result may be given in terms of the redshift factor, \displaystyle\frac{z_{\rm th}}{z_{0}}=O(2\times 10^{-3}\sim 4\times 10^{-2})\,, (106) using coupling constants and the number of particle species in standard and grand unified models. The redshift ratio ( 106 ) is larger than the ratio ( 105 ), which means that the termination of potential correction occurs after thermalization. Without potential correction the original inflaton potential is restored, and \chi field continues to roll down towards the field infinity in thermalized radiation-dominated universe. The field motion describes a quintessential dark energy. After thermal equilibrium is realized, there is an additional contribution to the field potential which restores the standard model gauge symmetry. One can set up an approximate set of time evolution equations for eJBD field, radiation energy density and the cosmic scale factor (or the redshift factor). But a more elaborate approach that takes into account thermalization effect is desirable to derive conclusive results, since thermalization and the recovery to the original potential proceeds simultaneously. VI Summary and outlook We have proposed a quantum theory of how the universe after inflation underwent the transition to hot big band universe that realizes a kind of phase transition to the late-time accelerating universe. Incorporation of particle physics into an extended class of Jordan-Brans-Dicke gravity is of central importance, and we generalized the original JBD idea of matter coupling maintaining five independent gauge invariant terms separately with different conformal factors. We determined the major matter coupling of eJBD inflaton to various gauge-boson pairs by the unambiguous trace anomaly formula. The eJBD field coupling via the trace anomaly, giving rise to gauge-boson pair decay, is of fundamental importance in enhancing super-radiant inflaton decay due to a high level of macro-coherence. In the initial preheating stage the inflaton motion is described by a harmonic oscillator, and interaction of inflaton to matter induces the inflaton decay into gauge bosons of correlated momentum pair, back-to-back. Quantum theory of inflaton and gauge-boson pairs have been fully developed. Standard particle physics techniques lead to an orchestrated particle pair production accompanied by tachyonic instability giving rise to an inflaton potential change. The tachyon instability gives a temporary epoch of destabilized inflaton potential, which pushes the inflaton trapped in the harmonic well to cross over a local potential maximum, ultimately towards the field infinity. This offers a conversion mechanism that bridges between inflationary acceleration and the dark energy acceleration. The present single eJBD field model is more economic, hence more restrictive than the model of two fields advocated in our previous work koy 24-1 . Our new theory is a tremendous simplification of calculation over the previous methods kls , my 95 and fkyy that relied on mode summation in instability bands of the Mathieu chart. The new approach bypasses this complicated procedure, yet one can derive important issues connected to the Dicke super-radiance model. Two extended types of Dicke model have been considered in detail. Type I Dicke model gives slightly larger escape probability and faster ending than type II, but not overwhelmingly better. Both types are acceptable to give large enough escape probabilities. Convenient tools of dedicated numerical computations need to be developed, but it should be clear that the scenario presented in this work is realized using an acceptable set of parameters in the model. Our model is described by a few parameters of eJBD gravity related to inflaton coupling of standard matter. These parameters greatly influence the late stage of cosmic evolution. As is made evident in koy 24-1 , the effective scalar mass after nucleo-synthesis is of the order of the inverse Hubble size, hence in problems of astrophysical interest, the scalar is essentially massless with its coupling to matter of the gravitational strength. This may or may not influence gravitational wave emission at merger events. It is expected that mergers of black hole and neutron star are a good target to search for deviation from the prediction of general relativity koy bh-ns merger . Thus, our model can be tested in forthcoming observations and laboratory experiments. These problems shall be worked out in separate publications. VII Appendix: Relevance of trace anomaly Figure 7: Three diagrams contributing to inflaton decay into massless gauge boson pair. Solid lines are for fermions attached to wavy gauge bosons and long dotted inflaton field, while black circles are the trace anomaly. We first discuss how a new conformal coupling term should be added to those of our previous works koy 24-1 , koy 24-2 . To see how the eJBD field couples to the trace of the energy-momentum tensor, we begin by writing the action as \displaystyle\hskip-8.5359ptS=\int d^{4}x\sqrt{-g_{J}}\left[-\frac{M_{P}^{2}}{2}F_{g}(\chi)R_{J}+{\cal L}_{\chi}+{\cal L}_{\rm SM}\right], (107) where {\cal L}_{\chi} and {\cal L}_{\rm SM} are assumed to include couplings of the eJBD field in a gauge invariant manner as in koy 23 . To transform from the Jordan metric frame to the Einstein metric frame, we perform the Weyl rescaling given by g_{J\mu\nu}=F^{-1}_{g}(\chi)g_{\mu\nu} . Suppose that \chi takes a nonzero homogeneous contribution \chi=\chi_{0}= constant and that we redefine \chi as a fluctuation on \chi_{0} , namely, \chi\to\chi_{0}+\chi with |\chi/\chi_{0}|\ll 1 . This setup allow us to expand as \displaystyle g_{J\mu\nu}=F^{-1}_{g}(\chi_{0})\left(1-\partial_{\chi}\ln F_{g}(\chi_{0})\chi+\cdots\right)g_{\mu\nu}. \displaystyle\equiv F^{-1}_{g}(\chi_{0})(g_{\mu\nu}+\delta g_{\mu\nu}). (108) Note that the overall constant factor may be removed by a simple rescaling, so we omit the factor F_{g}^{-1}(\chi_{0}) in the following argument. The relevant interaction of \chi can be identified by varying the action S with respect to \delta g_{\mu\nu} , to derive \displaystyle S_{\rm int}=\frac{1}{2}\int d^{4}x\sqrt{-g}T_{\mu\nu}\delta g^{\mu\nu} \displaystyle=-\int d^{4}x\sqrt{-g}\gamma_{J}\frac{\chi}{2M_{P}}T\,, (109) with T=g^{\mu\nu}T_{\mu\nu} . There is a piece of contribution to the trace of the energy-momentum tensor in the standard model, \displaystyle T_{\mu}^{\mu}=\sum_{A_{\mu}}\frac{4-d}{4}F^{\mu\nu}F_{\mu\nu}+(\cdots)\,. (110) The trace anomaly exists since the gauge boson pair contributes at one loop level. We find by computing the self-energy diagram of gauge boson that this gives rise to a finite contribution to the decay amplitude, the divergent contribution \propto 1/(4-d) in dimensional regularization scheme canceling 4-d in ( 110 ); \displaystyle-\frac{b_{A}\alpha_{A}}{8\pi}F^{{\rm A}\mu\nu}F^{\rm A}_{\mu\nu}\,,\hskip 8.5359pt\frac{d\alpha_{A}^{-1}}{d\ln\mu}=-\frac{b_{A}}{2\pi}\,. (111) b_{A} thus defined is the beta-function coefficient, with \mu being the renormalization scale. Acknowledgements. This research was partially supported by JSPS KAKENHI Nos. 21H01107 (KO) and 21K03575(MY).