Introduction
Cosmological backreaction is fundamentally a question about expansion: how inhomogeneous perturbations alter the local expansion and how these effects modify the effective evolution relative to a homogeneous Friedmann background. The expansion scalar Θ \Theta Θ of the hypersurface-normal congruence is the negative trace of the extrinsic curvature, Θ = − K \Theta= -K Θ = − K , and therefore controls the local volume expansion of the chosen foliation. Among the complete scalar gauges that we have studied so far, the uniform-expansion gauge is the one most directly adapted to this geometrical aspect of the backreaction problem. With the common spatial gauge choice E = 0 E = 0 E = 0 , the temporal condition of the uniform-expansion gauge fixes the perturbation of the extrinsic-curvature trace itself, κ = 0 \kappa= 0 κ = 0 , rather than a matter variable (δ ϕ = 0 \delta\phi= 0 δ ϕ = 0 in the comoving gauge or δ ρ = 0 \delta\rho= 0 δ ρ = 0 in the uniform-density gauge), the scalar shear (β = 0 \beta= 0 β = 0 in the longitudinal gauge), or the intrinsic spatial curvature (ψ = 0 \psi= 0 ψ = 0 in the spatially-flat gauge). Uniform-expansion slicing is closely related to constant-mean-curvature slicing, and has a natural role in nonlinear and long-wavelength descriptions of cosmological inhomogeneity [1 , 2 , 3 , 4 ] . This direct geometrical connection does not by itself make a gauge-fixed effective source an observable, but it makes the uniform-expansion sector especially informative when the object under discussion is cosmological backreaction.
At quadratic order the influence of linear perturbations can be collected into a quadratic effective energy-momentum tensor (2EMT), obtained by rearranging the second-order Einstein equation so that genuine second-order fields appear linearly while products of first-order perturbations act as an effective source [5 , 6 ] . The interpretation of such quadratic sources and of cosmological backreaction has long involved gauge subtleties [7 , 8 , 9 ] . In our previous calculations the scalar 2EMT was evaluated explicitly in several complete gauges, revealing structured rather than arbitrary dependence on slicing and threading [10 , 11 , 12 , 13 ] . For comparison, we use the results for the longitudinal, spatially-flat, comoving, and uniform-density gauges obtained in Ref. [13 ] , where the relevant background quantities were expressed exactly in terms of the slow-roll parameters before performing the slow-roll expansion. The UE results are evaluated using the same procedure.
A second principal aim of the present work is to address this gauge question directly at the level of the general quadratic source before any gauge condition is imposed. The early 2EMT formulations were constructed to describe backreaction in a gauge-invariant manner [5 , 6 ] , while subsequent analyses emphasized the gauge and observer dependence of second-order effective sources and their physical interpretation [7 , 8 , 9 ] . Previous calculations, including our own, primarily diagnosed the issue by comparing results obtained after different gauge choices. Here, no temporal or spatial gauge condition is imposed initially. We apply an arbitrary infinitesimal scalar gauge transformation ξ μ = ( ξ 0 , ∂ i ζ ) \xi^\mu= (\xi^0,\partial^i\zeta) ξ μ = ( ξ 0 , ∂ i ζ ) directly to the general quadratic source in configuration space and evaluate its component variation. The 00 and diagonal spatial components retain nonvanishing even-in-k k k contributions for general gauge functions, and therefore the general scalar 2EMT before gauge fixing is not gauge invariant. To our knowledge, this is the first explicit component-level evaluation of the general scalar-gauge variation of the quadratic source in this 2EMT framework. The calculation also separates two logically distinct questions: gauge invariance of the general quadratic source before gauge fixing, and uniqueness of the representation selected by a complete gauge fixing.
The most recent member of that sequence was the uniform-density analysis [13 ] . There the temporal slicing is defined by a slowly evolving matter clock, ρ \rho ρ , and the ultraviolet 2EMT contains a 1 / ϵ 1/\epsilon 1/ ϵ (ϵ \epsilon ϵ : slow-roll parameter) enhancement associated with the conditioning of that clock, together with an additional 1 / σ 2 2 1/\sigma_2^2 1/ σ 2 2 (σ 2 ≡ H / k \sigma_2 \equiv\mathcal{H}/k σ 2 ≡ H / k : short-wavelength parameter) gradient factor from the Laplacian contribution to the density constraint. The uniform-expansion gauge considered here is qualitatively different. It is a geometrical-expansion clock rather than a matter clock, and the solution of its constraint introduces the characteristic combination
k 2 + 3 ϵ H 2 . (1) k^2 + 3\epsilon\mathcal{H}^2. \tag*{(1)} k 2 + 3 ϵ H 2 . (1)
Thus the new problem is controlled by a competition between the perturbation-gradient scale k 2 k^2 k 2 and the slow-roll geometrical scale ϵ H 2 \epsilon\mathcal{H}^2 ϵ H 2 .
The UE results obtained in this work exhibit substantially different infrared behavior from those found in the four previously studied gauges. At exactly zero gradient both UE source components vanish. At small nonzero k k k , with σ 1 ≡ k / H \sigma_1 \equiv k/\mathcal{H} σ 1 ≡ k / H denoting the long-wavelength parameter, their first nonzero terms are
τ ^ 00 U E F L W ≃ 8 ϵ σ 1 2 , τ ^ i i U E F L W ≃ − 8 3 ϵ σ 1 2 , (2) \frac{\widehat{\tau}^{\mathrm{UE}}_{00}}{\mathcal{F}_{\mathrm{LW}}} \simeq\frac{8}{\epsilon}\sigma_1^2,\qquad
\frac{\widehat{\tau}^{\mathrm{UE}}_{ii}}{\mathcal{F}_{\mathrm{LW}}} \simeq-\frac{8}{3\epsilon}\sigma_1^2, \tag*{(2)} F LW τ 00 UE ≃ ϵ 8 σ 1 2 , F LW τ ii UE ≃ − 3 ϵ 8 σ 1 2 , (2)
at leading slow-roll order. Here we define the diagnostic ratio w r a w U E ≡ τ ^ i i U E / τ 00 U E w_{\mathrm{raw}}^{\mathrm{UE}} \equiv\widehat{\tau}^{\mathrm{UE}}_{ii}/\tau^{\mathrm{UE}}_{00} w raw UE ≡ τ ii UE / τ 00 UE , which approaches − 1 / 3 -1/3 − 1/3 in the strict infrared limit at leading slow-roll order. The first finite-gradient coefficients differ from the comoving and uniform-density sectors.
On the gradient-dominated side, sufficiently far from the transition that inverse-σ 1 \sigma_1 σ 1 corrections are subleading, the leading UE behavior approaches the spatially-flat sector. The same denominator makes the asymptotic analysis genuinely nonuniform: factors of ( k 2 + 3 ϵ H 2 ) − n (k^2+3\epsilon\mathcal{H}^2)^{-n} ( k 2 + 3 ϵ H 2 ) − n generate inequivalent series in the strict infrared hierarchy σ 1 2 ≪ ϵ \sigma_1^2 \ll\epsilon σ 1 2 ≪ ϵ and in the slow-roll-first hierarchy ϵ , ∣ δ ∣ ≪ σ 1 2 ≪ 1 \epsilon,|\delta|\ll\sigma_1^2\ll1 ϵ , ∣ δ ∣ ≪ σ 1 2 ≪ 1 (δ \delta δ : slow-roll parameter). Inverse powers of σ 1 \sigma_1 σ 1 at higher slow-roll orders are therefore not strict-infrared divergences; they signal nonuniformity as the transition is approached. The scale σ 1 2 ∼ 3 ϵ \sigma_1^2\sim3\epsilon σ 1 2 ∼ 3 ϵ marks the crossover of the UE constraint denominator, but it does not by itself define a sharp crossover of the full 2EMT. Near this region the full rational expressions, rather than a finite slow-roll-first truncation, must be used.
A second distinctive feature appears before the Fourier reduction. The condition κ = 0 \kappa=0 κ = 0 is a differential constraint for the shift variable Q ≡ β + E ′ Q\equiv\beta+E' Q ≡ β + E ′ in configuration space, and the gauge-fixed 2EMT retains Q Q Q , Q ′ Q' Q ′ , and Q ′ ′ Q'' Q ′′ in an unreduced derivative representation. In Fourier space the same constraint becomes an algebraic relation for each Fourier mode, allowing Q k Q_k Q k to be eliminated in favor of the Bardeen potential Ψ k \Psi_k Ψ k . This reduction is a convenient representation of the UE gauge-fixed result, but it is not by itself a gauge-invariance criterion. Once the gauge is completely fixed, equivalent variables may be used to rewrite the same uniquely specified gauge-fixed quantity.
In the strict ultraviolet limit, the UE source has the same leading coefficients as the longitudinal and spatially-flat sectors, τ ^ 00 U E ≃ 2 F S W \widehat{\tau}^{\mathrm{UE}}_{00}\simeq2\mathcal{F}_{\mathrm{SW}} τ 00 UE ≃ 2 F SW and τ ^ i i U E ≃ ( 2 / 3 ) F S W \widehat{\tau}^{\mathrm{UE}}_{ii}\simeq(2/3)\mathcal{F}_{\mathrm{SW}} τ ii UE ≃ ( 2/3 ) F SW , yielding w r a w U E → 1 / 3 w_{\mathrm{raw}}^{\mathrm{UE}}\to1/3 w raw UE → 1/3 .
The paper is organized as follows. In Sec. II we review the scalar perturbations, gauge-invariant variables, linear solutions, wavelength hierarchies, and the definition of the 2EMT, using the same conventions and presentation as the uniform-density companion paper. In Sec. III we first review the meaning of a scalar gauge transformation and the criterion for gauge invariance, and then apply the general transformation directly to the quadratic source before gauge fixing. We next relate the local expansion to the trace of the extrinsic curvature, impose the uniform-expansion gauge, derive the Fourier-space shift, and discuss complete gauge fixing and the uniqueness of the resulting gauge-fixed representation. Sections IV and V present the full long- and short-wavelength 2EMTs and analyze the strict and intermediate regimes. Section VI compares all five gauges and develops the scale-transition interpretation, including the e e e -fold evolution of a fixed mode. Section VII discusses the interpretation of the gauge-fixed result and its distinction from a fully specified relational observable. We conclude in Sec. VIII. Appendix A gives the explicit general scalar-gauge variation of the quadratic source before gauge fixing, while Appendix B records the unreduced uniform-expansion 2EMT before eliminating Q Q Q . The lengthy general 2EMT and the full expressions for the other four gauges are not repeated here; they are available in Refs. [11 , 13 ] .
Scalar perturbations and the quadratic effective source
The perturbative setup, notation, linear equations, Fourier conventions, and definition of the quadratic source in this section are the same as those given in Sec. II of the uniform-density companion paper Ref. [13 ] . We reproduce them here for convenience and to keep the present analysis self-contained.
We consider scalar perturbations of a spatially flat Friedmann background,
d s 2 = a 2 ( η ) { − ( 1 + 2 α ) d η 2 − 2 β , i d η d x i + [ ( 1 − 2 ψ ) δ i j + 2 E , i j ] d x i d x j } , (3) ds^2=a^2(\eta)\left\{-(1+2\alpha)d\eta^2-2\beta_{,i}d\eta\,dx^i+\left[(1-2\psi)\delta_{ij}+2E_{,ij}\right]dx^i dx^j\right\}, \tag*{(3)} d s 2 = a 2 ( η ) { − ( 1 + 2 α ) d η 2 − 2 β , i d η d x i + [ ( 1 − 2 ψ ) δ ij + 2 E , ij ] d x i d x j } , (3)
and a canonical scalar field
ϕ ( η , x ) = ϕ 0 ( η ) + δ ϕ ( η , x ) . (4) \phi(\eta,\mathbf{x})=\phi_0(\eta)+\delta\phi(\eta,\mathbf{x}). \tag*{(4)} ϕ ( η , x ) = ϕ 0 ( η ) + δ ϕ ( η , x ) . (4)
The energy-momentum tensor is
T μ ν = ϕ , μ ϕ , ν − g μ ν [ 1 2 g ρ σ ϕ , ρ ϕ , σ + V ( ϕ ) ] . (5) T_{\mu\nu}=\phi_{,\mu}\phi_{,\nu}-g_{\mu\nu}\left[\frac{1}{2}g^{\rho\sigma}\phi_{,\rho}\phi_{,\sigma}+V(\phi)\right]. \tag*{(5)} T μν = ϕ , μ ϕ , ν − g μν [ 2 1 g ρ σ ϕ , ρ ϕ , σ + V ( ϕ ) ] . (5)
The background equations are
H 2 = 8 π G 3 ( 1 2 ϕ 0 ′ 2 + a 2 V ) , (6) \mathcal{H}^2=\frac{8\pi G}{3}\left(\frac{1}{2}\phi_0^{\prime2}+a^2V\right), \tag*{(6)} H 2 = 3 8 π G ( 2 1 ϕ 0 ′2 + a 2 V ) , (6)
2 H ′ + H 2 = 8 π G ( − 1 2 ϕ 0 ′ 2 + a 2 V ) , (7) 2\mathcal{H}^{\prime}+\mathcal{H}^2=8\pi G\left(-\frac{1}{2}\phi_0^{\prime2}+a^2V\right), \tag*{(7)} 2 H ′ + H 2 = 8 π G ( − 2 1 ϕ 0 ′2 + a 2 V ) , (7)
ϕ 0 ′ ′ + 2 H ϕ 0 ′ + a 2 V ϕ = 0 , (8) \phi_0^{\prime\prime}+2\mathcal{H}\phi_0^{\prime}+a^2V_\phi=0, \tag*{(8)} ϕ 0 ′′ + 2 H ϕ 0 ′ + a 2 V ϕ = 0 , (8)
where H = a ′ / a \mathcal{H}=a^{\prime}/a H = a ′ / a and a prime denotes d / d η d/d\eta d / d η .
Under the infinitesimal transformation x μ → x μ + ξ μ x^\mu\to x^\mu+\xi^\mu x μ → x μ + ξ μ generated by ξ μ = ( ξ 0 , ∂ i ζ ) \xi^\mu=(\xi^0,\partial^i\zeta) ξ μ = ( ξ 0 , ∂ i ζ ) , the perturbation variables transform as
α ~ = α − ξ 0 ′ − H ξ 0 , (9) \widetilde{\alpha}=\alpha-\xi^{0\prime}-\mathcal{H}\xi^0, \tag*{(9)} α = α − ξ 0′ − H ξ 0 , (9)
β ~ = β − ξ 0 + ζ ′ , (10) \widetilde{\beta}=\beta-\xi^0+\zeta^{\prime}, \tag*{(10)} β = β − ξ 0 + ζ ′ , (10)
ψ ~ = ψ + H ξ 0 , (11) \widetilde{\psi}=\psi+\mathcal{H}\xi^0, \tag*{(11)} ψ = ψ + H ξ 0 , (11)
E ~ = E − ζ , (12) \widetilde{E}=E-\zeta, \tag*{(12)} E = E − ζ , (12)
δ ϕ ~ = δ ϕ − ϕ 0 ′ ξ 0 . (13) \widetilde{\delta\phi}=\delta\phi-\phi_0^{\prime}\xi^0. \tag*{(13)} δ ϕ = δ ϕ − ϕ 0 ′ ξ 0 . (13)
Introducing
Q = β + E ′ , Q ~ = Q − ξ 0 , (14) Q=\beta+E^{\prime},\qquad\widetilde{Q}=Q-\xi^0, \tag*{(14)} Q = β + E ′ , Q = Q − ξ 0 , (14)
we define the first-order gauge-invariant combinations in the standard scalar-perturbation framework [14 ]
Φ = α − Q ′ − H Q , (15) \Phi=\alpha-Q^{\prime}-\mathcal{H}Q, \tag*{(15)} Φ = α − Q ′ − H Q , (15)
Ψ = ψ + H Q , (16) \Psi=\psi+\mathcal{H}Q, \tag*{(16)} Ψ = ψ + H Q , (16)
δ ϕ g i = δ ϕ − ϕ 0 ′ Q . (17) \delta\phi_{\mathrm{gi}}=\delta\phi-\phi_0^{\prime}Q. \tag*{(17)} δ ϕ gi = δ ϕ − ϕ 0 ′ Q . (17)
Here, Φ \Phi Φ and Ψ \Psi Ψ are the so called Bardeen potentials. For a canonical scalar field without anisotropic stress, Φ = Ψ \Phi=\Psi Φ = Ψ .
The first-order equations give
δ ϕ g i = Ψ ′ + H Ψ 4 π G ϕ 0 ′ , (18) \delta\phi_{\mathrm{gi}}=\frac{\Psi^{\prime}+\mathcal{H}\Psi}{4\pi G\phi_0^{\prime}}, \tag*{(18)} δ ϕ gi = 4 π G ϕ 0 ′ Ψ ′ + H Ψ , (18)
δ ϕ g i ′ = Δ Ψ − K Ψ ′ − L Ψ 4 π G ϕ 0 ′ , (19) \delta\phi_{\mathrm{gi}}^{\prime}=\frac{\Delta\Psi-\mathcal{K}\Psi^{\prime}-\mathcal{L}\Psi}{4\pi G\phi_0^{\prime}}, \tag*{(19)} δ ϕ gi ′ = 4 π G ϕ 0 ′ ΔΨ − K Ψ ′ − L Ψ , (19)
where
K = 3 H + a 2 V ϕ ϕ 0 ′ , L = H ′ + 2 H 2 + a 2 H V ϕ ϕ 0 ′ . (20) \mathcal{K}=3\mathcal{H}+a^2\frac{V_\phi}{\phi_0^{\prime}},\qquad\mathcal{L}=\mathcal{H}^{\prime}+2\mathcal{H}^2+a^2\mathcal{H}\frac{V_\phi}{\phi_0^{\prime}}. \tag*{(20)} K = 3 H + a 2 ϕ 0 ′ V ϕ , L = H ′ + 2 H 2 + a 2 H ϕ 0 ′ V ϕ . (20)
The Bardeen potential obeys
Ψ ′ ′ − Δ Ψ + 2 K Ψ ′ + 2 L Ψ = 0. (21) \Psi^{\prime\prime}-\Delta\Psi+2\mathcal{K}\Psi^{\prime}+2\mathcal{L}\Psi=0. \tag*{(21)} Ψ ′′ − ΔΨ + 2 K Ψ ′ + 2 L Ψ = 0. (21)
Slow-roll and wavelength series expansions
We use the slow-roll parameters
ϵ ≡ − H ˙ H 2 = 4 π G ϕ ˙ 0 2 H 2 , δ ≡ − ϕ ¨ 0 H ϕ ˙ 0 = ϵ − ϵ ˙ 2 H ϵ . (22) \epsilon\equiv-\frac{\dot{H}}{H^2}=4\pi G\frac{\dot{\phi}_0^2}{H^2},\qquad\delta\equiv-\frac{\ddot{\phi}_0}{H\dot{\phi}_0}=\epsilon-\frac{\dot{\epsilon}}{2H\epsilon}. \tag*{(22)} ϵ ≡ − H 2 H ˙ = 4 π G H 2 ϕ ˙ 0 2 , δ ≡ − H ϕ ˙ 0 ϕ ¨ 0 = ϵ − 2 H ϵ ϵ ˙ . (22)
where H = a ˙ / a H=\dot{a}/a H = a ˙ / a and an overdot denotes the cosmic-time derivative. We introduce an additional dimensionless parameter characterizing the third cosmic-time derivative of the background inflaton:
ξ 3 ≡ ϕ ... 0 H 2 ϕ ˙ 0 . (23) \xi_3\equiv\frac{\dddot{\phi}_0}{H^2\dot{\phi}_0}. \tag*{(23)} ξ 3 ≡ H 2 ϕ ˙ 0 ϕ ... 0 . (23)
This parameter appears in the exact evolution equation for δ \delta δ , namely δ ˙ = H [ δ ( ϵ + δ ) − ξ 3 ] \dot{\delta}=H[\delta(\epsilon+\delta)-\xi_3] δ ˙ = H [ δ ( ϵ + δ ) − ξ 3 ] , as well as in the expression for V ϕ ϕ V_{\phi\phi} V ϕϕ . In the slow-roll hierarchy, ξ 3 \xi_3 ξ 3 is of second order, together with ϵ 2 \epsilon^2 ϵ 2 , ϵ δ \epsilon\delta ϵδ , and δ 2 \delta^2 δ 2 . The exact background relations used in our calculation are
ϵ ′ = 2 H ϵ ( ϵ − δ ) , δ ′ = H [ δ ( ϵ + δ ) − ξ 3 ] , H ′ = ( 1 − ϵ ) H 2 , ϕ 0 ′ = − 2 ϵ M p H , a 2 V = ( 3 − ϵ ) M p 2 H 2 , a 2 V ϕ = 2 ϵ ( 3 − δ ) M p H 2 , a 2 V ϕ ϕ = H 2 [ 3 ( ϵ + δ ) − ξ 3 ] . (24) \begin{aligned}
\epsilon'&=2\mathcal{H}\epsilon(\epsilon-\delta),&\qquad\delta'&=\mathcal{H}[\delta(\epsilon+\delta)-\xi_3],\\
\mathcal{H}'&=(1-\epsilon)\mathcal{H}^2,& \phi_0'&=-\sqrt{2\epsilon}\,M_p\mathcal{H},\\
a^2V&=(3-\epsilon)M_p^2\mathcal{H}^2,& a^2V_\phi&=\sqrt{2\epsilon}(3-\delta)M_p\mathcal{H}^2,\\
a^2V_{\phi\phi}&=\mathcal{H}^2[3(\epsilon+\delta)-\xi_3]. \tag*{(24)}
\end{aligned} ϵ ′ H ′ a 2 V a 2 V ϕϕ = 2 H ϵ ( ϵ − δ ) , = ( 1 − ϵ ) H 2 , = ( 3 − ϵ ) M p 2 H 2 , = H 2 [ 3 ( ϵ + δ ) − ξ 3 ] . δ ′ ϕ 0 ′ a 2 V ϕ = H [ δ ( ϵ + δ ) − ξ 3 ] , = − 2 ϵ M p H , = 2 ϵ ( 3 − δ ) M p H 2 , (24)
Here M p − 2 = 8 π G M_p^{-2}=8\pi G M p − 2 = 8 π G . Equivalently,
a 2 V ϕ ϕ H 2 = ( 3 − δ ) ( ϵ + δ ) + δ ′ H . \frac{a^2V_{\phi\phi}}{\mathcal{H}^2}=(3-\delta)(\epsilon+\delta)+\frac{\delta'}{\mathcal{H}}. H 2 a 2 V ϕϕ = ( 3 − δ ) ( ϵ + δ ) + H δ ′ .
Although ξ 3 \xi_3 ξ 3 is second order in slow roll, we retain it in the exact background relations until the final slow-roll expansion. This is particularly important when comparing with the comoving and uniform-density gauges, where inverse powers of ϵ \epsilon ϵ can promote nominally higher-order contributions into lower-order terms of the 2EMT. The conformal-time relation for H \mathcal{H} H can also be written as
H ′ − H 2 = − ϵ H 2 . (25) \mathcal{H}'-\mathcal{H}^2=-\epsilon\mathcal{H}^2. \tag*{(25)} H ′ − H 2 = − ϵ H 2 . (25)
The complete gauge-fixed 2EMT is assembled using Eq. (24) before the slow-roll truncation is imposed. We decompose the perturbation into Fourier modes using the convention
Ψ ( η , x ) = ∫ d 3 k Ψ k ( η ) e i k ⋅ x , (26) \Psi(\eta,\boldsymbol{x})=\int d^3k\,\Psi_k(\eta)e^{i\boldsymbol{k}\cdot\boldsymbol{x}}, \tag*{(26)} Ψ ( η , x ) = ∫ d 3 k Ψ k ( η ) e i k ⋅ x , (26)
Ψ k ( η ) = 1 ( 2 π ) 3 ∫ d 3 x Ψ ( η , x ) e − i k ⋅ x . (27) \Psi_k(\eta)=\frac{1}{(2\pi)^3}\int d^3x\,\Psi(\eta,\boldsymbol{x})e^{-i\boldsymbol{k}\cdot\boldsymbol{x}}. \tag*{(27)} Ψ k ( η ) = ( 2 π ) 3 1 ∫ d 3 x Ψ ( η , x ) e − i k ⋅ x . (27)
The first relation is the Fourier-mode expansion, whereas the second transforms the configuration-space field to momentum space. For each Fourier mode, the long- and short-wavelength solutions are written as [11 , 12 ]
Ψ k ( η ) ≃ { A 1 ϵ , ( k ≪ H : long-wavelength limit ) , 4 π G ϕ 0 ′ a [ c 1 sin ( k η ) + c 2 cos ( k η ) ] , ( k ≫ H : short-wavelength limit ) . (28) \Psi_k(\eta)\simeq
\begin{cases}
A_1\epsilon, & (k\ll\mathcal{H}:\ \text{long-wavelength limit}),\\
4\pi G\dfrac{\phi_0'}{a}\left[c_1\sin(k\eta)+c_2\cos(k\eta)\right], & (k\gg\mathcal{H}:\ \text{short-wavelength limit}).
\end{cases} \tag*{(28)} Ψ k ( η ) ≃ ⎩ ⎨ ⎧ A 1 ϵ , 4 π G a ϕ 0 ′ [ c 1 sin ( k η ) + c 2 cos ( k η ) ] , ( k ≪ H : long-wavelength limit ) , ( k ≫ H : short-wavelength limit ) . (28)
We use these as the leading asymptotic mode inputs. Explicit k 2 k^2 k 2 , k 4 k^4 k 4 , and higher-gradient coefficients are evaluated within this mode approximation. For short-wavelength products we retain the slowly varying envelope when differentiating, and only then take the fast-phase average,
sin 2 ( k η ) = cos 2 ( k η ) = 1 2 , sin ( k η ) cos ( k η ) = 0. (29) \sin^2(k\eta)=\cos^2(k\eta)=\frac{1}{2},\qquad\sin(k\eta)\cos(k\eta)=0. \tag*{(29)} sin 2 ( k η ) = cos 2 ( k η ) = 2 1 , sin ( k η ) cos ( k η ) = 0. (29)
We introduce independent wavelength parameters in the long- and short-wavelength domains respectively,
σ 1 ≡ k H ≪ 1 , σ 2 ≡ H k ≪ 1. (30) \sigma_1\equiv\frac{k}{\mathcal{H}}\ll1,\qquad\sigma_2\equiv\frac{\mathcal{H}}{k}\ll1. \tag*{(30)} σ 1 ≡ H k ≪ 1 , σ 2 ≡ k H ≪ 1. (30)
The slow-roll and gradient limits must be tracked separately. Throughout this paper we distinguish four asymptotic regimes. The strict IR and strict UV limits are obtained by performing the wavelength expansion first in σ 1 \sigma_1 σ 1 or σ 2 \sigma_2 σ 2 , respectively, at fixed admissible nonzero slow-roll parameters. The formal limit σ i → 0 \sigma_i \to0 σ i → 0 is taken afterward. The intermediate IR and intermediate UV regimes are defined by
ϵ , ∣ δ ∣ ≪ σ i 2 ≪ 1 , (31) \epsilon, \lvert\delta\rvert\ll\sigma_i^2 \ll1, \tag*{(31)} ϵ , ∣ δ ∣ ≪ σ i 2 ≪ 1 , (31)
and are organized by performing the slow-roll expansion first and then according to the wavelength hierarchy. After extracting the displayed overall factor, we retain zeroth- and first-order slow-roll terms in each wavelength coefficient. We denote the coefficientwise slow-roll remainder at fixed σ i \sigma_i σ i by O S R ( 2 ; σ i ) O_{\mathrm{SR}}(2;\sigma_i) O SR ( 2 ; σ i ) .
For the four gauges analyzed previously, the corresponding gauge-fixed Fourier-space 2EMT components contain only a finite set of polynomial or Laurent-polynomial terms in the wavelength parameters after the asymptotic mode solutions are substituted [13 ] . Thus, performing a wavelength or slow-roll expansion merely reorders the terms already present in the 2EMT, and does not generate an additional infinite sequence.
The uniform-expansion gauge is qualitatively different because its exact expressions contain factors of ( k 2 + 3 ϵ H 2 ) − n (k^2+3\epsilon\mathcal{H}^2)^{-n} ( k 2 + 3 ϵ H 2 ) − n , which mix the wavelength scale k 2 k^2 k 2 with the slow-roll geometrical scale ϵ H 2 \epsilon\mathcal{H}^2 ϵ H 2 . A series expansion of these factors generates an infinite sequence of terms, and expanding first in wavelength or first in slow roll gives different series corresponding to different asymptotic hierarchies. The UE series expansion therefore cannot be regarded simply as a finite reordering of terms already present in the original 2EMT.
Definition of the 2EMT
At second perturbative order, the Einstein equation can be arranged as
G μ ν ( 1 ) [ g ( 2 ) ] = 8 π G T μ ν ( 1 ) [ g ( 2 ) , δ ϕ ( 2 ) ] + 8 π G T μ ν ( 2 , e f f ) , (32) G^{(1)}_{\mu\nu}[g^{(2)}] = 8\pi G T^{(1)}_{\mu\nu}[g^{(2)},\delta\phi^{(2)}] + 8\pi G T^{(2,\mathrm{eff})}_{\mu\nu}, \tag*{(32)} G μν ( 1 ) [ g ( 2 ) ] = 8 π G T μν ( 1 ) [ g ( 2 ) , δ ϕ ( 2 ) ] + 8 π G T μν ( 2 , eff ) , (32)
where
T μ ν ( 2 , e f f ) ≡ T μ ν ( 2 ) [ g ( 1 ) , δ ϕ ( 1 ) ] − 1 8 π G G μ ν ( 2 ) [ g ( 1 ) ] . (33) T^{(2,\mathrm{eff})}_{\mu\nu} \equiv T^{(2)}_{\mu\nu}[g^{(1)},\delta\phi^{(1)}] - \frac{1}{8\pi G}G^{(2)}_{\mu\nu}[g^{(1)}]. \tag*{(33)} T μν ( 2 , eff ) ≡ T μν ( 2 ) [ g ( 1 ) , δ ϕ ( 1 ) ] − 8 π G 1 G μν ( 2 ) [ g ( 1 ) ] . (33)
The quadratic effective source is evaluated in Fourier space. We denote the contribution associated with a fixed Fourier pair k k k and − k -k − k by
τ ^ μ ν ( k ) ≡ 8 π G τ μ ν ( k ) = 8 π G ⟨ T μ ν ( 2 , e f f ) ⟩ k . (34) \widehat{\tau}_{\mu\nu}(k) \equiv8\pi G\tau_{\mu\nu}(k) = 8\pi G\left\langle T^{(2,\mathrm{eff})}_{\mu\nu}\right\rangle_k. \tag*{(34)} τ μν ( k ) ≡ 8 π G τ μν ( k ) = 8 π G ⟨ T μν ( 2 , eff ) ⟩ k . (34)
Here, ⟨ ⋯ ⟩ k \langle\cdots\rangle_k ⟨ ⋯ ⟩ k denotes the Fourier-transformed quadratic quantity obtained through the required spatial integration; it does not represent a separate averaging prescription. The integration pairs the modes k k k and − k -k − k . For a real perturbation, Ψ − k = Ψ k ∗ \Psi_{-k}=\Psi_k^* Ψ − k = Ψ k ∗ , and similarly for the other fields. Below we suppress the explicit argument k k k of the fixed-mode contribution.
The fixed Fourier pair is not itself isotropic. For the projected scalar sector we use, following Abramo et al. [6 ] ,
k i k j ⟶ k 2 3 δ i j , τ ^ i i ≡ 1 3 δ i j τ ^ i j . (35) k_i k_j \longrightarrow\frac{k^2}{3}\delta_{ij}, \qquad\widehat{\tau}_{ii} \equiv\frac{1}{3}\delta^{ij}\widehat{\tau}_{ij}. \tag*{(35)} k i k j ⟶ 3 k 2 δ ij , τ ii ≡ 3 1 δ ij τ ij . (35)
Thus τ ^ i i \widehat{\tau}_{ii} τ ii denotes one third of the projected spatial trace, and the off-diagonal spatial components vanish after this projection. The corresponding diagnostic energy density, pressure, and ratio are
ρ r a w = τ 00 a 2 , p r a w = τ i i a 2 , w r a w = p r a w ρ r a w . (36) \rho_{\mathrm{raw}} = \frac{\tau_{00}}{a^2}, \qquad p_{\mathrm{raw}} = \frac{\tau_{ii}}{a^2}, \qquad w_{\mathrm{raw}} = \frac{p_{\mathrm{raw}}}{\rho_{\mathrm{raw}}}. \tag*{(36)} ρ raw = a 2 τ 00 , p raw = a 2 τ ii , w raw = ρ raw p raw . (36)
The label “raw” emphasizes that these gauge-fixed quantities are not, by themselves, relational observables. In particular, w r a w w_{\mathrm{raw}} w raw is introduced only as a diagnostic component ratio and should not be identified with a physical equation-of-state parameter without a common observer and relational prescription.
The general Fourier-space expression for τ ^ μ ν \widehat{\tau}_{\mu\nu} τ μν depends on the gauge-invariant potential Ψ \Psi Ψ and on the gauge variables Q Q Q and E E E [11 , 13 ] . In the longitudinal, spatially-flat, comoving, and uniform-density gauges, the imposed conditions allow the final reduced expressions to be written directly in terms of Ψ \Psi Ψ . The uniform-expansion gauge requires one additional structural step: in configuration space its temporal condition is a differential equation for Q Q Q , while in Fourier space that equation becomes an algebraic relation for each Fourier mode.
We first recall the meaning of gauge transformation and gauge invariance in cosmological perturbation theory. A gauge transformation does not represent a physical change of the perturbed spacetime. Rather, it changes the identification map that associates points of the homogeneous background spacetime with points of the physical spacetime when perturbations are defined. An identification map may be written as
X : M 0 → M , (37) X:\mathcal{M}_0\to\mathcal{M}, \tag*{(37)} X : M 0 → M , (37)
where M 0 \mathcal{M}_0 M 0 denotes the homogeneous background spacetime and M \mathcal{M} M the physical perturbed spacetime. A change of identification map from X X X to Y Y Y generally associates the same point of the background spacetime with different points of the physical spacetime. This change gives rise to the gauge transformation laws of the perturbation variables. The physical spacetime and its fields remain unchanged; what changes is their identification with the background spacetime. For a fixed physical spacetime, the admissible identification maps X , Y , … : M 0 → M X,Y,\ldots:\mathcal{M}_0\to\mathcal{M} X , Y , … : M 0 → M generate gauge-related perturbative configurations, which form a gauge orbit.
At first order, the scalar part of a change of identification map is generated by
ξ μ = ( ξ 0 , ∂ i ζ ) , (38) \xi^\mu=(\xi^0,\partial^i\zeta), \tag*{(38)} ξ μ = ( ξ 0 , ∂ i ζ ) , (38)
and the corresponding transformations of α \alpha α , β \beta β , ψ \psi ψ , E E E , and δ ϕ \delta\phi δ ϕ are given in Eqs. (7) –(11) . In particular, Q Q Q and E E E transform as Q ~ = Q − ξ 0 \widetilde{Q}=Q-\xi^0 Q = Q − ξ 0 and E ~ = E − ζ \widetilde{E}=E-\zeta E = E − ζ , whereas the Bardeen potentials Φ \Phi Φ and Ψ \Psi Ψ remain gauge invariant. This is the standard scalar-perturbation notion of gauge freedom [14 , 15 ] .
Writing
X = { α , β , ψ , E , Q , δ ϕ , … } , (39) \mathcal{X}=\{\alpha,\beta,\psi,E,Q,\delta\phi,\ldots\}, \tag*{(39)} X = { α , β , ψ , E , Q , δ ϕ , … } , (39)
a perturbative functional F [ X ] F[\mathcal{X}] F [ X ] is gauge invariant if it has the same value for arbitrary gauge-related configurations,
F [ X ~ ] = F [ X ] for arbitrary ξ 0 and ζ , (40) F[\widetilde{\mathcal{X}}]=F[\mathcal{X}]
\qquad\text{for arbitrary }\xi^0\text{ and }\zeta, \tag*{(40)} F [ X ] = F [ X ] for arbitrary ξ 0 and ζ , (40)
or equivalently
Δ ξ F ≡ F [ X ~ ] − F [ X ] = 0. (41) \Delta_\xi F\equiv F[\widetilde{\mathcal{X}}]-F[\mathcal{X}]=0. \tag*{(41)} Δ ξ F ≡ F [ X ] − F [ X ] = 0. (41)
Equation (41) is the gauge-invariance criterion before any gauge condition is imposed.
For the quadratic effective source, this criterion must be applied to the complete general 2EMT. The presence of individually gauge-dependent variables such as Q Q Q and E E E does not by itself establish gauge dependence, since their transformed contributions could in principle cancel in the complete quadratic expression. The decisive test is therefore to apply an arbitrary scalar gauge transformation to the general quadratic source before gauge fixing, and evaluate its variation directly.
The gauge status of the quadratic effective source has been a long-standing issue in the backreaction literature. The early 2EMT program formulated the quadratic source with the aim of obtaining a gauge-invariant description [5 , 6 ] , whereas later analyses raised concerns about gauge dependence in the 2EMT framework [7 , 8 ] . Our previous studies evaluated the scalar 2EMT in different gauge choices [10 , 11 , 12 , 13 ] . Here, we perform the direct transformation test before imposing any temporal or spatial gauge condition. To our knowledge, this is the first explicit component-level evaluation of the general scalar-gauge variation of the quadratic source in the present 2EMT framework.
We start from the general scalar quadratic source T μ ν ( 2 , e f f ) T_{\mu\nu}^{(2,\mathrm{eff})} T μν ( 2 , eff ) given in Appendix A of Ref. [13 ] . Defining
S μ ν [ X ] ≡ 8 π G T μ ν ( 2 , e f f ) [ X ] , (42) \mathcal{S}_{\mu\nu}[\mathcal{X}]\equiv8\pi G T_{\mu\nu}^{(2,\mathrm{eff})}[\mathcal{X}], \tag*{(42)} S μν [ X ] ≡ 8 π G T μν ( 2 , eff ) [ X ] , (42)
we consider its variation in configuration space,
Δ ξ S μ ν ≡ S μ ν [ X ~ ] − S μ ν [ X ] . (43) \Delta_\xi\mathcal{S}_{\mu\nu}\equiv\mathcal{S}_{\mu\nu}[\widetilde{\mathcal{X}}]-\mathcal{S}_{\mu\nu}[\mathcal{X}]. \tag*{(43)} Δ ξ S μν ≡ S μν [ X ] − S μν [ X ] . (43)
No gauge condition is imposed, and the transformation parameters ξ 0 \xi^0 ξ 0 and ζ \zeta ζ remain arbitrary.
The explicit results for the 00 00 00 and i i ii ii components, with no sum over i i i , are given in Appendix A. In simplifying these expressions, terms that are odd under reversal of the integration momentum, including the total-spatial-derivative contributions encountered in the calculation, are omitted. The remaining expressions contain nonvanishing even-in-k k k contributions. Since these even and odd sectors cannot cancel identically, we obtain
Δ ξ S 00 ≠ 0 , Δ ξ S i i ≠ 0 in general (no sum over i ) . (44) \Delta_\xi\mathcal{S}_{00}\ne0,\qquad
\Delta_\xi\mathcal{S}_{ii}\ne0
\qquad\text{in general (no sum over }i\text{)}. \tag*{(44)} Δ ξ S 00 = 0 , Δ ξ S ii = 0 in general (no sum over i ) . (44)
Thus, the general scalar 2EMT before gauge fixing does not satisfy the gauge-invariance criterion in Eq. (40) . The general quadratic source itself is therefore gauge dependent under an arbitrary scalar gauge transformation.
The spatial off-diagonal components require a separate technical point. After the terms odd in the integration momentum are removed under symmetric momentum integration, the remaining off-diagonal tensor structures proportional to k i k j k_i k_j k i k j with i ≠ j i \ne j i = j vanish after isotropic angular integration. Consequently, the scalar spatial source obtained in the Fourier calculation is diagonal,
τ ^ i j ( k ) ⟶ δ i j τ ^ s ( k ) . (45) \widehat{\tau}_{ij}(k) \longrightarrow\delta_{ij}\widehat{\tau}_{s}(k). \tag*{(45)} τ ij ( k ) ⟶ δ ij τ s ( k ) . (45)
The same reduction applies to the corresponding gauge-variation terms. These cancellations are part of the Fourier integration used in constructing the quadratic source and do not constitute an additional spatial-averaging prescription.
The discussion above concerns the general quadratic source before gauge fixing. Gauge fixing addresses a different question: it selects an identification map, and hence a perturbative representation, by imposing specified conditions on the perturbation variables. We now turn to the uniform-expansion choice.
To make the geometrical meaning of the gauge explicit, let n μ n^\mu n μ be the future-directed unit normal to the constant-η \eta η hypersurfaces, and let
h μ ν = g μ ν + n μ n ν (46) h_{\mu\nu}=g_{\mu\nu}+n_\mu n_\nu \tag*{(46)} h μν = g μν + n μ n ν (46)
be the induced metric on these hypersurfaces.
We adopt the following sign convention for the extrinsic curvature:
K μ ν = − h μ ρ h ν σ ∇ ρ n σ . (47) K_{\mu\nu}=-h_\mu^{\ \rho}h_\nu^{\ \sigma}\nabla_\rho n_\sigma. \tag*{(47)} K μν = − h μ ρ h ν σ ∇ ρ n σ . (47)
Then the expansion scalar of the normal congruence is
Θ ≡ ∇ μ n μ = − K , K ≡ K μ μ . (48) \Theta\equiv\nabla_\mu n^\mu=-\mathcal{K},\qquad\mathcal{K}\equiv K^\mu_{\ \mu}. \tag*{(48)} Θ ≡ ∇ μ n μ = − K , K ≡ K μ μ . (48)
For the scalar metric of Eq. (3) , the linear spatial extrinsic curvature is
K j i = − H a δ j i + 1 a [ ( ψ ′ + H α ) δ j i − Q , j , i ] , (49) \mathcal{K}^{i}_{\ j}=-\frac{\mathcal{H}}{a}\delta^{i}_{\ j}+\frac{1}{a}\left[(\psi^\prime+\mathcal{H}\alpha)\delta^{i}_{\ j}-Q^{,i}_{\ ,j}\right], \tag*{(49)} K j i = − a H δ j i + a 1 [ ( ψ ′ + H α ) δ j i − Q , j , i ] , (49)
and its trace is conventionally written as
K = − 3 H a + κ a , κ ≡ 3 ( ψ ′ + H α ) − Δ Q . (50) \mathcal{K}=-\frac{3\mathcal{H}}{a}+\frac{\kappa}{a},\qquad\kappa\equiv3(\psi^\prime+\mathcal{H}\alpha)-\Delta Q. \tag*{(50)} K = − a 3 H + a κ , κ ≡ 3 ( ψ ′ + H α ) − Δ Q . (50)
Equivalently, at first perturbative order,
Θ = 3 H a − κ a , δ Θ ( 1 ) = − κ a . (51) \Theta=\frac{3\mathcal{H}}{a}-\frac{\kappa}{a},\qquad\delta\Theta^{(1)}=-\frac{\kappa}{a}. \tag*{(51)} Θ = a 3 H − a κ , δ Θ ( 1 ) = − a κ . (51)
The temporal uniform-expansion condition requires the first-order perturbation of the expansion scalar to vanish, while the remaining scalar spatial gauge freedom is fixed independently by E = 0 E=0 E = 0 . The uniform-expansion gauge is therefore
δ Θ ( 1 ) = 0 ( κ = 0 ) , E = 0. (52) \delta\Theta^{(1)}=0\quad(\kappa=0),\qquad E=0. \tag*{(52)} δ Θ ( 1 ) = 0 ( κ = 0 ) , E = 0. (52)
Cosmological backreaction concerns how inhomogeneous perturbations contribute to the effective evolution of a homogeneous background, particularly its expansion. The uniform-expansion gauge provides a geometrically defined temporal slicing on which the first-order perturbation of the expansion scalar of the hypersurface-normal congruence vanishes. This gauge condition does not eliminate the perturbations themselves or their quadratic contributions. Rather, it specifies the hypersurfaces on which the 2EMT is evaluated, allowing us to investigate how the quadratic effective source depends on a slicing defined by uniform local expansion.
At first order, the uniform-expansion condition δ Θ ( 1 ) = 0 \delta\Theta^{(1)}=0 δ Θ ( 1 ) = 0 (κ = 0 \kappa=0 κ = 0 ) corresponds to constant-mean-curvature (CMC) slicing, since the trace of the extrinsic curvature is spatially uniform on each constant-η \eta η hypersurface, although it may evolve in time. The variable κ \kappa κ and the extrinsic-curvature trace are standard in gauge-ready formulations of cosmological perturbation theory [4 , 14 , 15 ] .
From the definitions of Φ \Phi Φ and Ψ \Psi Ψ , together with Φ = Ψ \Phi=\Psi Φ = Ψ , we have
ψ = Ψ − H Q , α = Ψ + Q ′ + H Q . (53) \psi=\Psi-\mathcal{H}Q,\qquad\alpha=\Psi+Q^{\prime}+\mathcal{H}Q. \tag*{(53)} ψ = Ψ − H Q , α = Ψ + Q ′ + H Q . (53)
Then Eq. (50) becomes
κ = 3 ( Ψ ′ + H Ψ ) + [ 3 ( H 2 − H ′ ) − Δ ] Q . (54) \kappa=3(\Psi^{\prime}+\mathcal{H}\Psi)+\left[3(\mathcal{H}^{2}-\mathcal{H}^{\prime})-\Delta\right]Q. \tag*{(54)} κ = 3 ( Ψ ′ + H Ψ ) + [ 3 ( H 2 − H ′ ) − Δ ] Q . (54)
The condition κ = 0 \kappa=0 κ = 0 therefore gives
( Δ − 3 ϵ H 2 ) Q U E = 3 ( Ψ ′ + H Ψ ) , (55) (\Delta-3\epsilon\mathcal{H}^{2})Q_{\mathrm{UE}}=3(\Psi^{\prime}+\mathcal{H}\Psi), \tag*{(55)} ( Δ − 3 ϵ H 2 ) Q UE = 3 ( Ψ ′ + H Ψ ) , (55)
where Eq. (25) has been used.
(55) is the first structural distinction from the gauges considered in the companion analysis. In configuration space, it is a differential equation for Q Q Q , rather than an algebraic relation. Consequently, an unreduced derivative representation of the 2EMT can retain Q Q Q , Q ′ Q^{\prime} Q ′ , and Q ′ ′ Q^{\prime\prime} Q ′′ ; the explicit expressions are collected in Appendix B.
For a Fourier mode, Δ → − k 2 \Delta\to-k^{2} Δ → − k 2 , and Eq. (55) becomes
Q U E ( k ) = − 3 ( Ψ ′ + H Ψ ) k 2 + 3 ϵ H 2 . (56) Q_{\mathrm{UE}}(k)=-\frac{3(\Psi^{\prime}+\mathcal{H}\Psi)}{k^{2}+3\epsilon\mathcal{H}^{2}}. \tag*{(56)} Q UE ( k ) = − k 2 + 3 ϵ H 2 3 ( Ψ ′ + H Ψ ) . (56)
The remaining perturbations are therefore
E U E = 0 , β U E = Q U E , (57) E_{\mathrm{UE}}=0,\qquad\beta_{\mathrm{UE}}=Q_{\mathrm{UE}}, \tag*{(57)} E UE = 0 , β UE = Q UE , (57)
ψ U E = Ψ − H Q U E , (58) \psi_{\mathrm{UE}}=\Psi-\mathcal{H}Q_{\mathrm{UE}}, \tag*{(58)} ψ UE = Ψ − H Q UE , (58)
α U E = Ψ + Q U E ′ + H Q U E , (59) \alpha_{\mathrm{UE}}=\Psi+Q_{\mathrm{UE}}^{\prime}+\mathcal{H}Q_{\mathrm{UE}}, \tag*{(59)} α UE = Ψ + Q UE ′ + H Q UE , (59)
δ ϕ U E = δ ϕ g i + ϕ 0 ′ Q U E . (60) \delta\phi_{\mathrm{UE}}=\delta\phi_{\mathrm{gi}}+\phi_{0}^{\prime}Q_{\mathrm{UE}}. \tag*{(60)} δ ϕ UE = δ ϕ gi + ϕ 0 ′ Q UE . (60)
Equation (56) immediately exposes the scale competition in Eq. (1) . It also gives useful limiting checks. Since
Q C = Ψ ′ + H Ψ H ′ − H 2 = − Ψ ′ + H Ψ ϵ H 2 , (61) Q_{\mathrm{C}}=\frac{\Psi^{\prime}+\mathcal{H}\Psi}{\mathcal{H}^{\prime}-\mathcal{H}^{2}}=-\frac{\Psi^{\prime}+\mathcal{H}\Psi}{\epsilon\mathcal{H}^{2}}, \tag*{(61)} Q C = H ′ − H 2 Ψ ′ + H Ψ = − ϵ H 2 Ψ ′ + H Ψ , (61)
we have
Q U E ⟶ Q C ( k 2 ≪ ϵ H 2 ) . (62) Q_{\mathrm{UE}}\longrightarrow Q_{\mathrm{C}}\qquad(k^{2}\ll\epsilon\mathcal{H}^{2}). \tag*{(62)} Q UE ⟶ Q C ( k 2 ≪ ϵ H 2 ) . (62)
For k 2 ≫ ϵ H 2 k^{2}\gg\epsilon\mathcal{H}^{2} k 2 ≫ ϵ H 2 , by contrast, Q U E = O ( k − 2 ) Q_{\mathrm{UE}}=O(k^{-2}) Q UE = O ( k − 2 ) , and the shift becomes gradient suppressed, which anticipates the approach to the spatially-flat result.
Complete gauge fixing and the gauge-fixed 2EMT
Gauge fixing selects, within a given gauge orbit, an identification map for which the specified gauge conditions are satisfied. Starting from an arbitrary identification map, the gauge conditions determine the transformation parameters required to reach the selected representative. A complete gauge fixing removes any residual gauge freedom and therefore makes this representative unique, subject to the appropriate boundary or regularity conditions.
For the uniform-expansion gauge, the transformation laws
κ ~ = κ + ( Δ − 3 ϵ H 2 ) ξ 0 , E ~ = E − ζ (63) \widetilde{\kappa}=\kappa+(\Delta-3\epsilon\mathcal{H}^{2})\xi^{0},\qquad\widetilde{E}=E-\zeta \tag*{(63)} κ = κ + ( Δ − 3 ϵ H 2 ) ξ 0 , E = E − ζ (63)
show directly that the fixing is complete. If both the original and transformed configurations satisfy the same gauge conditions, κ = κ ~ = 0 \kappa=\widetilde{\kappa}=0 κ = κ = 0 and E = E ~ = 0 E=\widetilde{E}=0 E = E = 0 , then ζ = 0 \zeta=0 ζ = 0 and
( Δ − 3 ϵ H 2 ) ξ 0 = 0. (64) (\Delta-3\epsilon\mathcal{H}^{2})\xi^{0}=0. \tag*{(64)} ( Δ − 3 ϵ H 2 ) ξ 0 = 0. (64)
With the boundary or regularity conditions that remove the homogeneous solution, ξ 0 = 0 \xi^{0}=0 ξ 0 = 0 , and no nontrivial residual scalar gauge transformation remains.
Once the complete gauge has been fixed, the perturbation variables and the corresponding 2EMT are evaluated within this selected representation. The absence of residual gauge freedom means that the gauge-fixed result is unambiguously specified for a given perturbative solution.
In particular, the variables appearing in the final gauge-fixed expression do not determine its gauge status. For example, in the spatially-flat gauge ψ S F = 0 \psi_{\mathrm{SF}}=0 ψ SF = 0 , and hence Ψ = H Q S F \Psi=\mathcal{H}Q_{\mathrm{SF}} Ψ = H Q SF . A gauge-fixed 2EMT written in terms of Ψ \Psi Ψ may therefore equally be written in terms of Q S F Q_{\mathrm{SF}} Q SF . The two expressions represent the same quantity in the same fixed gauge. Thus, the appearance of a gauge-invariant variable such as Ψ \Psi Ψ , or of a gauge variable such as Q Q Q , does not by itself implement gauge invariance or imply gauge dependence of the gauge-fixed 2EMT.
In the following sections, the uniform-expansion conditions are maintained throughout the evaluation and interpretation of the corresponding quadratic source.
Uniform-expansion 2EMT in the long-wavelength domain
We substitute the UE constraint and the long-wavelength mode into the general scalar 2EMT. We use σ 1 = k / H ≪ 1 \sigma_1 = k/\mathcal{H} \ll1 σ 1 = k / H ≪ 1 . The exact result is rational in k 2 + 3 ϵ H 2 k^2 + 3\epsilon\mathcal{H}^2 k 2 + 3 ϵ H 2 . The second representation below is the wavelength-first expansion for the strict infrared. The third is the slow-roll-first expansion for the intermediate infrared.2
For comparison with the other gauges, we define
F L W ≡ H 2 ∣ A 1 ∣ 2 . (65) \mathcal{F}_{\mathrm{LW}} \equiv\mathcal{H}^2 |A_1|^2 . \tag*{(65)} F LW ≡ H 2 ∣ A 1 ∣ 2 . (65)
τ ^ 00 L W = k 2 ϵ ∣ A 1 ∣ 2 H 2 ( k 2 + 3 ϵ H 2 ) 4 { k 10 + H 2 k 8 [ 1 − 2 δ + ϵ ( 29 − 4 δ ) + 4 ϵ 2 ] + H 4 k 6 [ δ ( 1 − 2 δ ) 2 ( δ + 3 ) + 24 ϵ 3 + ϵ 2 ( 4 δ 2 + 40 δ + 250 ) − 2 ϵ ( 4 δ 3 + 36 δ 2 − δ + 39 ) ] + 6 H 6 ϵ k 4 [ 12 δ 4 − 8 δ 3 + 107 δ 2 − 95 δ + 21 + 3 ϵ 2 ( 4 δ 2 + 26 δ + 61 ) − ϵ ( 24 δ 3 + 70 δ 2 + 135 δ + 44 ) ] − 9 H 8 ϵ k 2 [ 24 ϵ 4 − 3 ϵ 3 ( 12 δ 2 + 32 δ + 111 ) + 6 ϵ 2 ( 12 δ 3 + 8 δ 2 + 95 δ − 9 ) − ϵ ( 36 δ 4 + 24 δ 3 + 349 δ 2 − 361 δ + 88 ) + 12 ( 1 − 2 δ ) 2 ( δ 2 − 1 ) ] − 81 H 10 ϵ 2 [ 4 ϵ 4 − ϵ 3 ( 8 δ + 41 ) + ϵ 2 ( 4 δ 2 + 108 δ − 31 ) − 2 ϵ ( 52 δ 2 − 56 δ + 15 ) + 8 ( 1 − 2 δ ) 2 ( δ − 1 ) ] − H 4 k 2 ϵ 3 ( k 2 + 3 ϵ H 2 ) 2 ( 1 + 2 ϵ − 2 δ ) 2 } \widehat{\tau}^{\mathrm{LW}}_{00} =
\frac{k^2 \epsilon|A_1|^2}{\mathcal{H}^2 (k^2 + 3\epsilon\mathcal{H}^2)^4}
\Bigl\{
k^{10}+\mathcal{H}^2 k^8[1-2\delta+\epsilon(29-4\delta)+4\epsilon^2]
+\mathcal{H}^4 k^6[\delta(1-2\delta)^2(\delta+3)+24\epsilon^3
+\epsilon^2(4\delta^2+40\delta+250)-2\epsilon(4\delta^3+36\delta^2-\delta+39)]
+6\mathcal{H}^6\epsilon k^4[12\delta^4-8\delta^3+107\delta^2-95\delta+21
+3\epsilon^2(4\delta^2+26\delta+61)-\epsilon(24\delta^3+70\delta^2+135\delta+44)]
-9\mathcal{H}^8\epsilon k^2[24\epsilon^4-3\epsilon^3(12\delta^2+32\delta+111)
+6\epsilon^2(12\delta^3+8\delta^2+95\delta-9)-\epsilon(36\delta^4+24\delta^3+349\delta^2-361\delta+88)+12(1-2\delta)^2(\delta^2-1)]
-81\mathcal{H}^{10}\epsilon^2[4\epsilon^4-\epsilon^3(8\delta+41)+\epsilon^2(4\delta^2+108\delta-31)-2\epsilon(52\delta^2-56\delta+15)+8(1-2\delta)^2(\delta-1)]
-\mathcal{H}^4 k^2\epsilon_3(k^2+3\epsilon\mathcal{H}^2)^2(1+2\epsilon-2\delta)^2
\Bigr\} τ 00 LW = H 2 ( k 2 + 3 ϵ H 2 ) 4 k 2 ϵ ∣ A 1 ∣ 2 { k 10 + H 2 k 8 [ 1 − 2 δ + ϵ ( 29 − 4 δ ) + 4 ϵ 2 ] + H 4 k 6 [ δ ( 1 − 2 δ ) 2 ( δ + 3 ) + 24 ϵ 3 + ϵ 2 ( 4 δ 2 + 40 δ + 250 ) − 2 ϵ ( 4 δ 3 + 36 δ 2 − δ + 39 )] + 6 H 6 ϵ k 4 [ 12 δ 4 − 8 δ 3 + 107 δ 2 − 95 δ + 21 + 3 ϵ 2 ( 4 δ 2 + 26 δ + 61 ) − ϵ ( 24 δ 3 + 70 δ 2 + 135 δ + 44 )] − 9 H 8 ϵ k 2 [ 24 ϵ 4 − 3 ϵ 3 ( 12 δ 2 + 32 δ + 111 ) + 6 ϵ 2 ( 12 δ 3 + 8 δ 2 + 95 δ − 9 ) − ϵ ( 36 δ 4 + 24 δ 3 + 349 δ 2 − 361 δ + 88 ) + 12 ( 1 − 2 δ ) 2 ( δ 2 − 1 )] − 81 H 10 ϵ 2 [ 4 ϵ 4 − ϵ 3 ( 8 δ + 41 ) + ϵ 2 ( 4 δ 2 + 108 δ − 31 ) − 2 ϵ ( 52 δ 2 − 56 δ + 15 ) + 8 ( 1 − 2 δ ) 2 ( δ − 1 )] − H 4 k 2 ϵ 3 ( k 2 + 3 ϵ H 2 ) 2 ( 1 + 2 ϵ − 2 δ ) 2 }
≃ F L W { σ 1 2 ϵ [ 8 + ( − 40 δ + 30 ϵ ) + ( 64 δ 2 − 112 δ ϵ + 31 ϵ 2 ) + ( − 32 δ 3 + 104 δ 2 ϵ − 108 δ ϵ 2 + 41 ϵ 3 ) + ( − 48 δ 4 + 24 δ 3 ϵ + 4 δ ϵ 3 + 24 ϵ 4 − 4 ϵ 2 ϵ 3 ) ] + σ 1 4 9 ϵ 2 [ − 84 + ( 432 δ − 272 ϵ ) + ( − 732 δ 2 + 983 δ ϵ − 318 ϵ 2 ) + ( 432 δ 3 − 899 δ 2 ϵ + 726 δ ϵ 2 − 159 ϵ 3 − ϵ ϵ 3 ) + ( − 48 δ 4 + 24 δ 3 ϵ + 4 δ ϵ 3 + 24 ϵ 4 − 4 ϵ 2 ϵ 3 ) + ( 36 δ 4 ϵ − 72 δ 3 ϵ 2 + 36 δ 2 ϵ 3 − 4 δ 2 ϵ ϵ 3 + 8 δ ϵ 2 ϵ 3 − 4 ϵ 3 ϵ 3 ) ] + σ 1 6 27 ϵ 3 [ 192 + ( − 1008 δ + 590 ϵ ) + ( 1776 δ 2 − 2106 δ ϵ + 626 ϵ 2 ) + ( − 1152 δ 3 + 1938 δ 2 ϵ − 1230 δ ϵ 2 + 264 ϵ 3 + 2 ϵ ϵ 3 ) + ( 192 δ 4 − 112 δ 3 ϵ − 68 δ 2 ϵ 2 + 12 δ ϵ 3 − 8 δ ϵ ϵ 3 − 24 ϵ 4 + 8 ϵ 2 ϵ 3 ) + ( − 120 δ 4 ϵ + 240 δ 3 ϵ 2 − 120 δ 2 ϵ 3 + 8 δ 2 ϵ ϵ 3 − 16 δ ϵ 2 ϵ 3 + 8 ϵ 3 ϵ 3 ) ] + σ 1 8 81 ϵ 4 [ − 360 + ( 1920 δ − 1088 ϵ ) + ( − 3480 δ 2 + 3873 δ ϵ − 1046 ϵ 2 ) + ( 2400 δ 3 − 3617 δ 2 ϵ + 1862 δ ϵ 2 − 344 ϵ 3 − 3 ϵ ϵ 3 ) + ( − 480 δ 4 + 312 δ 3 ϵ + 248 δ 2 ϵ 2 − 104 δ ϵ 3 + 12 δ ϵ ϵ 3 + 24 ϵ 4 − 12 ϵ 2 ϵ 3 ) + ( 268 δ 4 ϵ − 536 δ 3 ϵ 2 + 268 δ 2 ϵ 3 − 12 δ 2 ϵ ϵ 3 + 24 δ ϵ 2 ϵ 3 − 12 ϵ 3 ϵ 3 ) ] + σ 1 10 243 ϵ 5 [ 600 + ( − 3240 δ + 1810 ϵ ) + ( 6000 δ 2 − 6452 δ ϵ + 1610 ϵ 2 ) + ( − 4320 δ 3 + 6124 δ 2 ϵ − 2654 δ ϵ 2 + 392 ϵ 3 + 4 ϵ ϵ 3 ) + ( 960 δ 4 − 672 δ 3 ϵ − 572 δ 2 ϵ 2 + 308 δ ϵ 3 − 16 δ ϵ ϵ 3 − 24 ϵ 4 + 16 ϵ 2 ϵ 3 ) + ( − 496 δ 4 ϵ + 992 δ 3 ϵ 2 − 496 δ 2 ϵ 3 + 16 δ 2 ϵ ϵ 3 − 32 δ ϵ 2 ϵ 3 + 16 ϵ 3 ϵ 3 ) ] } . \begin{aligned}
\simeq\mathcal{F}_{\mathrm{LW}}\Biggl\{&
\frac{\sigma_1^2}{\epsilon}\Bigl[8+(-40\delta+30\epsilon)+(64\delta^2-112\delta\epsilon+31\epsilon^2)\\
&\quad+(-32\delta^3+104\delta^2\epsilon-108\delta\epsilon^2+41\epsilon^3)+(-48\delta^4+24\delta^3\epsilon+4\delta\epsilon_3+24\epsilon^4-4\epsilon^2\epsilon_3)\Bigr]\\
&+\frac{\sigma_1^4}{9\epsilon^2}\Bigl[-84+(432\delta-272\epsilon)+(-732\delta^2+983\delta\epsilon-318\epsilon^2)\\
&\quad+(432\delta^3-899\delta^2\epsilon+726\delta\epsilon^2-159\epsilon^3-\epsilon\epsilon_3)\\
&\quad+(-48\delta^4+24\delta^3\epsilon+4\delta\epsilon_3+24\epsilon^4-4\epsilon^2\epsilon_3)\\
&\quad+(36\delta^4\epsilon-72\delta^3\epsilon^2+36\delta^2\epsilon^3-4\delta^2\epsilon\epsilon_3+8\delta\epsilon^2\epsilon_3-4\epsilon^3\epsilon_3)\Bigr]\\
&+\frac{\sigma_1^6}{27\epsilon^3}\Bigl[192+(-1008\delta+590\epsilon)+(1776\delta^2-2106\delta\epsilon+626\epsilon^2)\\
&\quad+(-1152\delta^3+1938\delta^2\epsilon-1230\delta\epsilon^2+264\epsilon^3+2\epsilon\epsilon_3)\\
&\quad+(192\delta^4-112\delta^3\epsilon-68\delta^2\epsilon^2+12\delta\epsilon^3-8\delta\epsilon\epsilon_3-24\epsilon^4+8\epsilon^2\epsilon_3)\\
&\quad+(-120\delta^4\epsilon+240\delta^3\epsilon^2-120\delta^2\epsilon^3+8\delta^2\epsilon\epsilon_3-16\delta\epsilon^2\epsilon_3+8\epsilon^3\epsilon_3)\Bigr]\\
&+\frac{\sigma_1^8}{81\epsilon^4}\Bigl[-360+(1920\delta-1088\epsilon)+(-3480\delta^2+3873\delta\epsilon-1046\epsilon^2)\\
&\quad+(2400\delta^3-3617\delta^2\epsilon+1862\delta\epsilon^2-344\epsilon^3-3\epsilon\epsilon_3)\\
&\quad+(-480\delta^4+312\delta^3\epsilon+248\delta^2\epsilon^2-104\delta\epsilon^3+12\delta\epsilon\epsilon_3+24\epsilon^4-12\epsilon^2\epsilon_3)\\
&\quad+(268\delta^4\epsilon-536\delta^3\epsilon^2+268\delta^2\epsilon^3-12\delta^2\epsilon\epsilon_3+24\delta\epsilon^2\epsilon_3-12\epsilon^3\epsilon_3)\Bigr]\\
&+\frac{\sigma_1^{10}}{243\epsilon^5}\Bigl[600+(-3240\delta+1810\epsilon)+(6000\delta^2-6452\delta\epsilon+1610\epsilon^2)\\
&\quad+(-4320\delta^3+6124\delta^2\epsilon-2654\delta\epsilon^2+392\epsilon^3+4\epsilon\epsilon_3)\\
&\quad+(960\delta^4-672\delta^3\epsilon-572\delta^2\epsilon^2+308\delta\epsilon^3-16\delta\epsilon\epsilon_3-24\epsilon^4+16\epsilon^2\epsilon_3)\\
&\quad+(-496\delta^4\epsilon+992\delta^3\epsilon^2-496\delta^2\epsilon^3+16\delta^2\epsilon\epsilon_3-32\delta\epsilon^2\epsilon_3+16\epsilon^3\epsilon_3)\Bigr]\Biggr\}.
\end{aligned} ≃ F LW { ϵ σ 1 2 [ 8 + ( − 40 δ + 30 ϵ ) + ( 64 δ 2 − 112 δ ϵ + 31 ϵ 2 ) + ( − 32 δ 3 + 104 δ 2 ϵ − 108 δ ϵ 2 + 41 ϵ 3 ) + ( − 48 δ 4 + 24 δ 3 ϵ + 4 δ ϵ 3 + 24 ϵ 4 − 4 ϵ 2 ϵ 3 ) ] + 9 ϵ 2 σ 1 4 [ − 84 + ( 432 δ − 272 ϵ ) + ( − 732 δ 2 + 983 δ ϵ − 318 ϵ 2 ) + ( 432 δ 3 − 899 δ 2 ϵ + 726 δ ϵ 2 − 159 ϵ 3 − ϵ ϵ 3 ) + ( − 48 δ 4 + 24 δ 3 ϵ + 4 δ ϵ 3 + 24 ϵ 4 − 4 ϵ 2 ϵ 3 ) + ( 36 δ 4 ϵ − 72 δ 3 ϵ 2 + 36 δ 2 ϵ 3 − 4 δ 2 ϵ ϵ 3 + 8 δ ϵ 2 ϵ 3 − 4 ϵ 3 ϵ 3 ) ] + 27 ϵ 3 σ 1 6 [ 192 + ( − 1008 δ + 590 ϵ ) + ( 1776 δ 2 − 2106 δ ϵ + 626 ϵ 2 ) + ( − 1152 δ 3 + 1938 δ 2 ϵ − 1230 δ ϵ 2 + 264 ϵ 3 + 2 ϵ ϵ 3 ) + ( 192 δ 4 − 112 δ 3 ϵ − 68 δ 2 ϵ 2 + 12 δ ϵ 3 − 8 δ ϵ ϵ 3 − 24 ϵ 4 + 8 ϵ 2 ϵ 3 ) + ( − 120 δ 4 ϵ + 240 δ 3 ϵ 2 − 120 δ 2 ϵ 3 + 8 δ 2 ϵ ϵ 3 − 16 δ ϵ 2 ϵ 3 + 8 ϵ 3 ϵ 3 ) ] + 81 ϵ 4 σ 1 8 [ − 360 + ( 1920 δ − 1088 ϵ ) + ( − 3480 δ 2 + 3873 δ ϵ − 1046 ϵ 2 ) + ( 2400 δ 3 − 3617 δ 2 ϵ + 1862 δ ϵ 2 − 344 ϵ 3 − 3 ϵ ϵ 3 ) + ( − 480 δ 4 + 312 δ 3 ϵ + 248 δ 2 ϵ 2 − 104 δ ϵ 3 + 12 δ ϵ ϵ 3 + 24 ϵ 4 − 12 ϵ 2 ϵ 3 ) + ( 268 δ 4 ϵ − 536 δ 3 ϵ 2 + 268 δ 2 ϵ 3 − 12 δ 2 ϵ ϵ 3 + 24 δ ϵ 2 ϵ 3 − 12 ϵ 3 ϵ 3 ) ] + 243 ϵ 5 σ 1 10 [ 600 + ( − 3240 δ + 1810 ϵ ) + ( 6000 δ 2 − 6452 δ ϵ + 1610 ϵ 2 ) + ( − 4320 δ 3 + 6124 δ 2 ϵ − 2654 δ ϵ 2 + 392 ϵ 3 + 4 ϵ ϵ 3 ) + ( 960 δ 4 − 672 δ 3 ϵ − 572 δ 2 ϵ 2 + 308 δ ϵ 3 − 16 δ ϵ ϵ 3 − 24 ϵ 4 + 16 ϵ 2 ϵ 3 ) + ( − 496 δ 4 ϵ + 992 δ 3 ϵ 2 − 496 δ 2 ϵ 3 + 16 δ 2 ϵ ϵ 3 − 32 δ ϵ 2 ϵ 3 + 16 ϵ 3 ϵ 3 ) ] } .
+ O ( σ 1 12 ) (strict IR) ≃ ϵ F L W { [ σ 1 2 + σ 1 4 ] + [ 3 δ − 9 ϵ + 126 ϵ σ 1 2 + 108 ϵ σ 1 4 + ( − 2 δ + 17 ϵ ) σ 1 2 ] + O S R ( 2 ; σ 1 ) } (intermediate IR) . (66) \begin{aligned}
+ \mathcal{O}(\sigma_1^{12})
\qquad\text{(strict IR)} \\
&\simeq\epsilon\mathcal{F}_{\mathrm{LW}}\left\{\left[\sigma_1^2+\sigma_1^4\right]+\left[3\delta-9\epsilon+\frac{126\epsilon}{\sigma_1^2}+\frac{108\epsilon}{\sigma_1^4}+(-2\delta+17\epsilon)\sigma_1^2\right]+\mathcal{O}_{\mathrm{SR}}(2;\sigma_1)\right\}
\qquad\text{(intermediate IR)}. \tag*{(66)}
\end{aligned} + O ( σ 1 12 ) (strict IR) ≃ ϵ F LW { [ σ 1 2 + σ 1 4 ] + [ 3 δ − 9 ϵ + σ 1 2 126 ϵ + σ 1 4 108 ϵ + ( − 2 δ + 17 ϵ ) σ 1 2 ] + O SR ( 2 ; σ 1 ) } (intermediate IR) . (66)
T ^ i i L W = k 2 ϵ ∣ A 1 ∣ 2 3 H 2 ( k 2 + 3 ϵ H 2 ) 5 { 3 k 12 − H 2 k 10 [ 1 − 10 δ + 16 δ 2 + 4 ϵ 2 − 2 ϵ ( 10 δ + 23 ) ] − 3 H 4 k 8 [ − δ ( 1 − 2 δ ) 2 ( δ − 3 ) + 28 ϵ 3 − ϵ 2 ( 4 δ 2 + 84 δ + 239 ) + ϵ ( 8 δ 3 + 40 δ 2 + 228 δ − 145 ) ] − 9 H 6 ϵ k 6 [ 4 δ 4 − 272 δ 3 + 597 δ 2 − 323 δ + 46 + 48 ϵ 3 + 4 ϵ 2 ( δ 2 − 67 δ − 111 ) − 4 ϵ ( 2 δ 3 − 123 δ 2 + 13 δ + 44 ) ] − 27 H 8 ϵ k 4 [ 16 ϵ 4 + ϵ 3 ( 36 δ 2 − 356 δ − 367 ) − 4 ϵ 2 ( 10 δ 3 − 141 δ 2 + 94 δ + 20 ) + ϵ ( − 28 δ 4 − 80 δ 3 + 629 δ 2 − 447 δ + 78 ) + 4 ( 1 − 2 δ ) 2 ( 2 δ 3 − 7 δ 2 − 4 δ + 7 ) ] + 81 H 10 ϵ k 2 [ 20 ϵ 4 + ϵ 3 ( − 28 δ 2 + 80 δ + 230 ) + ϵ 2 ( 152 δ 3 − 244 δ 2 + 274 δ − 39 ) + ϵ ( − 220 δ 4 + 416 δ 3 − 915 δ 2 + 745 δ − 182 ) + 4 ( 1 − 2 δ ) 2 ( 6 δ 3 − 11 δ 2 + 26 δ − 17 ) ] + 243 H 12 ϵ 3 [ 12 ϵ 4 + ϵ 3 ( 105 − 72 δ ) + ϵ 2 ( 236 δ 2 − 196 δ + 39 ) − 2 ϵ ( 136 δ 3 − 128 δ 2 + 8 δ + 11 ) + 8 ( 1 − 2 δ ) 2 ( 3 δ 2 − δ − 1 ) ] + 3 H 4 ξ 3 ( k 2 + 3 ϵ H 2 ) [ − 216 H 6 ϵ 2 { 6 ϵ 2 + ϵ ( 5 − 10 δ ) + ( 1 − 2 δ ) 2 } ] + 9 H 4 ϵ k 2 { − 8 δ ( 1 − 2 δ ) 2 + 4 ϵ 3 − 20 ϵ 2 ( 2 δ + 3 ) + ϵ ( 68 δ 2 + 28 δ − 31 ) } + 6 H 2 ϵ k 4 { 4 δ 2 + 20 δ + 4 ϵ 2 − 4 ϵ ( 2 δ + 5 ) − 11 } + k 6 ( 1 + 2 ϵ − 2 δ ) 2 } ≃ F L W { σ 1 2 3 ϵ [ − 8 + ( 24 δ − 22 ϵ ) + ( 24 δ 2 − 16 δ ϵ + 39 ϵ 2 − 8 ξ 3 ) + ( − 128 δ 3 + 256 δ 2 ϵ − 196 δ ϵ 2 + 32 δ ξ 3 + 105 ϵ 3 − 40 ϵ ξ 3 ) + ( 96 δ 4 − 272 δ 3 ϵ + 236 δ 2 ϵ 2 − 32 δ 2 ξ 3 − 72 δ ϵ 3 + 80 δ ϵ ξ 3 + 12 ϵ 4 − 48 ϵ 2 ξ 3 ) ] + σ 1 4 9 ϵ 2 [ − 28 + ( 256 δ − 72 ϵ ) + ( − 852 δ 2 + 825 δ ϵ − 234 ϵ 2 + 32 ξ 3 ) + ( 1256 δ 3 − 2195 δ 2 ϵ + 1254 δ ϵ 2 − 136 δ ξ 3 − 295 ϵ 3 + 129 ϵ ξ 3 ) + ( − 752 δ 4 + 1776 δ 3 ϵ − 1424 δ 2 ϵ 2 + 160 δ 2 ξ 3 + 440 δ ϵ 3 − 292 δ ϵ ξ 3 − 40 ϵ 4 + 132 ϵ 2 ξ 3 ) + ( 960 δ 5 − 2208 δ 4 ϵ + 1528 δ 3 ϵ 2 − 320 δ 3 ξ 3 − 280 δ 2 ϵ 2 + 68 δ 2 ϵ ξ 3 − 40 δ ϵ 2 ξ 3 + 4 ϵ 3 ξ 3 ) ] + σ 1 6 27 ϵ 3 [ 192 + ( − 1392 δ + 502 ϵ ) + ( 3872 δ 2 − 3518 δ ϵ + 860 ϵ 2 − 80 ξ 3 ) + ( − 5056 δ 3 + 7786 δ 2 ϵ − 3934 δ ϵ 2 + 352 δ ξ 3 + 792 ϵ 3 − 298 ϵ ξ 3 ) + ( 2944 δ 4 − 6080 δ 3 ϵ + 4196 δ 2 ϵ 2 − 448 δ 2 ξ 3 − 1124 δ ϵ 3 + 728 δ ϵ ξ 3 + 64 ϵ 4 − 280 ϵ 2 ξ 3 ) + ( − 512 δ 5 + 1128 δ 4 ϵ − 720 δ 3 ϵ 2 + 128 δ 3 ξ 3 + 104 δ 2 ϵ 3 − 264 δ 2 ϵ ξ 3 + 144 δ ϵ 2 ξ 3 − 8 ϵ 3 ξ 3 ) ] + σ 1 8 81 ϵ 4 [ − 600 + ( 4160 δ − 1616 ϵ ) + ( − 11080 δ 2 + 9823 δ ϵ − 2174 ϵ 2 + 160 ξ 3 ) + ( 14000 δ 3 − 20137 δ 2 ϵ + 9142 δ ϵ 2 − 720 δ ξ 3 − 1616 ϵ 3 + 579 ϵ ξ 3 ) + ( − 8160 δ 4 + 15632 δ 3 ϵ − 9592 δ 2 ϵ 2 + 960 δ 2 ξ 3 + 2208 δ ϵ 3 − 1484 δ ϵ ξ 3 − 88 ϵ 4 + 524 ϵ 2 ξ 3 ) + ( 1600 δ 5 − 3444 δ 4 ϵ + 2088 δ 3 ϵ 2 − 320 δ 3 ξ 3 − 244 δ 2 ϵ 3 + 652 δ 2 ϵ ξ 3 − 344 δ ϵ 2 ξ 3 + 12 ϵ 3 ξ 3 ) ] + σ 1 10 243 ϵ 5 [ 1400 + ( − 9560 δ + 3890 ϵ ) + ( 25080 δ 2 − 22108 δ ϵ + 4560 ϵ 2 − 280 ξ 3 ) + ( − 31360 δ 3 + 43508 δ 2 ϵ − 18158 δ ϵ 2 + 1280 δ ξ 3 + 2824 ϵ 3 − 1004 ϵ ξ 3 ) ] \begin{aligned}
\widehat{\mathcal{T}}_{ii}^{\mathrm{LW}}=
\frac{k^2\epsilon\lvert A_1\rvert^2}{3\mathcal{H}^2(k^2+3\epsilon\mathcal{H}^2)^5}
\left\{
3k^{12}-\mathcal{H}^2k^{10}\left[1-10\delta+16\delta^2+4\epsilon^2-2\epsilon(10\delta+23)\right]
-3\mathcal{H}^4k^8\left[-\delta(1-2\delta)^2(\delta-3)+28\epsilon^3-\epsilon^2(4\delta^2+84\delta+239)+\epsilon(8\delta^3+40\delta^2+228\delta-145)\right]
-9\mathcal{H}^6\epsilon k^6\left[4\delta^4-272\delta^3+597\delta^2-323\delta+46+48\epsilon^3+4\epsilon^2(\delta^2-67\delta-111)-4\epsilon(2\delta^3-123\delta^2+13\delta+44)\right]
-27\mathcal{H}^8\epsilon k^4\left[16\epsilon^4+\epsilon^3(36\delta^2-356\delta-367)-4\epsilon^2(10\delta^3-141\delta^2+94\delta+20)+\epsilon(-28\delta^4-80\delta^3+629\delta^2-447\delta+78)+4(1-2\delta)^2(2\delta^3-7\delta^2-4\delta+7)\right]
+81\mathcal{H}^{10}\epsilon k^2\left[20\epsilon^4+\epsilon^3(-28\delta^2+80\delta+230)+\epsilon^2(152\delta^3-244\delta^2+274\delta-39)+\epsilon(-220\delta^4+416\delta^3-915\delta^2+745\delta-182)+4(1-2\delta)^2(6\delta^3-11\delta^2+26\delta-17)\right]
+243\mathcal{H}^{12}\epsilon^3\left[12\epsilon^4+\epsilon^3(105-72\delta)+\epsilon^2(236\delta^2-196\delta+39)-2\epsilon(136\delta^3-128\delta^2+8\delta+11)+8(1-2\delta)^2(3\delta^2-\delta-1)\right]
+3\mathcal{H}^4\xi_3(k^2+3\epsilon\mathcal{H}^2)\left[-216\mathcal{H}^6\epsilon^2\{6\epsilon^2+\epsilon(5-10\delta)+(1-2\delta)^2\}\right]
+9\mathcal{H}^4\epsilon k^2\left\{-8\delta(1-2\delta)^2+4\epsilon^3-20\epsilon^2(2\delta+3)+\epsilon(68\delta^2+28\delta-31)\right\}
+6\mathcal{H}^2\epsilon k^4\left\{4\delta^2+20\delta+4\epsilon^2-4\epsilon(2\delta+5)-11\right\}
+k^6(1+2\epsilon-2\delta)^2
\right\} \\
&\simeq\mathcal{F}_{\mathrm{LW}}\{
\frac{\sigma_1^2}{3\epsilon}[-8+(24\delta-22\epsilon)+(24\delta^2-16\delta\epsilon+39\epsilon^2-8\xi_3)+(-128\delta^3+256\delta^2\epsilon-196\delta\epsilon^2+32\delta\xi_3+105\epsilon^3-40\epsilon\xi_3)+(96\delta^4-272\delta^3\epsilon+236\delta^2\epsilon^2-32\delta^2\xi_3-72\delta\epsilon^3+80\delta\epsilon\xi_3+12\epsilon^4-48\epsilon^2\xi_3)]
+\frac{\sigma_1^4}{9\epsilon^2}[-28+(256\delta-72\epsilon)+(-852\delta^2+825\delta\epsilon-234\epsilon^2+32\xi_3)+(1256\delta^3-2195\delta^2\epsilon+1254\delta\epsilon^2-136\delta\xi_3-295\epsilon^3+129\epsilon\xi_3)+(-752\delta^4+1776\delta^3\epsilon-1424\delta^2\epsilon^2+160\delta^2\xi_3+440\delta\epsilon^3-292\delta\epsilon\xi_3-40\epsilon^4+132\epsilon^2\xi_3)+(960\delta^5-2208\delta^4\epsilon+1528\delta^3\epsilon^2-320\delta^3\xi_3-280\delta^2\epsilon^2+68\delta^2\epsilon\xi_3-40\delta\epsilon^2\xi_3+4\epsilon^3\xi_3)]
+\frac{\sigma_1^6}{27\epsilon^3}[192+(-1392\delta+502\epsilon)+(3872\delta^2-3518\delta\epsilon+860\epsilon^2-80\xi_3)+(-5056\delta^3+7786\delta^2\epsilon-3934\delta\epsilon^2+352\delta\xi_3+792\epsilon^3-298\epsilon\xi_3)+(2944\delta^4-6080\delta^3\epsilon+4196\delta^2\epsilon^2-448\delta^2\xi_3-1124\delta\epsilon^3+728\delta\epsilon\xi_3+64\epsilon^4-280\epsilon^2\xi_3)+(-512\delta^5+1128\delta^4\epsilon-720\delta^3\epsilon^2+128\delta^3\xi_3+104\delta^2\epsilon^3-264\delta^2\epsilon\xi_3+144\delta\epsilon^2\xi_3-8\epsilon^3\xi_3)]
+\frac{\sigma_1^8}{81\epsilon^4}[-600+(4160\delta-1616\epsilon)+(-11080\delta^2+9823\delta\epsilon-2174\epsilon^2+160\xi_3)+(14000\delta^3-20137\delta^2\epsilon+9142\delta\epsilon^2-720\delta\xi_3-1616\epsilon^3+579\epsilon\xi_3)+(-8160\delta^4+15632\delta^3\epsilon-9592\delta^2\epsilon^2+960\delta^2\xi_3+2208\delta\epsilon^3-1484\delta\epsilon\xi_3-88\epsilon^4+524\epsilon^2\xi_3)+(1600\delta^5-3444\delta^4\epsilon+2088\delta^3\epsilon^2-320\delta^3\xi_3-244\delta^2\epsilon^3+652\delta^2\epsilon\xi_3-344\delta\epsilon^2\xi_3+12\epsilon^3\xi_3)]
+\frac{\sigma_1^{10}}{243\epsilon^5}[1400+(-9560\delta+3890\epsilon)+(25080\delta^2-22108\delta\epsilon+4560\epsilon^2-280\xi_3)+(-31360\delta^3+43508\delta^2\epsilon-18158\delta\epsilon^2+1280\delta\xi_3+2824\epsilon^3-1004\epsilon\xi_3)]
\end{aligned} T ii LW = 3 H 2 ( k 2 + 3 ϵ H 2 ) 5 k 2 ϵ ∣ A 1 ∣ 2 { 3 k 12 − H 2 k 10 [ 1 − 10 δ + 16 δ 2 + 4 ϵ 2 − 2 ϵ ( 10 δ + 23 ) ] − 3 H 4 k 8 [ − δ ( 1 − 2 δ ) 2 ( δ − 3 ) + 28 ϵ 3 − ϵ 2 ( 4 δ 2 + 84 δ + 239 ) + ϵ ( 8 δ 3 + 40 δ 2 + 228 δ − 145 ) ] − 9 H 6 ϵ k 6 [ 4 δ 4 − 272 δ 3 + 597 δ 2 − 323 δ + 46 + 48 ϵ 3 + 4 ϵ 2 ( δ 2 − 67 δ − 111 ) − 4 ϵ ( 2 δ 3 − 123 δ 2 + 13 δ + 44 ) ] − 27 H 8 ϵ k 4 [ 16 ϵ 4 + ϵ 3 ( 36 δ 2 − 356 δ − 367 ) − 4 ϵ 2 ( 10 δ 3 − 141 δ 2 + 94 δ + 20 ) + ϵ ( − 28 δ 4 − 80 δ 3 + 629 δ 2 − 447 δ + 78 ) + 4 ( 1 − 2 δ ) 2 ( 2 δ 3 − 7 δ 2 − 4 δ + 7 ) ] + 81 H 10 ϵ k 2 [ 20 ϵ 4 + ϵ 3 ( − 28 δ 2 + 80 δ + 230 ) + ϵ 2 ( 152 δ 3 − 244 δ 2 + 274 δ − 39 ) + ϵ ( − 220 δ 4 + 416 δ 3 − 915 δ 2 + 745 δ − 182 ) + 4 ( 1 − 2 δ ) 2 ( 6 δ 3 − 11 δ 2 + 26 δ − 17 ) ] + 243 H 12 ϵ 3 [ 12 ϵ 4 + ϵ 3 ( 105 − 72 δ ) + ϵ 2 ( 236 δ 2 − 196 δ + 39 ) − 2 ϵ ( 136 δ 3 − 128 δ 2 + 8 δ + 11 ) + 8 ( 1 − 2 δ ) 2 ( 3 δ 2 − δ − 1 ) ] + 3 H 4 ξ 3 ( k 2 + 3 ϵ H 2 ) [ − 216 H 6 ϵ 2 { 6 ϵ 2 + ϵ ( 5 − 10 δ ) + ( 1 − 2 δ ) 2 } ] + 9 H 4 ϵ k 2 { − 8 δ ( 1 − 2 δ ) 2 + 4 ϵ 3 − 20 ϵ 2 ( 2 δ + 3 ) + ϵ ( 68 δ 2 + 28 δ − 31 ) } + 6 H 2 ϵ k 4 { 4 δ 2 + 20 δ + 4 ϵ 2 − 4 ϵ ( 2 δ + 5 ) − 11 } + k 6 ( 1 + 2 ϵ − 2 δ ) 2 } ≃ F LW { 3 ϵ σ 1 2 [ − 8 + ( 24 δ − 22 ϵ ) + ( 24 δ 2 − 16 δ ϵ + 39 ϵ 2 − 8 ξ 3 ) + ( − 128 δ 3 + 256 δ 2 ϵ − 196 δ ϵ 2 + 32 δ ξ 3 + 105 ϵ 3 − 40 ϵ ξ 3 ) + ( 96 δ 4 − 272 δ 3 ϵ + 236 δ 2 ϵ 2 − 32 δ 2 ξ 3 − 72 δ ϵ 3 + 80 δ ϵ ξ 3 + 12 ϵ 4 − 48 ϵ 2 ξ 3 )] + 9 ϵ 2 σ 1 4 [ − 28 + ( 256 δ − 72 ϵ ) + ( − 852 δ 2 + 825 δ ϵ − 234 ϵ 2 + 32 ξ 3 ) + ( 1256 δ 3 − 2195 δ 2 ϵ + 1254 δ ϵ 2 − 136 δ ξ 3 − 295 ϵ 3 + 129 ϵ ξ 3 ) + ( − 752 δ 4 + 1776 δ 3 ϵ − 1424 δ 2 ϵ 2 + 160 δ 2 ξ 3 + 440 δ ϵ 3 − 292 δ ϵ ξ 3 − 40 ϵ 4 + 132 ϵ 2 ξ 3 ) + ( 960 δ 5 − 2208 δ 4 ϵ + 1528 δ 3 ϵ 2 − 320 δ 3 ξ 3 − 280 δ 2 ϵ 2 + 68 δ 2 ϵ ξ 3 − 40 δ ϵ 2 ξ 3 + 4 ϵ 3 ξ 3 )] + 27 ϵ 3 σ 1 6 [ 192 + ( − 1392 δ + 502 ϵ ) + ( 3872 δ 2 − 3518 δ ϵ + 860 ϵ 2 − 80 ξ 3 ) + ( − 5056 δ 3 + 7786 δ 2 ϵ − 3934 δ ϵ 2 + 352 δ ξ 3 + 792 ϵ 3 − 298 ϵ ξ 3 ) + ( 2944 δ 4 − 6080 δ 3 ϵ + 4196 δ 2 ϵ 2 − 448 δ 2 ξ 3 − 1124 δ ϵ 3 + 728 δ ϵ ξ 3 + 64 ϵ 4 − 280 ϵ 2 ξ 3 ) + ( − 512 δ 5 + 1128 δ 4 ϵ − 720 δ 3 ϵ 2 + 128 δ 3 ξ 3 + 104 δ 2 ϵ 3 − 264 δ 2 ϵ ξ 3 + 144 δ ϵ 2 ξ 3 − 8 ϵ 3 ξ 3 )] + 81 ϵ 4 σ 1 8 [ − 600 + ( 4160 δ − 1616 ϵ ) + ( − 11080 δ 2 + 9823 δ ϵ − 2174 ϵ 2 + 160 ξ 3 ) + ( 14000 δ 3 − 20137 δ 2 ϵ + 9142 δ ϵ 2 − 720 δ ξ 3 − 1616 ϵ 3 + 579 ϵ ξ 3 ) + ( − 8160 δ 4 + 15632 δ 3 ϵ − 9592 δ 2 ϵ 2 + 960 δ 2 ξ 3 + 2208 δ ϵ 3 − 1484 δ ϵ ξ 3 − 88 ϵ 4 + 524 ϵ 2 ξ 3 ) + ( 1600 δ 5 − 3444 δ 4 ϵ + 2088 δ 3 ϵ 2 − 320 δ 3 ξ 3 − 244 δ 2 ϵ 3 + 652 δ 2 ϵ ξ 3 − 344 δ ϵ 2 ξ 3 + 12 ϵ 3 ξ 3 )] + 243 ϵ 5 σ 1 10 [ 1400 + ( − 9560 δ + 3890 ϵ ) + ( 25080 δ 2 − 22108 δ ϵ + 4560 ϵ 2 − 280 ξ 3 ) + ( − 31360 δ 3 + 43508 δ 2 ϵ − 18158 δ ϵ 2 + 1280 δ ξ 3 + 2824 ϵ 3 − 1004 ϵ ξ 3 )]
+ ( 18400 δ 4 − 33776 δ 3 ϵ + 19020 δ 2 ϵ 2 − 1760 δ 2 ξ 3 − 3756 δ ϵ 3 + 2656 δ ϵ ξ 3 + 112 ϵ 4 − 896 ϵ 2 ξ 3 ) + ( − 3840 δ 5 + 8144 δ 4 ϵ − 4768 δ 3 ϵ 2 + 640 δ 3 ξ 3 + 464 δ 2 ϵ 3 − 1296 δ 2 ϵ ξ 3 + 672 δ ϵ 2 ξ 3 − 16 ϵ 3 ξ 3 ) + O ( σ 1 12 ) (strict IR) ≃ ϵ F L W { [ − 1 3 σ 1 2 + σ 1 4 ] + [ − 3 δ + 150 ϵ − 138 ϵ σ 1 2 − 252 ϵ σ 1 4 + 10 δ + ϵ 3 σ 1 2 ] + O S R ( 2 ; σ 1 ) } (intermediate IR) . (67) \begin{aligned}
&+ \left(18400\delta^{4}-33776\delta^{3}\epsilon+19020\delta^{2}\epsilon^{2}-1760\delta^{2}\xi_{3}-3756\delta\epsilon^{3}+2656\delta\epsilon\xi_{3}+112\epsilon^{4}-896\epsilon^{2}\xi_{3}\right)\\
&+\left(-3840\delta^{5}+8144\delta^{4}\epsilon-4768\delta^{3}\epsilon^{2}+640\delta^{3}\xi_{3}+464\delta^{2}\epsilon^{3}-1296\delta^{2}\epsilon\xi_{3}+672\delta\epsilon^{2}\xi_{3}-16\epsilon^{3}\xi_{3}\right)\\
&+\mathcal{O}\left(\sigma_{1}^{12}\right)\qquad\text{(strict IR)}\\
&\simeq\epsilon\mathcal{F}_{\mathrm{LW}}\left\{\left[-\frac{1}{3}\sigma_{1}^{2}+\sigma_{1}^{4}\right]+\left[-3\delta+150\epsilon-\frac{138\epsilon}{\sigma_{1}^{2}}-\frac{252\epsilon}{\sigma_{1}^{4}}+\frac{10\delta+\epsilon}{3}\sigma_{1}^{2}\right]+\mathcal{O}_{\mathrm{SR}}(2;\sigma_{1})\right\}\qquad\text{(intermediate IR)}. \tag*{(67)}
\end{aligned} + ( 18400 δ 4 − 33776 δ 3 ϵ + 19020 δ 2 ϵ 2 − 1760 δ 2 ξ 3 − 3756 δ ϵ 3 + 2656 δ ϵ ξ 3 + 112 ϵ 4 − 896 ϵ 2 ξ 3 ) + ( − 3840 δ 5 + 8144 δ 4 ϵ − 4768 δ 3 ϵ 2 + 640 δ 3 ξ 3 + 464 δ 2 ϵ 3 − 1296 δ 2 ϵ ξ 3 + 672 δ ϵ 2 ξ 3 − 16 ϵ 3 ξ 3 ) + O ( σ 1 12 ) (strict IR) ≃ ϵ F LW { [ − 3 1 σ 1 2 + σ 1 4 ] + [ − 3 δ + 150 ϵ − σ 1 2 138 ϵ − σ 1 4 252 ϵ + 3 10 δ + ϵ σ 1 2 ] + O SR ( 2 ; σ 1 ) } (intermediate IR) . (67)
Strict infrared
The strict infrared is defined by σ 1 → 0 \sigma_{1}\to0 σ 1 → 0 at fixed admissible nonzero slow-roll parameters. Neither component contains a σ 1 0 \sigma_{1}^{0} σ 1 0 term. Therefore
τ ^ 00 U E , τ ^ i i U E ⟶ 0 ( σ 1 → 0 ) . (68) \widehat{\tau}^{\mathrm{UE}}_{00},\widehat{\tau}^{\mathrm{UE}}_{ii}\longrightarrow0\qquad(\sigma_{1}\to0). \tag*{(68)} τ 00 UE , τ ii UE ⟶ 0 ( σ 1 → 0 ) . (68)
The first nonzero terms are
τ ^ 00 U E F L W = σ 1 2 ϵ [ 8 + ( 30 ϵ − 40 δ ) + O S R ( 2 ) ] + O ( σ 1 4 ) , \frac{\widehat{\tau}^{\mathrm{UE}}_{00}}{\mathcal{F}_{\mathrm{LW}}}=\frac{\sigma_{1}^{2}}{\epsilon}\left[8+(30\epsilon-40\delta)+\mathcal{O}_{\mathrm{SR}}(2)\right]+\mathcal{O}(\sigma_{1}^{4}), F LW τ 00 UE = ϵ σ 1 2 [ 8 + ( 30 ϵ − 40 δ ) + O SR ( 2 ) ] + O ( σ 1 4 ) ,
τ ^ i i U E F L W = − σ 1 2 3 ϵ [ 8 + ( 22 ϵ − 24 δ ) + O S R ( 2 ) ] + O ( σ 1 4 ) . (69) \frac{\widehat{\tau}^{\mathrm{UE}}_{ii}}{\mathcal{F}_{\mathrm{LW}}}=-\frac{\sigma_{1}^{2}}{3\epsilon}\left[8+(22\epsilon-24\delta)+\mathcal{O}_{\mathrm{SR}}(2)\right]+\mathcal{O}(\sigma_{1}^{4}). \tag*{(69)} F LW τ ii UE = − 3 ϵ σ 1 2 [ 8 + ( 22 ϵ − 24 δ ) + O SR ( 2 ) ] + O ( σ 1 4 ) . (69)
Hence, we have
lim σ 1 → 0 w r a w U E = − 1 3 + ϵ 3 − 2 δ 3 + O S R ( 2 ) . (70) \lim_{\sigma_{1}\to0}w_{\mathrm{raw}}^{\mathrm{UE}}=-\frac{1}{3}+\frac{\epsilon}{3}-\frac{2\delta}{3}+\mathcal{O}_{\mathrm{SR}}(2). \tag*{(70)} σ 1 → 0 lim w raw UE = − 3 1 + 3 ϵ − 3 2 δ + O SR ( 2 ) . (70)
At leading slow-roll order, the ratio tends to − 1 / 3 -1/3 − 1/3 .
This zero-gradient cancellation is shared by the comoving and uniform-density gauges, but the first gradient coefficients are not. Their corresponding leading pairs are ( − 10 , 10 / 3 ) σ 1 2 (-10,10/3)\sigma_{1}^{2} ( − 10 , 10/3 ) σ 1 2 and ( − 18 , 6 ) σ 1 2 (-18,6)\sigma_{1}^{2} ( − 18 , 6 ) σ 1 2 , respectively, in units of F L W \mathcal{F}_{\mathrm{LW}} F LW . The UE pair instead is ( 8 / ϵ , − 8 / ( 3 ϵ ) ) σ 1 2 (8/\epsilon,-8/(3\epsilon))\sigma_{1}^{2} ( 8/ ϵ , − 8/ ( 3 ϵ )) σ 1 2 . Thus equality at k = 0 k=0 k = 0 does not imply equality at finite wavelength.
The distinction can also be seen directly from the shift. Using the comoving result,
Q C = − Ψ ′ + H Ψ ϵ H 2 , (71) Q_{\mathrm{C}}=-\frac{\Psi^{\prime}+\mathcal{H}\Psi}{\epsilon\mathcal{H}^{2}}, \tag*{(71)} Q C = − ϵ H 2 Ψ ′ + H Ψ , (71)
the UE shift satisfies
Q U E Q C = 3 ϵ H 2 k 2 + 3 ϵ H 2 . (72) \frac{Q_{\mathrm{UE}}}{Q_{\mathrm{C}}}=\frac{3\epsilon\mathcal{H}^{2}}{k^{2}+3\epsilon\mathcal{H}^{2}}. \tag*{(72)} Q C Q UE = k 2 + 3 ϵ H 2 3 ϵ H 2 . (72)
Therefore
Q U E − Q C = − Q C k 2 k 2 + 3 ϵ H 2 ≃ − Q C σ 1 2 3 ϵ . (73) Q_{\mathrm{UE}}-Q_{\mathrm{C}}=-Q_{\mathrm{C}}\frac{k^{2}}{k^{2}+3\epsilon\mathcal{H}^{2}}\simeq-Q_{\mathrm{C}}\frac{\sigma_{1}^{2}}{3\epsilon}. \tag*{(73)} Q UE − Q C = − Q C k 2 + 3 ϵ H 2 k 2 ≃ − Q C 3 ϵ σ 1 2 . (73)
The representatives coincide at exactly zero gradient, while their first gradient corrections need not produce the same 2EMT.
For ϵ , ∣ δ ∣ ≪ σ 1 2 ≪ 1 \epsilon,|\delta|\ll\sigma_{1}^{2}\ll1 ϵ , ∣ δ ∣ ≪ σ 1 2 ≪ 1 , the formal lowest slow-roll terms in Eqs. (66) and (67) are
τ ^ 00 U E ≃ ϵ ( σ 1 2 + σ 1 4 ) F L W , τ ^ i i U E ≃ ϵ ( − 1 3 σ 1 2 + σ 1 4 ) F L W . (74) \widehat{\tau}^{\mathrm{UE}}_{00}\simeq\epsilon(\sigma_{1}^{2}+\sigma_{1}^{4})\mathcal{F}_{\mathrm{LW}},\qquad
\widehat{\tau}^{\mathrm{UE}}_{ii}\simeq\epsilon\left(-\frac{1}{3}\sigma_{1}^{2}+\sigma_{1}^{4}\right)\mathcal{F}_{\mathrm{LW}}. \tag*{(74)} τ 00 UE ≃ ϵ ( σ 1 2 + σ 1 4 ) F LW , τ ii UE ≃ ϵ ( − 3 1 σ 1 2 + σ 1 4 ) F LW . (74)
These terms are not uniformly dominant over the entire intermediate interval. Higher slow-roll orders contain increasing inverse powers of σ 1 \sigma_1 σ 1 . For example, the energy density contains 108 ϵ 2 / σ 1 4 108\epsilon^2/\sigma_1^4 108 ϵ 2 / σ 1 4 . Compared with the nominal leading term ϵ σ 1 2 \epsilon\sigma_1^2 ϵ σ 1 2 , its relative size is
108 ϵ σ 1 6 (75) \frac{108\epsilon}{\sigma_1^6} \tag*{(75)} σ 1 6 108 ϵ (75)
Ignoring numerical coefficients, a representative parametric condition for suppressing this correction is
σ 1 6 ≫ ϵ , or equivalently σ 1 2 ≫ ϵ 1 / 3 . (76) \sigma_1^6 \gg\epsilon,\qquad\text{or equivalently}\qquad\sigma_1^2 \gg\epsilon^{1/3}. \tag*{(76)} σ 1 6 ≫ ϵ , or equivalently σ 1 2 ≫ ϵ 1/3 . (76)
This is stronger than the basic intermediate hierarchy. It should be viewed as a sufficient parametric separation, not as a sharp boundary.
Sufficiently deep in this gradient-dominated sector, the leading gradient terms become
τ ^ 00 U E ≃ ϵ σ 1 2 F L W , τ ^ i i U E ≃ − 1 3 ϵ σ 1 2 F L W , w r a w U E ≃ − 1 3 . (77) \widehat{\tau}^{\mathrm{UE}}_{00} \simeq\epsilon\sigma_1^2\mathcal{F}_{\mathrm{LW}},\qquad\widehat{\tau}^{\mathrm{UE}}_{ii} \simeq-\frac{1}{3}\epsilon\sigma_1^2\mathcal{F}_{\mathrm{LW}},\qquad w^{\mathrm{UE}}_{\mathrm{raw}}\simeq-\frac{1}{3}. \tag*{(77)} τ 00 UE ≃ ϵ σ 1 2 F LW , τ ii UE ≃ − 3 1 ϵ σ 1 2 F LW , w raw UE ≃ − 3 1 . (77)
This is the regime where the UE result approaches the leading spatially-flat behavior.
As σ 1 \sigma_1 σ 1 decreases, the inverse-σ 1 \sigma_1 σ 1 terms grow. Several nominally higher slow-roll terms then become comparable to, or larger than, the lowest-order contribution. Near the denominator scale
σ 1 2 ∼ 3 ϵ , (78) \sigma_1^2 \sim3\epsilon, \tag*{(78)} σ 1 2 ∼ 3 ϵ , (78)
the wavelength and slow-roll scales also compete directly inside the UE constraint. No finite slow-roll-first truncation is uniformly reliable through this transition region. The full rational expressions must be used there.
Although w r a w U E w^{\mathrm{UE}}_{\mathrm{raw}} w raw UE tends to − 1 / 3 -1/3 − 1/3 in both asymptotic sectors, this does not imply that the ratio remains − 1 / 3 -1/3 − 1/3 through the intervening transition.
Uniform-expansion 2EMT in the short-wavelength domain
For k ≫ H k \gg\mathcal{H} k ≫ H , the short-wavelength mode in Eq. (28) is oscillatory. We retain the slowly varying envelope during differentiation, and then apply the phase average in Eq. (29) . We write
C k ≡ ∣ c 1 ∣ 2 + ∣ c 2 ∣ 2 , (79) C_k \equiv\lvert c_1\rvert^2+\lvert c_2\rvert^2, \tag*{(79)} C k ≡ ∣ c 1 ∣ 2 + ∣ c 2 ∣ 2 , (79)
and define
F S W ≡ 2 π G k 4 a 2 C k . (80) \mathcal{F}_{\mathrm{SW}}\equiv\frac{2\pi G k^4}{a^2}C_k. \tag*{(80)} F SW ≡ a 2 2 π G k 4 C k . (80)
The exact UE expressions, their wavelength-first strict-UV expansions, and their slow-roll-first intermediate-UV forms are given below.
τ ^ 00 S W = 2 π G k 2 C k a 2 ( k 2 + 3 ϵ H 2 ) 4 { 2 k 10 + H 2 k 8 ( 1 + 31 ϵ + 3 δ ) + H 4 k 6 [ δ ( δ − 1 ) 2 ( δ + 3 ) + 202 ϵ 2 − 2 ϵ ( 7 δ 2 − 10 δ + 12 ) ] + 3 H 6 ϵ k 4 [ 6 δ 4 − 10 δ 3 + 56 δ 2 − 130 δ + 78 + 212 ϵ 2 + ϵ ( − 47 δ 2 + 15 δ − 88 ) ] + 9 H 8 ϵ k 2 [ 100 ϵ 3 − 2 ϵ 2 ( 24 δ 2 − 5 δ + 50 ) + ϵ ( 9 δ 4 − 27 δ 3 + 163 δ 2 − 257 δ + 112 ) − 12 ( δ − 1 ) 3 ( δ + 1 ) ] + 81 H 10 ϵ 2 [ 5 ϵ 3 − ϵ 2 ( δ + 3 ) 2 + 2 ϵ ( 7 δ 2 − 6 δ − 1 ) − 8 ( δ − 1 ) 3 ] − H 2 k 2 ξ 3 ( k 2 + 3 ϵ H 2 ) 2 [ k 2 + H 2 ( δ − 1 ) 2 ] } ≃ 2 π G C k H 4 a 2 { 2 σ 2 − 4 + σ 2 − 2 [ 1 + ( 3 δ + 7 ϵ ) − ξ 3 ] + [ 3 δ − 36 ϵ + ( − 5 δ 2 − 16 δ ϵ + 10 ϵ 2 − ξ 3 ) + ( δ 3 − 14 δ 2 ϵ + 2 δ ξ 3 + 6 ϵ ξ 3 ) + ( δ 4 − δ 2 ξ 3 ) ] + σ 2 2 [ 234 ϵ + ( − 426 δ ϵ + 114 ϵ 2 ) + ( 228 δ 2 ϵ + 75 δ ϵ 2 − 78 ϵ 3 + 6 ϵ ξ 3 ) ] \begin{aligned}
\widehat{\tau}^{\mathrm{SW}}_{00}
={}&\frac{2\pi G k^2 C_k}{a^2(k^2+3\epsilon\mathcal{H}^2)^4}
\Bigl\{2k^{10}+\mathcal{H}^2k^8(1+31\epsilon+3\delta)\\
&+\mathcal{H}^4k^6\bigl[\delta(\delta-1)^2(\delta+3)+202\epsilon^2-2\epsilon(7\delta^2-10\delta+12)\bigr]\\
&+3\mathcal{H}^6\epsilon k^4\bigl[6\delta^4-10\delta^3+56\delta^2-130\delta+78+212\epsilon^2+\epsilon(-47\delta^2+15\delta-88)\bigr]\\
&+9\mathcal{H}^8\epsilon k^2\bigl[100\epsilon^3-2\epsilon^2(24\delta^2-5\delta+50)+\epsilon(9\delta^4-27\delta^3+163\delta^2-257\delta+112)-12(\delta-1)^3(\delta+1)\bigr]\\
&+81\mathcal{H}^{10}\epsilon^2\bigl[5\epsilon^3-\epsilon^2(\delta+3)^2+2\epsilon(7\delta^2-6\delta-1)-8(\delta-1)^3\bigr]\\
&-\mathcal{H}^2k^2\xi_3(k^2+3\epsilon\mathcal{H}^2)^2[k^2+\mathcal{H}^2(\delta-1)^2]\Bigr\}\\
\simeq{}&\frac{2\pi G C_k\mathcal{H}^4}{a^2}\Bigl\{2\sigma_2^{-4}+\sigma_2^{-2}[1+(3\delta+7\epsilon)-\xi_3]\\
&+\bigl[3\delta-36\epsilon+(-5\delta^2-16\delta\epsilon+10\epsilon^2-\xi_3)+(\delta^3-14\delta^2\epsilon+2\delta\xi_3+6\epsilon\xi_3)+(\delta^4-\delta^2\xi_3)\bigr]\\
&+\sigma_2^2\bigl[234\epsilon+(-426\delta\epsilon+114\epsilon^2)+(228\delta^2\epsilon+75\delta\epsilon^2-78\epsilon^3+6\epsilon\xi_3)\bigr]
\end{aligned} τ 00 SW = ≃ a 2 ( k 2 + 3 ϵ H 2 ) 4 2 π G k 2 C k { 2 k 10 + H 2 k 8 ( 1 + 31 ϵ + 3 δ ) + H 4 k 6 [ δ ( δ − 1 ) 2 ( δ + 3 ) + 202 ϵ 2 − 2 ϵ ( 7 δ 2 − 10 δ + 12 ) ] + 3 H 6 ϵ k 4 [ 6 δ 4 − 10 δ 3 + 56 δ 2 − 130 δ + 78 + 212 ϵ 2 + ϵ ( − 47 δ 2 + 15 δ − 88 ) ] + 9 H 8 ϵ k 2 [ 100 ϵ 3 − 2 ϵ 2 ( 24 δ 2 − 5 δ + 50 ) + ϵ ( 9 δ 4 − 27 δ 3 + 163 δ 2 − 257 δ + 112 ) − 12 ( δ − 1 ) 3 ( δ + 1 ) ] + 81 H 10 ϵ 2 [ 5 ϵ 3 − ϵ 2 ( δ + 3 ) 2 + 2 ϵ ( 7 δ 2 − 6 δ − 1 ) − 8 ( δ − 1 ) 3 ] − H 2 k 2 ξ 3 ( k 2 + 3 ϵ H 2 ) 2 [ k 2 + H 2 ( δ − 1 ) 2 ] } a 2 2 π G C k H 4 { 2 σ 2 − 4 + σ 2 − 2 [ 1 + ( 3 δ + 7 ϵ ) − ξ 3 ] + [ 3 δ − 36 ϵ + ( − 5 δ 2 − 16 δ ϵ + 10 ϵ 2 − ξ 3 ) + ( δ 3 − 14 δ 2 ϵ + 2 δ ξ 3 + 6 ϵ ξ 3 ) + ( δ 4 − δ 2 ξ 3 ) ] + σ 2 2 [ 234 ϵ + ( − 426 δ ϵ + 114 ϵ 2 ) + ( 228 δ 2 ϵ + 75 δ ϵ 2 − 78 ϵ 3 + 6 ϵ ξ 3 ) ]
+ ( − 42 δ ε 3 + 27 δ 2 ε 2 − 12 δ ε ξ 3 − 27 ε 2 ξ 3 ) + ( 6 δ 4 ε + 6 δ 2 ε ξ 3 ) + σ 2 4 [ 108 ε + ( − 216 δ ε − 1800 ε 2 ) + ( 2637 δ ε 2 − 432 ε 3 ) + ( 216 δ 3 ε − 999 δ 2 ε 2 − 270 δ ε 3 + 378 ε 4 − 27 ε 2 ξ 3 ) + ( − 108 δ 4 ε + 207 δ 3 ε 2 + 54 δ 2 ε 2 ξ 3 + 108 ε 3 ξ 3 ) + ( − 45 δ 4 ε 2 − 27 δ 2 ε 2 ξ 3 ) ] + σ 2 6 [ − 648 ε 2 + ( 648 δ ε 2 + 8802 ε 3 ) + ( 1944 δ 2 ε 2 − 9936 δ ε 3 + 2106 ε 4 ) + ( − 3240 δ 3 ε 2 + 1350 δ 2 ε 3 + 189 δ ε 4 − 1566 ε 5 + 108 ε 3 ξ 3 ) + ( 1296 δ 4 ε 2 − 324 δ 3 ε 3 − 27 δ 2 ε 4 − 216 δ ε 3 ξ 3 − 405 ε 4 ξ 3 ) + ( 108 δ 4 ε 3 + 108 δ 2 ε 3 ξ 3 ) ] + σ 2 8 [ 1944 ε 3 + ( 3888 δ ε 3 − 33696 ε 4 ) + ( − 23328 δ 2 ε 3 + 22599 δ ε 4 − 11340 ε 5 ) + ( 27216 δ 3 ε 3 + 13527 δ 2 ε 4 + 5508 δ ε 5 + 5994 ε 6 − 405 ε 4 ξ 3 ) + ( − 9720 δ 4 ε 3 − 2835 δ 3 ε 4 − 1458 δ 2 ε 5 + 810 δ ε 4 ξ 3 + 1458 ε 5 ξ 3 ) + ( 405 δ 4 ε 4 − 405 δ 2 ε 4 ξ 3 ) ] + σ 2 10 [ − 58320 δ 4 ε 4 + 104490 ε 5 + ( 174960 δ 4 ε 2 + 15066 δ ε 5 + 59778 ε 6 ) + ( − 174960 δ 3 ε 4 − 145800 δ 2 ε 5 − 53217 δ ε 6 − 21870 ε 7 + 1458 ε 5 ξ 3 ) + ( 58320 δ 4 ε 4 + 32562 δ 3 ε 5 + 16767 δ 2 ε 6 − 2916 δ ε 5 ξ 3 − 5103 ε 6 ξ 3 ) + ( − 6318 δ 5 ε 4 + 1458 δ 2 ε 5 ξ 3 ) + O ( σ 2 12 ) } (strict UV) \begin{aligned}
&+ \left(-42\delta\varepsilon^{3}+27\delta^{2}\varepsilon^{2}-12\delta\varepsilon\xi_{3}-27\varepsilon^{2}\xi_{3}\right)+\left(6\delta^{4}\varepsilon+6\delta^{2}\varepsilon\xi_{3}\right)\\
&+\sigma_{2}^{4}\left[108\varepsilon+\left(-216\delta\varepsilon-1800\varepsilon^{2}\right)+\left(2637\delta\varepsilon^{2}-432\varepsilon^{3}\right)\right.\\
&\qquad+\left(216\delta^{3}\varepsilon-999\delta^{2}\varepsilon^{2}-270\delta\varepsilon^{3}+378\varepsilon^{4}-27\varepsilon^{2}\xi_{3}\right)\\
&\qquad+\left.\left(-108\delta^{4}\varepsilon+207\delta^{3}\varepsilon^{2}+54\delta^{2}\varepsilon^{2}\xi_{3}+108\varepsilon^{3}\xi_{3}\right)+\left(-45\delta^{4}\varepsilon^{2}-27\delta^{2}\varepsilon^{2}\xi_{3}\right)\right]\\
&+\sigma_{2}^{6}\left[-648\varepsilon^{2}+\left(648\delta\varepsilon^{2}+8802\varepsilon^{3}\right)+\left(1944\delta^{2}\varepsilon^{2}-9936\delta\varepsilon^{3}+2106\varepsilon^{4}\right)\right.\\
&\qquad+\left(-3240\delta^{3}\varepsilon^{2}+1350\delta^{2}\varepsilon^{3}+189\delta\varepsilon^{4}-1566\varepsilon^{5}+108\varepsilon^{3}\xi_{3}\right)\\
&\qquad+\left(1296\delta^{4}\varepsilon^{2}-324\delta^{3}\varepsilon^{3}-27\delta^{2}\varepsilon^{4}-216\delta\varepsilon^{3}\xi_{3}-405\varepsilon^{4}\xi_{3}\right)\\
&\qquad+\left.\left(108\delta^{4}\varepsilon^{3}+108\delta^{2}\varepsilon^{3}\xi_{3}\right)\right]\\
&+\sigma_{2}^{8}\left[1944\varepsilon^{3}+\left(3888\delta\varepsilon^{3}-33696\varepsilon^{4}\right)+\left(-23328\delta^{2}\varepsilon^{3}+22599\delta\varepsilon^{4}-11340\varepsilon^{5}\right)\right.\\
&\qquad+\left(27216\delta^{3}\varepsilon^{3}+13527\delta^{2}\varepsilon^{4}+5508\delta\varepsilon^{5}+5994\varepsilon^{6}-405\varepsilon^{4}\xi_{3}\right)\\
&\qquad+\left(-9720\delta^{4}\varepsilon^{3}-2835\delta^{3}\varepsilon^{4}-1458\delta^{2}\varepsilon^{5}+810\delta\varepsilon^{4}\xi_{3}+1458\varepsilon^{5}\xi_{3}\right)\\
&\qquad+\left.\left(405\delta^{4}\varepsilon^{4}-405\delta^{2}\varepsilon^{4}\xi_{3}\right)\right]\\
&+\sigma_{2}^{10}\left[-58320\delta^{4}\varepsilon^{4}+104490\varepsilon^{5}+\left(174960\delta^{4}\varepsilon^{2}+15066\delta\varepsilon^{5}+59778\varepsilon^{6}\right)\right.\\
&\qquad+\left(-174960\delta^{3}\varepsilon^{4}-145800\delta^{2}\varepsilon^{5}-53217\delta\varepsilon^{6}-21870\varepsilon^{7}+1458\varepsilon^{5}\xi_{3}\right)\\
&\qquad+\left(58320\delta^{4}\varepsilon^{4}+32562\delta^{3}\varepsilon^{5}+16767\delta^{2}\varepsilon^{6}-2916\delta\varepsilon^{5}\xi_{3}-5103\varepsilon^{6}\xi_{3}\right)\\
&\qquad+\left.\left(-6318\delta^{5}\varepsilon^{4}+1458\delta^{2}\varepsilon^{5}\xi_{3}\right)+\mathcal{O}(\sigma_{2}^{12})\right\}\qquad\text{(strict UV)}
\end{aligned} + ( − 42 δ ε 3 + 27 δ 2 ε 2 − 12 δ ε ξ 3 − 27 ε 2 ξ 3 ) + ( 6 δ 4 ε + 6 δ 2 ε ξ 3 ) + σ 2 4 [ 108 ε + ( − 216 δ ε − 1800 ε 2 ) + ( 2637 δ ε 2 − 432 ε 3 ) + ( 216 δ 3 ε − 999 δ 2 ε 2 − 270 δ ε 3 + 378 ε 4 − 27 ε 2 ξ 3 ) + ( − 108 δ 4 ε + 207 δ 3 ε 2 + 54 δ 2 ε 2 ξ 3 + 108 ε 3 ξ 3 ) + ( − 45 δ 4 ε 2 − 27 δ 2 ε 2 ξ 3 ) ] + σ 2 6 [ − 648 ε 2 + ( 648 δ ε 2 + 8802 ε 3 ) + ( 1944 δ 2 ε 2 − 9936 δ ε 3 + 2106 ε 4 ) + ( − 3240 δ 3 ε 2 + 1350 δ 2 ε 3 + 189 δ ε 4 − 1566 ε 5 + 108 ε 3 ξ 3 ) + ( 1296 δ 4 ε 2 − 324 δ 3 ε 3 − 27 δ 2 ε 4 − 216 δ ε 3 ξ 3 − 405 ε 4 ξ 3 ) + ( 108 δ 4 ε 3 + 108 δ 2 ε 3 ξ 3 ) ] + σ 2 8 [ 1944 ε 3 + ( 3888 δ ε 3 − 33696 ε 4 ) + ( − 23328 δ 2 ε 3 + 22599 δ ε 4 − 11340 ε 5 ) + ( 27216 δ 3 ε 3 + 13527 δ 2 ε 4 + 5508 δ ε 5 + 5994 ε 6 − 405 ε 4 ξ 3 ) + ( − 9720 δ 4 ε 3 − 2835 δ 3 ε 4 − 1458 δ 2 ε 5 + 810 δ ε 4 ξ 3 + 1458 ε 5 ξ 3 ) + ( 405 δ 4 ε 4 − 405 δ 2 ε 4 ξ 3 ) ] + σ 2 10 [ − 58320 δ 4 ε 4 + 104490 ε 5 + ( 174960 δ 4 ε 2 + 15066 δ ε 5 + 59778 ε 6 ) + ( − 174960 δ 3 ε 4 − 145800 δ 2 ε 5 − 53217 δ ε 6 − 21870 ε 7 + 1458 ε 5 ξ 3 ) + ( 58320 δ 4 ε 4 + 32562 δ 3 ε 5 + 16767 δ 2 ε 6 − 2916 δ ε 5 ξ 3 − 5103 ε 6 ξ 3 ) + ( − 6318 δ 5 ε 4 + 1458 δ 2 ε 5 ξ 3 ) + O ( σ 2 12 ) } (strict UV)
≃ 2 π G C k H 4 a 2 { 2 σ 2 − 4 + σ 2 − 2 + [ ( 3 δ + 7 ε ) σ 2 − 2 + ( 3 δ − 36 ε ) + 234 ε σ 2 2 + 108 ε σ 2 4 ] + O S R ( 2 ; σ 2 ) } (intermediate UV). (81) \begin{aligned}
&\simeq\frac{2\pi G C_k \mathcal{H}^{4}}{a^{2}}\left\{2\sigma_{2}^{-4}+\sigma_{2}^{-2}\right.\\
&\qquad+\left[(3\delta+7\varepsilon)\sigma_{2}^{-2}+(3\delta-36\varepsilon)+234\varepsilon\sigma_{2}^{2}+108\varepsilon\sigma_{2}^{4}\right]\\
&\qquad\left.+\mathcal{O}_{\mathrm{SR}}(2;\sigma_{2})\right\}\qquad\text{(intermediate UV).} \tag*{(81)}
\end{aligned} ≃ a 2 2 π G C k H 4 { 2 σ 2 − 4 + σ 2 − 2 + [ ( 3 δ + 7 ε ) σ 2 − 2 + ( 3 δ − 36 ε ) + 234 ε σ 2 2 + 108 ε σ 2 4 ] + O SR ( 2 ; σ 2 ) } (intermediate UV). (81)
τ ^ i i S W = 2 π G k 2 C k 3 a 2 ( k 2 + 3 ε H 2 ) 5 { 2 k 12 + H 2 k 10 ( − 1 + 29 ε − δ − 4 δ 2 ) + 3 H 4 k 8 [ δ ( δ − 3 ) ( δ − 1 ) 2 + 39 ε 2 + ε ( − 30 δ 2 − 5 δ + 31 ) ] − 9 H 6 ε k 6 [ δ 4 − 69 δ 3 + 201 δ 2 − 287 δ + 130 + 16 ε 2 + ε ( 39 δ 2 + 39 δ − 158 ) ] − 27 H 8 ε k 4 [ 82 ε 3 + ε 2 ( − 19 δ 2 + 93 δ − 254 ) + ε ( 9 δ 4 − 137 δ 3 + 353 δ 2 − 443 δ + 170 ) + 4 ( δ − 1 ) 2 ( 2 δ 3 − 7 δ 2 − 4 δ + 7 ) ] − 81 H 10 ε 2 k 2 [ 69 ε 3 + ε 2 ( − 43 δ 2 + 36 δ − 127 ) + ε ( 7 δ 4 + 65 δ 3 − 177 δ 2 + 147 δ − 66 ) − 4 ( δ − 1 ) 2 ( 6 δ 3 − 11 δ 2 + 26 δ − 17 ) ] − 243 H 12 ε [ 19 ε 3 + ε 2 ( − 11 δ 2 − 22 δ + 1 ) + 2 ε ( 16 δ 3 − 17 δ 2 + 6 δ − 5 ) − 8 ( δ − 1 ) 2 ( 3 δ 2 − δ − 1 ) ] + 3 H 2 ξ 3 ( k 2 + 3 ε H 2 ) [ 216 H 8 ε 2 ( δ − 1 ) ( 1 + ε − δ ) + 9 H 6 ε k 2 { − 8 δ ( δ − 1 ) 2 + ε ( δ 2 + 30 δ − 55 ) } + 3 H 4 ε k 4 ( 2 δ 2 − 4 δ + 3 ε − 22 ) + H 2 k 6 { 6 ε + ( δ − 1 ) 2 } + k 8 ] } \begin{aligned}
\widehat{\tau}^{\mathrm{SW}}_{ii}
={}&\frac{2\pi G k^{2}C_k}{3a^{2}(k^{2}+3\varepsilon\mathcal{H}^{2})^{5}}\left\{2k^{12}+\mathcal{H}^{2}k^{10}(-1+29\varepsilon-\delta-4\delta^{2})\right.\\
&+3\mathcal{H}^{4}k^{8}\left[\delta(\delta-3)(\delta-1)^{2}+39\varepsilon^{2}+\varepsilon(-30\delta^{2}-5\delta+31)\right]\\
&-9\mathcal{H}^{6}\varepsilon k^{6}\left[\delta^{4}-69\delta^{3}+201\delta^{2}-287\delta+130+16\varepsilon^{2}+\varepsilon(39\delta^{2}+39\delta-158)\right]\\
&-27\mathcal{H}^{8}\varepsilon k^{4}\left[82\varepsilon^{3}+\varepsilon^{2}(-19\delta^{2}+93\delta-254)+\varepsilon(9\delta^{4}-137\delta^{3}+353\delta^{2}-443\delta+170)\right.\\
&\qquad\left.+4(\delta-1)^{2}(2\delta^{3}-7\delta^{2}-4\delta+7)\right]\\
&-81\mathcal{H}^{10}\varepsilon^{2}k^{2}\left[69\varepsilon^{3}+\varepsilon^{2}(-43\delta^{2}+36\delta-127)\right.\\
&\qquad\left.+\varepsilon(7\delta^{4}+65\delta^{3}-177\delta^{2}+147\delta-66)-4(\delta-1)^{2}(6\delta^{3}-11\delta^{2}+26\delta-17)\right]\\
&-243\mathcal{H}^{12}\varepsilon\left[19\varepsilon^{3}+\varepsilon^{2}(-11\delta^{2}-22\delta+1)+2\varepsilon(16\delta^{3}-17\delta^{2}+6\delta-5)\right.\\
&\qquad\left.-8(\delta-1)^{2}(3\delta^{2}-\delta-1)\right]\\
&\left.+3\mathcal{H}^{2}\xi_{3}(k^{2}+3\varepsilon\mathcal{H}^{2})\left[216\mathcal{H}^{8}\varepsilon^{2}(\delta-1)(1+\varepsilon-\delta)+9\mathcal{H}^{6}\varepsilon k^{2}\{-8\delta(\delta-1)^{2}+\varepsilon(\delta^{2}+30\delta-55)\}\right.\right.\\
&\qquad\left.\left.+3\mathcal{H}^{4}\varepsilon k^{4}(2\delta^{2}-4\delta+3\varepsilon-22)+\mathcal{H}^{2}k^{6}\{6\varepsilon+(\delta-1)^{2}\}+k^{8}\right]\right\}
\end{aligned} τ ii SW = 3 a 2 ( k 2 + 3 ε H 2 ) 5 2 π G k 2 C k { 2 k 12 + H 2 k 10 ( − 1 + 29 ε − δ − 4 δ 2 ) + 3 H 4 k 8 [ δ ( δ − 3 ) ( δ − 1 ) 2 + 39 ε 2 + ε ( − 30 δ 2 − 5 δ + 31 ) ] − 9 H 6 ε k 6 [ δ 4 − 69 δ 3 + 201 δ 2 − 287 δ + 130 + 16 ε 2 + ε ( 39 δ 2 + 39 δ − 158 ) ] − 27 H 8 ε k 4 [ 82 ε 3 + ε 2 ( − 19 δ 2 + 93 δ − 254 ) + ε ( 9 δ 4 − 137 δ 3 + 353 δ 2 − 443 δ + 170 ) + 4 ( δ − 1 ) 2 ( 2 δ 3 − 7 δ 2 − 4 δ + 7 ) ] − 81 H 10 ε 2 k 2 [ 69 ε 3 + ε 2 ( − 43 δ 2 + 36 δ − 127 ) + ε ( 7 δ 4 + 65 δ 3 − 177 δ 2 + 147 δ − 66 ) − 4 ( δ − 1 ) 2 ( 6 δ 3 − 11 δ 2 + 26 δ − 17 ) ] − 243 H 12 ε [ 19 ε 3 + ε 2 ( − 11 δ 2 − 22 δ + 1 ) + 2 ε ( 16 δ 3 − 17 δ 2 + 6 δ − 5 ) − 8 ( δ − 1 ) 2 ( 3 δ 2 − δ − 1 ) ] + 3 H 2 ξ 3 ( k 2 + 3 ε H 2 ) [ 216 H 8 ε 2 ( δ − 1 ) ( 1 + ε − δ ) + 9 H 6 ε k 2 { − 8 δ ( δ − 1 ) 2 + ε ( δ 2 + 30 δ − 55 )} + 3 H 4 ε k 4 ( 2 δ 2 − 4 δ + 3 ε − 22 ) + H 2 k 6 { 6 ε + ( δ − 1 ) 2 } + k 8 ] }
τ ^ i i S W ≃ 2 π G C k H 4 a 2 { 2 3 σ 2 − 4 + σ 2 − 2 [ − 1 3 − δ + ε 3 − 4 δ 2 3 + ξ 3 ] + [ − 3 δ + 36 ε + ( 7 δ 2 − 16 ε 2 + ξ 3 ) + ( − 5 δ 3 − 10 δ 2 ε − 2 δ ξ 3 − 6 ε ξ 3 ) + ( δ 4 + δ 2 ξ 3 ) ] + σ 2 2 [ − 390 ε + ( 906 δ ε − 36 ε 2 ) + ( − 708 δ 2 ε − 87 δ ε 2 + 42 ε 3 − 78 ε ξ 3 ) + ( 282 δ 3 ε + 153 δ 2 ε 2 + 12 δ ε ξ 3 + 27 ε 2 ξ 3 ) + ( − 18 δ 4 ε − 6 δ 2 ε ξ 3 ) ] + σ 2 4 [ − 252 ε + ( 648 δ ε + 4320 ε 2 ) + ( − 288 δ 2 ε − 9333 δ ε 2 − 324 ε 3 ) + ( − 432 δ 3 ε + 6813 δ 2 ε 2 + 558 δ ε 3 − 72 δ ε ξ 3 − 108 ε 4 + 387 ε 2 ξ 3 ) + ( 396 δ 4 ε − 2547 δ 3 ε 2 − 864 δ 2 ε 3 + 144 δ 2 ε ξ 3 + 234 δ ε 2 ξ 3 − 108 ε 3 ξ 3 ) \begin{aligned}
\widehat{\tau}^{\mathrm{SW}}_{ii}
\simeq{}&\frac{2\pi G C_k\mathcal{H}^{4}}{a^{2}}\left\{\frac{2}{3}\sigma_{2}^{-4}+\sigma_{2}^{-2}\left[-\frac{1}{3}-\frac{\delta+\varepsilon}{3}-\frac{4\delta^{2}}{3}+\xi_{3}\right]\right.\\
&+\left[-3\delta+36\varepsilon+(7\delta^{2}-16\varepsilon^{2}+\xi_{3})+(-5\delta^{3}-10\delta^{2}\varepsilon-2\delta\xi_{3}-6\varepsilon\xi_{3})+(\delta^{4}+\delta^{2}\xi_{3})\right]\\
&+\sigma_{2}^{2}\left[-390\varepsilon+(906\delta\varepsilon-36\varepsilon^{2})+(-708\delta^{2}\varepsilon-87\delta\varepsilon^{2}+42\varepsilon^{3}-78\varepsilon\xi_{3})\right.\\
&\qquad\left.+(282\delta^{3}\varepsilon+153\delta^{2}\varepsilon^{2}+12\delta\varepsilon\xi_{3}+27\varepsilon^{2}\xi_{3})+(-18\delta^{4}\varepsilon-6\delta^{2}\varepsilon\xi_{3})\right]\\
&+\sigma_{2}^{4}\left[-252\varepsilon+(648\delta\varepsilon+4320\varepsilon^{2})+(-288\delta^{2}\varepsilon-9333\delta\varepsilon^{2}-324\varepsilon^{3})\right.\\
&\qquad+(-432\delta^{3}\varepsilon+6813\delta^{2}\varepsilon^{2}+558\delta\varepsilon^{3}-72\delta\varepsilon\xi_{3}-108\varepsilon^{4}+387\varepsilon^{2}\xi_{3})\\
&\qquad\left.+(396\delta^{4}\varepsilon-2547\delta^{3}\varepsilon^{2}-864\delta^{2}\varepsilon^{3}+144\delta^{2}\varepsilon\xi_{3}+234\delta\varepsilon^{2}\xi_{3}-108\varepsilon^{3}\xi_{3})\right.
\end{aligned} τ ii SW ≃ a 2 2 π G C k H 4 { 3 2 σ 2 − 4 + σ 2 − 2 [ − 3 1 − 3 δ + ε − 3 4 δ 2 + ξ 3 ] + [ − 3 δ + 36 ε + ( 7 δ 2 − 16 ε 2 + ξ 3 ) + ( − 5 δ 3 − 10 δ 2 ε − 2 δ ξ 3 − 6 ε ξ 3 ) + ( δ 4 + δ 2 ξ 3 ) ] + σ 2 2 [ − 390 ε + ( 906 δ ε − 36 ε 2 ) + ( − 708 δ 2 ε − 87 δ ε 2 + 42 ε 3 − 78 ε ξ 3 ) + ( 282 δ 3 ε + 153 δ 2 ε 2 + 12 δ ε ξ 3 + 27 ε 2 ξ 3 ) + ( − 18 δ 4 ε − 6 δ 2 ε ξ 3 ) ] + σ 2 4 [ − 252 ε + ( 648 δ ε + 4320 ε 2 ) + ( − 288 δ 2 ε − 9333 δ ε 2 − 324 ε 3 ) + ( − 432 δ 3 ε + 6813 δ 2 ε 2 + 558 δ ε 3 − 72 δ ε ξ 3 − 108 ε 4 + 387 ε 2 ξ 3 ) + ( 396 δ 4 ε − 2547 δ 3 ε 2 − 864 δ 2 ε 3 + 144 δ 2 ε ξ 3 + 234 δ ε 2 ξ 3 − 108 ε 3 ξ 3 )
+ ( − 72 δ 5 ϵ + 99 δ 4 ϵ 2 − 72 δ 3 ϵ ξ 3 + 27 δ 2 ϵ 2 ξ 3 ) + σ 2 6 [ 1944 ϵ 2 + ( − 3240 δ ϵ 2 − 27918 ϵ 3 ) + ( − 4320 δ 2 ϵ 2 + 55296 δ ϵ 3 + 1944 ϵ 4 − 216 ϵ 2 ξ 3 ) + ( 12312 δ 3 ϵ 2 − 35586 δ 2 ϵ 3 − 1377 δ ϵ 4 + 1296 δ ϵ 2 ξ 3 + 270 ϵ 5 − 756 ϵ 3 ξ 3 ) + ( − 8424 δ 4 ϵ 2 + 12420 δ 3 ϵ 3 + 3591 δ 2 ϵ 4 − 1944 δ 2 ϵ 2 ξ 3 − 3024 δ ϵ 3 ξ 3 + 405 ϵ 4 ξ 3 ) + ( 1728 δ 5 ϵ 2 − 324 δ 4 ϵ 3 + 864 δ 3 ϵ 2 ξ 3 − 108 δ 2 ϵ 3 ξ 3 ) ] + σ 2 8 [ − 7128 ϵ 3 + ( − 9072 δ ϵ 3 + 136080 ϵ 4 ) + ( 93312 δ 2 ϵ 3 − 233847 δ ϵ 4 − 4860 ϵ 5 + 2592 ϵ 3 ξ 3 ) + ( − 150336 δ 3 ϵ 3 + 111699 δ 2 ϵ 4 − 4212 δ ϵ 5 − 11664 δ ϵ 3 ξ 3 − 648 ϵ 6 − 3483 ϵ 4 ξ 3 ) + ( 92664 δ 4 ϵ 3 − 33777 δ 3 ϵ 4 − 12150 δ 2 ϵ 5 + 15552 δ 2 ϵ 3 ξ 3 + 22518 δ ϵ 4 ξ 3 − 1458 ϵ 5 ξ 3 ) + ( − 19440 δ 5 ϵ 3 + 405 δ 4 ϵ 4 − 6480 δ 3 ϵ 3 ξ 3 + 405 δ 2 ϵ 4 ξ 3 ) ] + σ 2 10 [ 252720 δ ϵ 4 − 537030 ϵ 5 + ( − 933120 δ 2 ϵ 4 + 684774 δ ϵ 5 − 8748 ϵ 6 − 19440 ϵ 4 ξ 3 ) + ( 1263600 δ 3 ϵ 4 − 27216 δ 2 ϵ 5 + 71685 δ ϵ 6 + 77760 δ ϵ 4 ξ 3 + 1458 ϵ 7 + 47142 ϵ 5 ξ 3 ) + ( − 738720 δ 4 ϵ 4 − 36450 δ 3 ϵ 5 + 32805 δ 2 ϵ 6 − 97200 δ 2 ϵ 4 ξ 3 − 133164 δ ϵ 5 ξ 3 + 5103 ϵ 6 ξ 3 ) + ( 155520 δ 5 ϵ 4 + 3402 δ 4 ϵ 5 + 38880 δ 3 ϵ 4 ξ 3 − 1458 δ 2 ϵ 5 ξ 3 ) ] + O ( σ 2 12 ) \begin{aligned}
&+\left(-72\delta^{5}\epsilon+99\delta^{4}\epsilon^{2}-72\delta^{3}\epsilon\xi_{3}+27\delta^{2}\epsilon^{2}\xi_{3}\right)\\
&+\sigma_{2}^{6}\left[1944\epsilon^{2}+\left(-3240\delta\epsilon^{2}-27918\epsilon^{3}\right)+\left(-4320\delta^{2}\epsilon^{2}+55296\delta\epsilon^{3}+1944\epsilon^{4}-216\epsilon^{2}\xi_{3}\right)\right.\\
&\quad+\left(12312\delta^{3}\epsilon^{2}-35586\delta^{2}\epsilon^{3}-1377\delta\epsilon^{4}+1296\delta\epsilon^{2}\xi_{3}+270\epsilon^{5}-756\epsilon^{3}\xi_{3}\right)\\
&\quad+\left(-8424\delta^{4}\epsilon^{2}+12420\delta^{3}\epsilon^{3}+3591\delta^{2}\epsilon^{4}-1944\delta^{2}\epsilon^{2}\xi_{3}-3024\delta\epsilon^{3}\xi_{3}+405\epsilon^{4}\xi_{3}\right)\\
&\quad\left.+\left(1728\delta^{5}\epsilon^{2}-324\delta^{4}\epsilon^{3}+864\delta^{3}\epsilon^{2}\xi_{3}-108\delta^{2}\epsilon^{3}\xi_{3}\right)\right]\\
&+\sigma_{2}^{8}\left[-7128\epsilon^{3}+\left(-9072\delta\epsilon^{3}+136080\epsilon^{4}\right)+\left(93312\delta^{2}\epsilon^{3}-233847\delta\epsilon^{4}-4860\epsilon^{5}+2592\epsilon^{3}\xi_{3}\right)\right.\\
&\quad+\left(-150336\delta^{3}\epsilon^{3}+111699\delta^{2}\epsilon^{4}-4212\delta\epsilon^{5}-11664\delta\epsilon^{3}\xi_{3}-648\epsilon^{6}-3483\epsilon^{4}\xi_{3}\right)\\
&\quad+\left(92664\delta^{4}\epsilon^{3}-33777\delta^{3}\epsilon^{4}-12150\delta^{2}\epsilon^{5}+15552\delta^{2}\epsilon^{3}\xi_{3}+22518\delta\epsilon^{4}\xi_{3}-1458\epsilon^{5}\xi_{3}\right)\\
&\quad\left.+\left(-19440\delta^{5}\epsilon^{3}+405\delta^{4}\epsilon^{4}-6480\delta^{3}\epsilon^{3}\xi_{3}+405\delta^{2}\epsilon^{4}\xi_{3}\right)\right]\\
&+\sigma_{2}^{10}\left[252720\delta\epsilon^{4}-537030\epsilon^{5}+\left(-933120\delta^{2}\epsilon^{4}+684774\delta\epsilon^{5}-8748\epsilon^{6}-19440\epsilon^{4}\xi_{3}\right)\right.\\
&\quad+\left(1263600\delta^{3}\epsilon^{4}-27216\delta^{2}\epsilon^{5}+71685\delta\epsilon^{6}+77760\delta\epsilon^{4}\xi_{3}+1458\epsilon^{7}+47142\epsilon^{5}\xi_{3}\right)\\
&\quad+\left(-738720\delta^{4}\epsilon^{4}-36450\delta^{3}\epsilon^{5}+32805\delta^{2}\epsilon^{6}-97200\delta^{2}\epsilon^{4}\xi_{3}-133164\delta\epsilon^{5}\xi_{3}+5103\epsilon^{6}\xi_{3}\right)\\
&\quad\left.+\left(155520\delta^{5}\epsilon^{4}+3402\delta^{4}\epsilon^{5}+38880\delta^{3}\epsilon^{4}\xi_{3}-1458\delta^{2}\epsilon^{5}\xi_{3}\right)\right]\\
&+\mathcal{O}(\sigma_{2}^{12})
\end{aligned} + ( − 72 δ 5 ϵ + 99 δ 4 ϵ 2 − 72 δ 3 ϵ ξ 3 + 27 δ 2 ϵ 2 ξ 3 ) + σ 2 6 [ 1944 ϵ 2 + ( − 3240 δ ϵ 2 − 27918 ϵ 3 ) + ( − 4320 δ 2 ϵ 2 + 55296 δ ϵ 3 + 1944 ϵ 4 − 216 ϵ 2 ξ 3 ) + ( 12312 δ 3 ϵ 2 − 35586 δ 2 ϵ 3 − 1377 δ ϵ 4 + 1296 δ ϵ 2 ξ 3 + 270 ϵ 5 − 756 ϵ 3 ξ 3 ) + ( − 8424 δ 4 ϵ 2 + 12420 δ 3 ϵ 3 + 3591 δ 2 ϵ 4 − 1944 δ 2 ϵ 2 ξ 3 − 3024 δ ϵ 3 ξ 3 + 405 ϵ 4 ξ 3 ) + ( 1728 δ 5 ϵ 2 − 324 δ 4 ϵ 3 + 864 δ 3 ϵ 2 ξ 3 − 108 δ 2 ϵ 3 ξ 3 ) ] + σ 2 8 [ − 7128 ϵ 3 + ( − 9072 δ ϵ 3 + 136080 ϵ 4 ) + ( 93312 δ 2 ϵ 3 − 233847 δ ϵ 4 − 4860 ϵ 5 + 2592 ϵ 3 ξ 3 ) + ( − 150336 δ 3 ϵ 3 + 111699 δ 2 ϵ 4 − 4212 δ ϵ 5 − 11664 δ ϵ 3 ξ 3 − 648 ϵ 6 − 3483 ϵ 4 ξ 3 ) + ( 92664 δ 4 ϵ 3 − 33777 δ 3 ϵ 4 − 12150 δ 2 ϵ 5 + 15552 δ 2 ϵ 3 ξ 3 + 22518 δ ϵ 4 ξ 3 − 1458 ϵ 5 ξ 3 ) + ( − 19440 δ 5 ϵ 3 + 405 δ 4 ϵ 4 − 6480 δ 3 ϵ 3 ξ 3 + 405 δ 2 ϵ 4 ξ 3 ) ] + σ 2 10 [ 252720 δ ϵ 4 − 537030 ϵ 5 + ( − 933120 δ 2 ϵ 4 + 684774 δ ϵ 5 − 8748 ϵ 6 − 19440 ϵ 4 ξ 3 ) + ( 1263600 δ 3 ϵ 4 − 27216 δ 2 ϵ 5 + 71685 δ ϵ 6 + 77760 δ ϵ 4 ξ 3 + 1458 ϵ 7 + 47142 ϵ 5 ξ 3 ) + ( − 738720 δ 4 ϵ 4 − 36450 δ 3 ϵ 5 + 32805 δ 2 ϵ 6 − 97200 δ 2 ϵ 4 ξ 3 − 133164 δ ϵ 5 ξ 3 + 5103 ϵ 6 ξ 3 ) + ( 155520 δ 5 ϵ 4 + 3402 δ 4 ϵ 5 + 38880 δ 3 ϵ 4 ξ 3 − 1458 δ 2 ϵ 5 ξ 3 ) ] + O ( σ 2 12 )
\begin{aligned}
&\simeq\frac{2\pi G C_k H^4}{a^2}\left\{\left[\frac{2}{3}\sigma_{2}^{-4}-\frac{1}{3}\sigma_{2}^{-2}\right]\right.\\
&\left.\qquad+\left[-\frac{\delta+\epsilon}{3}\sigma_{2}^{-2}+(-3\delta+36\epsilon)-390\epsilon\sigma_{2}^{2}-252\epsilon\sigma_{2}^{4}+\mathcal{O}_{\mathrm{SR}}(2;\sigma_{2})\right]\right\}
\end{aligned}
\qquad\text{(intermediate UV).} \tag*{(82)}\end{aligned}
Strict ultraviolet
Taking σ 2 → 0 \sigma_{2}\to0 σ 2 → 0 at fixed admissible nonzero slow-roll parameters gives
τ ^ 00 U E F S W = 2 + [ 1 + ( 7 ϵ + 3 δ ) − ξ 3 ] σ 2 2 + O ( σ 2 4 ) , \frac{\hat{\tau}^{\mathrm{UE}}_{00}}{\mathcal{F}_{\mathrm{SW}}}
=2+\left[1+(7\epsilon+3\delta)-\xi_{3}\right]\sigma_{2}^{2}+\mathcal{O}(\sigma_{2}^{4}), F SW τ ^ 00 UE = 2 + [ 1 + ( 7 ϵ + 3 δ ) − ξ 3 ] σ 2 2 + O ( σ 2 4 ) ,
τ ^ i i U E F S W = 2 3 − 1 3 [ 1 + ( ϵ + δ ) + 4 δ 2 − 3 ξ 3 ] σ 2 2 + O ( σ 2 4 ) . (83) \frac{\hat{\tau}^{\mathrm{UE}}_{ii}}{\mathcal{F}_{\mathrm{SW}}}
=\frac{2}{3}-\frac{1}{3}\left[1+(\epsilon+\delta)+4\delta^{2}-3\xi_{3}\right]\sigma_{2}^{2}+\mathcal{O}(\sigma_{2}^{4}). \tag*{(83)} F SW τ ^ ii UE = 3 2 − 3 1 [ 1 + ( ϵ + δ ) + 4 δ 2 − 3 ξ 3 ] σ 2 2 + O ( σ 2 4 ) . (83)
Therefore, we have
τ ^ 00 U E ≃ 2 F S W , τ ^ i i U E ≃ 2 3 F S W , w r a w U E ≃ 1 3 . (84) \hat{\tau}^{\mathrm{UE}}_{00}\simeq2\mathcal{F}_{\mathrm{SW}},\qquad
\hat{\tau}^{\mathrm{UE}}_{ii}\simeq\frac{2}{3}\mathcal{F}_{\mathrm{SW}},\qquad
w_{\mathrm{raw}}^{\mathrm{UE}}\simeq\frac{1}{3}. \tag*{(84)} τ ^ 00 UE ≃ 2 F SW , τ ^ ii UE ≃ 3 2 F SW , w raw UE ≃ 3 1 . (84)
The leading coefficients coincide with the longitudinal and spatially-flat results.
For ϵ , ∣ δ ∣ ≪ σ 2 2 ≪ 1 \epsilon,|\delta|\ll\sigma_{2}^{2}\ll1 ϵ , ∣ δ ∣ ≪ σ 2 2 ≪ 1 , Eqs. (81) and (82) give
τ ^ 00 U E F S W ≃ 2 + σ 2 2 + ( 3 δ + 7 ϵ ) σ 2 2 + ( 3 δ − 36 ϵ ) σ 2 4 + 234 ϵ σ 2 6 + 108 ϵ σ 2 8 + O S R ( 2 ; σ 2 ) , (85) \frac{\hat{\tau}^{\mathrm{UE}}_{00}}{\mathcal{F}_{\mathrm{SW}}}
\simeq2+\sigma_{2}^{2}+(3\delta+7\epsilon)\sigma_{2}^{2}+(3\delta-36\epsilon)\sigma_{2}^{4}+234\epsilon\sigma_{2}^{6}+108\epsilon\sigma_{2}^{8}+\mathcal{O}_{\mathrm{SR}}(2;\sigma_{2}), \tag*{(85)} F SW τ ^ 00 UE ≃ 2 + σ 2 2 + ( 3 δ + 7 ϵ ) σ 2 2 + ( 3 δ − 36 ϵ ) σ 2 4 + 234 ϵ σ 2 6 + 108 ϵ σ 2 8 + O SR ( 2 ; σ 2 ) , (85)
τ ^ i i U E F S W ≃ 2 3 − 1 3 σ 2 2 − δ + ϵ 3 σ 2 2 + ( − 3 δ + 36 ϵ ) σ 2 4 − 390 ϵ σ 2 6 − 252 ϵ σ 2 8 + O S R ( 2 ; σ 2 ) . (86) \frac{\hat{\tau}^{\mathrm{UE}}_{ii}}{\mathcal{F}_{\mathrm{SW}}}
\simeq\frac{2}{3}-\frac{1}{3}\sigma_{2}^{2}-\frac{\delta+\epsilon}{3}\sigma_{2}^{2}+(-3\delta+36\epsilon)\sigma_{2}^{4}-390\epsilon\sigma_{2}^{6}-252\epsilon\sigma_{2}^{8}+\mathcal{O}_{\mathrm{SR}}(2;\sigma_{2}). \tag*{(86)} F SW τ ^ ii UE ≃ 3 2 − 3 1 σ 2 2 − 3 δ + ϵ σ 2 2 + ( − 3 δ + 36 ϵ ) σ 2 4 − 390 ϵ σ 2 6 − 252 ϵ σ 2 8 + O SR ( 2 ; σ 2 ) . (86)
Thus the leading slow-roll-independent structures are
τ ^ 00 U E F S W ≃ 2 + σ 2 2 + ⋯ , τ ^ i i U E F S W ≃ 2 3 − 1 3 σ 2 2 + ⋯ . (87) \frac{\hat{\tau}^{\mathrm{UE}}_{00}}{\mathcal{F}_{\mathrm{SW}}}\simeq2+\sigma_{2}^{2}+\cdots,\qquad
\frac{\hat{\tau}^{\mathrm{UE}}_{ii}}{\mathcal{F}_{\mathrm{SW}}}\simeq\frac{2}{3}-\frac{1}{3}\sigma_{2}^{2}+\cdots. \tag*{(87)} F SW τ ^ 00 UE ≃ 2 + σ 2 2 + ⋯ , F SW τ ^ ii UE ≃ 3 2 − 3 1 σ 2 2 + ⋯ . (87)
They coincide with the longitudinal and spatially-flat results at this displayed order. The UV expansion has no analogue of the infrared nonuniformity. Indeed,
k 2 + 3 ϵ H 2 = k 2 ( 1 + 3 ϵ σ 2 2 ) , (88) k^2+3\epsilon\mathcal{H}^2=k^2\left(1+3\epsilon\sigma_2^2\right), \tag*{(88)} k 2 + 3 ϵ H 2 = k 2 ( 1 + 3 ϵ σ 2 2 ) , (88)
and 3 ϵ σ 2 2 ≪ 1 3\epsilon\sigma_2^2\ll1 3 ϵ σ 2 2 ≪ 1 throughout both ultraviolet regimes.
Perturbative validity
In the ultraviolet, Eq. (56) gives parametrically
Q U E = O ( Ψ k ) . (89) Q_{\mathrm{UE}}=\mathcal{O}\left(\frac{\Psi}{k}\right). \tag*{(89)} Q UE = O ( k Ψ ) . (89)
Consequently,
∣ ψ U E ∣ , ∣ α U E ∣ , ∣ k β U E ∣ = O ( ∣ Ψ ∣ ) . (90) \lvert\psi_{\mathrm{UE}}\rvert,\lvert\alpha_{\mathrm{UE}}\rvert,\lvert k\beta_{\mathrm{UE}}\rvert=\mathcal{O}(\lvert\Psi\rvert). \tag*{(90)} ∣ ψ UE ∣ , ∣ α UE ∣ , ∣ k β UE ∣ = O (∣ Ψ ∣) . (90)
A necessary small-metric condition is therefore
∣ Ψ ∣ ≪ 1. (91) \lvert\Psi\rvert\ll1. \tag*{(91)} ∣ Ψ ∣ ≪ 1. (91)
Unlike the spatially-flat, comoving, and uniform-density gauges, the UE gauge introduces no additional wavelength-dependent metric lower bound in the UV. This condition ensures the smallness of the first-order metric perturbations, but does not by itself guarantee the validity of the higher-order perturbative expansion. For the intermediate UV, Eq. (91) must also be accompanied by
ϵ , ∣ δ ∣ ≪ σ 2 ≪ 1. (92) \sqrt{\epsilon},\sqrt{\lvert\delta\rvert}\ll\sigma_2\ll1. \tag*{(92)} ϵ , ∣ δ ∣ ≪ σ 2 ≪ 1. (92)
The comparison with the longitudinal, spatially-flat, comoving, and uniform-density gauges uses the results of Ref. [13 ] . Those results employ the same background prescription used in the present UE calculation. The long-wavelength results are summarized first in Table I .
Table I. Leading gauge-fixed scalar 2EMTs in the strict and intermediate infrared regimes. The strict-IR C, UD, and UE entries retain their first nonzero gradient terms. Their displayed w r a w w_{\mathrm{raw}} w raw values denote the corresponding ratio limits. The UE intermediate entry applies only sufficiently deep in the gradient-dominated sector, where the inverse-σ 1 \sigma_1 σ 1 corrections are parametrically subleading.
The longitudinal and spatially-flat gauges retain slow-roll-suppressed constant terms. The comoving, uniform-density, and uniform-expansion sources instead vanish at exactly zero gradient (σ 1 = 0 \sigma_1=0 σ 1 = 0 ). Their first finite-gradient coefficients, however, are different. In particular, the UE coefficient pair is
( 8 ϵ , − 8 3 ϵ ) σ 1 2 F L W , (93) \left(\frac{8}{\epsilon},-\frac{8}{3\epsilon}\right)\sigma_1^2F_{\mathrm{LW}}, \tag*{(93)} ( ϵ 8 , − 3 ϵ 8 ) σ 1 2 F LW , (93)
whereas the C and UD pairs are respectively
( − 10 , 10 3 ) σ 1 2 F L W , ( − 18 , 6 ) σ 1 2 F L W . (94) \left(-10,\frac{10}{3}\right)\sigma_1^2F_{\mathrm{LW}},\qquad(-18,6)\sigma_1^2F_{\mathrm{LW}}. \tag*{(94)} ( − 10 , 3 10 ) σ 1 2 F LW , ( − 18 , 6 ) σ 1 2 F LW . (94)
Thus the common zero-gradient cancellation does not imply equality of the first finite-gradient source. The component-ratio limit w r a w → − 1 / 3 w_{\mathrm{raw}} \to-1/3 w raw → − 1/3 is also common to C, UD, and UE. This agreement should not be confused with agreement of the 2EMT, since both the signs and the slow-roll scaling of the coefficients differ.
In the intermediate infrared, the leading terms of the UE slow-roll-first expansion do not dominate throughout the interval ϵ , ∣ δ ∣ ≪ σ 1 2 ≪ 1 \epsilon, \lvert\delta\rvert\ll\sigma_1^2 \ll1 ϵ , ∣ δ ∣ ≪ σ 1 2 ≪ 1 . Only sufficiently deep in the gradient-dominated sector does its leading contribution reduce to the spatially-flat result. Closer to the transition, inverse powers of σ 1 \sigma_1 σ 1 promote nominally higher slow-roll terms. The complete rational UE expression must then be used. Thus, the intermediate UE result in Table I is applicable only sufficiently deep in the gradient-dominated regime, not throughout the entire interval.
The characteristic UE denominator k 2 + 3 ϵ H 2 k^2 + 3\epsilon\mathcal{H}^2 k 2 + 3 ϵ H 2 defines a structural transition scale,
k 2 ∼ 3 ϵ H 2 , σ 1 2 ∼ 3 ϵ . (95) k^2 \sim3\epsilon\mathcal{H}^2,\qquad\sigma_1^2 \sim3\epsilon. \tag*{(95)} k 2 ∼ 3 ϵ H 2 , σ 1 2 ∼ 3 ϵ . (95)
At this scale, the gradient and background terms in the UE constraint become comparable. However, this structural transition should be distinguished from the loss of accuracy of a finite slow-roll-first expansion in the intermediate IR. For example, comparing a leading contribution of order ϵ σ 1 2 \epsilon\sigma_1^2 ϵ σ 1 2 with a correction of order ϵ 2 / σ 1 4 \epsilon^2/\sigma_1^4 ϵ 2 / σ 1 4 shows that the latter remains subleading only when σ 1 6 ≫ ϵ \sigma_1^6 \gg\epsilon σ 1 6 ≫ ϵ , up to numerical coefficients. Thus, the truncated expansion may become inaccurate at larger values of σ 1 \sigma_1 σ 1 , well before the structural transition at σ 1 2 ∼ 3 ϵ \sigma_1^2 \sim3\epsilon σ 1 2 ∼ 3 ϵ is reached.
The short-wavelength comparison is in Table II .
Table II. Leading gauge-fixed scalar 2EMTs in the strict and intermediate ultraviolet regimes. The intermediate columns display the first slow-roll-independent finite-wavelength structures. The UE, longitudinal, and spatially-flat gauges share the same displayed terms.
The UE ultraviolet behavior is therefore qualitatively simple. Its leading strict-UV coefficients coincide with the longitudinal and spatially-flat values. The same agreement persists for the displayed slow-roll-independent intermediate corrections. There is no 1 / ϵ 1/\epsilon 1/ ϵ matter-clock enhancement, and no additional 1 / σ 2 2 1/\sigma_2^2 1/ σ 2 2 density-gradient enhancement.
For completeness, the necessary UV small-metric conditions for all five gauges are summarized in Table III .
Table III. Necessary UV small-metric conditions at fixed oscillation amplitude ∣ Ψ ∣ \lvert\Psi\rvert ∣ Ψ ∣ and fixed background parameters. These conditions ensure the smallness of the first-order metric perturbations, but do not by themselves guarantee the validity of the higher-order perturbative expansion.
In the intermediate UV, in addition to the conditions in Table III , each row must also satisfy
ϵ , ∣ δ ∣ ≪ σ 2 ≪ 1. (96) \sqrt{\epsilon},\sqrt{\lvert\delta\rvert}\ll\sigma_2 \ll1. \tag*{(96)} ϵ , ∣ δ ∣ ≪ σ 2 ≪ 1. (96)
For the spatially-flat, comoving, and uniform-density gauges, the UV small-metric conditions in Table III impose lower bounds on σ 2 \sigma_2 σ 2 at fixed nonzero ∣ Ψ ∣ \lvert\Psi\rvert ∣ Ψ ∣ and fixed background parameters. Thus, the strict-UV limit σ 2 → 0 \sigma_2 \to0 σ 2 → 0 eventually violates these conditions, and the corresponding coefficients in Table II should be regarded as formal asymptotic results. These results can be applied at finite wavelengths only where the small-metric conditions are satisfied.
The five gauges can be classified according to their slicing conditions:
geometrical gauges: L , S F , U E , matter-clock gauges: C , U D . (97) \begin{aligned}
&\text{geometrical gauges:}\quad\mathrm{L},\ \mathrm{SF},\ \mathrm{UE},\\
&\text{matter-clock gauges:}\quad\mathrm{C},\ \mathrm{UD}. \tag*{(97)}
\end{aligned} geometrical gauges: L , SF , UE , matter-clock gauges: C , UD . (97)
The longitudinal (L), spatially-flat (SF), and uniform-expansion (UE) gauges are defined by vanishing scalar shear, spatial curvature perturbation, and local expansion perturbation, respectively. In contrast, the comoving (C) and uniform-density (UD) gauges use the inflaton and energy density as matter clocks.
The L, SF, and UE gauges exhibit the same leading 2EMT coefficients in the gradient-dominated intermediate IR and strict UV regimes, as well as the same displayed finite-wavelength corrections in the intermediate UV. This agreement may be related to their common geometrical character. However, their strict-IR behaviors differ. The L and SF sources retain slow-roll-suppressed zero-gradient terms, whereas the UE source vanishes at zero gradient and its first nonzero coefficients are enhanced by 1 / ϵ 1/\epsilon 1/ ϵ .
Unlike L and SF, UE uses the local expansion scalar as a geometrical clock, which evolves slowly during inflation, Θ ˙ 0 = − 3 ϵ H 2 \dot{\Theta}_0=-3\epsilon H^2 Θ ˙ 0 = − 3 ϵ H 2 . This may be related to the 1 / ϵ 1/\epsilon 1/ ϵ enhancement of its first nonzero strict-IR coefficients, analogous to, but distinct from, the matter-clock enhancements in C and UD. Although Q U E → Q C Q_{\mathrm{UE}}\to Q_{\mathrm{C}} Q UE → Q C as k → 0 k\to0 k → 0 , their first finite-gradient 2EMT coefficients differ.
Scale evolution of a fixed mode
The hierarchy has a direct dynamical interpretation for a fixed comoving mode. At Hubble crossing, k = a ∗ H ∗ k=a_*H_* k = a ∗ H ∗ and σ 1 ∗ = 1 \sigma_{1*}=1 σ 1 ∗ = 1 . Since
d ln ( a H ) d N = 1 − ϵ , (98) \frac{d\ln(aH)}{dN}=1-\epsilon, \tag*{(98)} d N d ln ( a H ) = 1 − ϵ , (98)
the mode evolves as
σ 1 ( Δ N ) = exp [ − ∫ N ∗ N ∗ + Δ N ( 1 − ϵ ) d N ] ≃ e − ( 1 − ϵ ) Δ N ≃ e − Δ N , (99) \sigma_1(\Delta N)=\exp\left[-\int_{N_*}^{N_*+\Delta N}(1-\epsilon)dN\right]\simeq e^{-(1-\epsilon)\Delta N}\simeq e^{-\Delta N}, \tag*{(99)} σ 1 ( Δ N ) = exp [ − ∫ N ∗ N ∗ + Δ N ( 1 − ϵ ) d N ] ≃ e − ( 1 − ϵ ) Δ N ≃ e − Δ N , (99)
for slowly varying ϵ \epsilon ϵ .
The denominator scale σ 1 2 ∼ 3 ϵ \sigma_1^2\sim3\epsilon σ 1 2 ∼ 3 ϵ is reached, for approximately constant ϵ \epsilon ϵ , after
Δ N d e n ≃ 1 2 ( 1 − ϵ ) ln ( 1 3 ϵ ) . (100) \Delta N_{\mathrm{den}}\simeq\frac{1}{2(1-\epsilon)}\ln\left(\frac{1}{3\epsilon}\right). \tag*{(100)} Δ N den ≃ 2 ( 1 − ϵ ) 1 ln ( 3 ϵ 1 ) . (100)
For ϵ = 10 − 2 , 10 − 3 , 10 − 4 \epsilon=10^{-2},10^{-3},10^{-4} ϵ = 1 0 − 2 , 1 0 − 3 , 1 0 − 4 , this gives approximately 1.8, 2.9, and 4.1 e-folds, respectively. Therefore, the long-wavelength regime σ 1 ≪ 1 \sigma_1\ll1 σ 1 ≪ 1 is reached within only a few e-folds after horizon crossing.
The slow-roll-first intermediate expansion need not remain uniform all the way to this scale. For example, the correction in Eq. (75) becomes important parametrically when σ 1 2 \sigma_1^2 σ 1 2 is no longer large relative to ϵ \epsilon ϵ . This criterion concerns the reliability of that truncation; it is not a second physical crossover scale.
A fixed mode therefore evolves schematically as
Hubble exit / gradient dominated ⟶ deep intermediate UE, SF-like at leading order ⟶ nonuniform transition region ⟶ strict-IR UE finite-gradient sector. (101) \boxed{\begin{gathered}
\text{Hubble exit / gradient dominated}\\
\longrightarrow\ \text{deep intermediate UE, SF-like at leading order}\\
\longrightarrow\ \text{nonuniform transition region}\\
\longrightarrow\ \text{strict-IR UE finite-gradient sector.}
\end{gathered}} \tag*{(101)} Hubble exit / gradient dominated ⟶ deep intermediate UE, SF-like at leading order ⟶ nonuniform transition region ⟶ strict-IR UE finite-gradient sector. (101)
The final strict-IR UE sector has the enhanced coefficient σ 1 2 / ϵ \sigma_1^2/\epsilon σ 1 2 / ϵ , while still vanishing as σ 1 → 0 \sigma_1\to0 σ 1 → 0 .
Interpretation and observability
The UE calculation has three structural features. First, the temporal gauge condition is differential for Q Q Q in configuration space, but algebraic for each Fourier mode. Second, the denominator k 2 + 3 ϵ H 2 k^2+3\epsilon H^2 k 2 + 3 ϵ H 2 makes the wavelength-first and slow-roll-first expansions genuinely noncommuting. Third, in the strict infrared, the zero-gradient UE source vanishes, while its first finite-gradient term is enhanced by 1 / ϵ 1/\epsilon 1/ ϵ .
The zero-gradient behavior requires a careful interpretation. The UE shift Q U E Q_{\mathrm{UE}} Q UE tends to the comoving shift Q C Q_{\mathrm{C}} Q C as k → 0 k\to0 k → 0 , but this equality does not fix the first gradient coefficient of the quadratic source. The C, UD, and UE sources all vanish at exactly zero gradient, yet their leading σ 1 2 \sigma_1^2 σ 1 2 coefficients are different. The strict-IR ratio w r a w → − 1 / 3 w_{\mathrm{raw}}\to-1/3 w raw → − 1/3 therefore carries much less information than the individual components.
The intermediate infrared also illustrates why asymptotic ordering matters. At formal lowest slow-roll order, the UE source has the same leading gradient structure as the spatially-flat source. This correspondence is reliable only sufficiently deep in the gradient-dominated sector. Toward the transition, higher slow-roll terms with inverse powers of σ 1 \sigma_1 σ 1 become comparable to, or larger than, the nominal leading term. The full rational result, rather than a finite reordered series, then controls the UE source.
In the ultraviolet, the UE behavior is more direct. The UE metric perturbations remain of order ∣ Ψ ∣ |\Psi| ∣Ψ∣ , so the necessary small-metric condition is simply ∣ Ψ ∣ ≪ 1 |\Psi| \ll1 ∣Ψ∣ ≪ 1 . The leading UE 2EMT coincides with the longitudinal and spatially-flat result, without the matter-clock 1 / ϵ 1/\epsilon 1/ ϵ enhancement of the comoving gauge or the additional density-gradient enhancement of the uniform-density gauge.
These results also clarify the status of Q Q Q . In the unreduced derivative form, Q Q Q remains because the UE condition is differential in configuration space. In Fourier space, Eq. (56) eliminates Q Q Q for each mode. The resulting expression may then be written in terms of the gauge-invariant potential Ψ \Psi Ψ . This is a useful representation, but it does not by itself assign an additional gauge-invariance property to the gauge-fixed 2EMT. Equivalent gauge-fixed variables describe the same selected representation.
The gauge-fixed 2EMT should also not be confused with a fully specified observable. A physical measurement requires an observer congruence, a physical hypersurface, and a measurement or coarse-graining prescription. An observer-centered construction could involve quantities such as τ ^ μ ν ( 2 ) u μ u ν \hat{\tau}^{(2)}_{\mu\nu}u^\mu u^\nu τ ^ μν ( 2 ) u μ u ν together with clock and observer corrections. A unique relational observable lies beyond the present work. Here we isolate the gauge-fixed quadratic source and its scale-dependent structure.
Conclusions
We have calculated the scalar quadratic effective energy-momentum tensor on uniform-expansion hypersurfaces during slow-roll inflation. The calculation first uses the exact mutually consistent background identities, and only afterward applies the slow-roll truncation. This prescription is the same as that used in the four-gauge comparison of [13 ] .
The complete scalar UE gauge is defined by κ = 0 \kappa= 0 κ = 0 and E = 0 E = 0 E = 0 . Its temporal constraint is differential in configuration space, whereas in Fourier space
Q U E ( k ) = − 3 ( Ψ ′ + H Ψ ) k 2 + 3 ϵ H 2 . Q_{\mathrm{UE}}(k) = -\frac{3(\Psi' + \mathcal{H}\Psi)}{k^2 + 3\epsilon\mathcal{H}^2}. Q UE ( k ) = − k 2 + 3 ϵ H 2 3 ( Ψ ′ + H Ψ ) .
The denominator k 2 + 3 ϵ H 2 k^2 + 3\epsilon\mathcal{H}^2 k 2 + 3 ϵ H 2 produces inequivalent wavelength-first and slow-roll-first asymptotic expansions.
In the strict infrared, the UE source has no zero-gradient term. Its first nonzero contributions are
τ ^ 00 U E F L W ≃ 8 ϵ σ 1 2 , τ ^ i i U E F L W ≃ − 8 3 ϵ σ 1 2 , (102) \frac{\hat{\tau}^{\mathrm{UE}}_{00}}{\mathcal{F}_{\mathrm{LW}}} \simeq\frac{8}{\epsilon}\sigma_1^2,\qquad\frac{\hat{\tau}^{\mathrm{UE}}_{ii}}{\mathcal{F}_{\mathrm{LW}}} \simeq-\frac{8}{3\epsilon}\sigma_1^2, \tag*{(102)} F LW τ ^ 00 UE ≃ ϵ 8 σ 1 2 , F LW τ ^ ii UE ≃ − 3 ϵ 8 σ 1 2 , (102)
at leading slow-roll order. Thus w r a w U E → − 1 / 3 w_{\mathrm{raw}}^{\mathrm{UE}} \to-1/3 w raw UE → − 1/3 as σ 1 → 0 \sigma_1 \to0 σ 1 → 0 . The comoving and uniform-density sources also vanish at zero gradient, but their first gradient coefficients are different. Consequently, the equality Q U E → Q C Q_{\mathrm{UE}} \to Q_{\mathrm{C}} Q UE → Q C at zero gradient does not imply equality of the corresponding finite-gradient 2EMTs.
In the intermediate infrared, the formal lowest slow-roll contribution is
τ ^ 00 U E ≃ ϵ ( σ 1 2 + σ 1 4 ) F L W , τ ^ i i U E ≃ ϵ ( − 1 3 σ 1 2 + σ 1 4 ) F L W . (103) \hat{\tau}^{\mathrm{UE}}_{00} \simeq\epsilon(\sigma_1^2 + \sigma_1^4)\mathcal{F}_{\mathrm{LW}},\qquad\hat{\tau}^{\mathrm{UE}}_{ii} \simeq\epsilon\left(-\frac{1}{3}\sigma_1^2 + \sigma_1^4\right)\mathcal{F}_{\mathrm{LW}}. \tag*{(103)} τ ^ 00 UE ≃ ϵ ( σ 1 2 + σ 1 4 ) F LW , τ ^ ii UE ≃ ϵ ( − 3 1 σ 1 2 + σ 1 4 ) F LW . (103)
Only sufficiently deep in the gradient-dominated sector do the leading terms reduce to the spatially-flat form with w r a w U E ≃ − 1 / 3 w_{\mathrm{raw}}^{\mathrm{UE}} \simeq-1/3 w raw UE ≃ − 1/3 . As the transition is approached, inverse powers of σ 1 \sigma_1 σ 1 promote higher slow-roll orders. No finite slow-roll-first truncation is uniformly reliable there, and the full rational expressions must be used.
In the strict ultraviolet, the UE source becomes
τ ^ 00 U E ≃ 2 F S W , τ ^ i i U E ≃ 2 3 F S W , w r a w U E ≃ 1 3 . (104) \hat{\tau}^{\mathrm{UE}}_{00} \simeq2\mathcal{F}_{\mathrm{SW}},\qquad\hat{\tau}^{\mathrm{UE}}_{ii} \simeq\frac{2}{3}\mathcal{F}_{\mathrm{SW}},\qquad w_{\mathrm{raw}}^{\mathrm{UE}} \simeq\frac{1}{3}. \tag*{(104)} τ ^ 00 UE ≃ 2 F SW , τ ^ ii UE ≃ 3 2 F SW , w raw UE ≃ 3 1 . (104)
The leading intermediate-UV finite-wavelength structure also agrees with the longitudinal and spatially-flat gauges. The UE reconstructed metric requires only the necessary condition ∣ Ψ ∣ ≪ 1 |\Psi| \ll1 ∣Ψ∣ ≪ 1 , with no additional wavelength-dependent UV lower bound.
The five-gauge comparison reveals a regime-dependent pattern associated with the choice of slicing. The geometrical gauges L, SF, and UE share the leading 2EMT coefficients in the gradient-dominated intermediate IR and strict UV, as well as the displayed finite-wavelength corrections in the intermediate UV. However, their strict-IR behaviors differ. The 1 / ϵ 1/\epsilon 1/ ϵ enhancement of the first nonzero UE coefficients may be related to the slow evolution of its geometrical expansion clock, analogous to, but distinct from, the matter-clock effects in C and UD. Although Q U E → Q C Q_{\mathrm{UE}} \to Q_{\mathrm{C}} Q UE → Q C as k → 0 k \to0 k → 0 , their first finite-gradient 2EMT coefficients differ. The UE constraint also defines a structural transition scale k 2 ∼ 3 ϵ H 2 k^2 \sim3\epsilon\mathcal{H}^2 k 2 ∼ 3 ϵ H 2 , distinct from the scale at which a truncated intermediate-IR expansion may lose accuracy.
We examine the gauge dependence of the general scalar quadratic 2EMT before imposing any gauge condition. Under an arbitrary scalar gauge transformation, its 00 and diagonal spatial components acquire nonvanishing variations. In the Fourier representation, the spatial integration of quadratic products pairs each mode k \mathbf{k} k with − k -\mathbf{k} − k , which we refer to as Fourier-mode pairing. Under symmetric momentum integration, terms odd in k \mathbf{k} k vanish, whereas even terms generally survive. The calculated gauge variations contain nonvanishing even-in-k k k contributions, demonstrating that the general scalar 2EMT remains gauge dependent after Fourier-mode pairing. This explicit transformation test thus settles the gauge-invariance question for the general scalar quadratic 2EMT considered here, a long-standing issue in the cosmological backreaction literature [5 , 6 , 7 , 8 , 9 ] .
In contrast, complete UE gauge fixing removes the residual scalar gauge freedom and selects a unique representative of each gauge orbit under the assumed boundary conditions. This uniqueness does not imply gauge invariance of the 2EMT.
Finally, a gauge-fixed 2EMT is not by itself a fully specified observable. A relational observable requires additional physical information about the hypersurface, observer, and measurement prescription. The present results provide the UE source and its scale-dependent asymptotic structure for such future constructions.
The gauge classification also suggests a natural next step. In analogy with comparisons within the matter-clock sector, the UE slicing can be compared with another geometrically defined clock. A particularly natural counterpart is uniform spacetime curvature, δ R = 0 \delta R=0 δ R = 0 , where R R R is the four-dimensional Ricci scalar. Such a comparison can test whether the scale-transition structure found here is specific to local expansion or is more general among geometrical-clock gauges.
Acknowledgments
The author is grateful to Jaichan Hwang for very helpful discussions. The author used ChatGPT (OpenAI) as an assistive tool during the analysis and preparation of this manuscript. All AI-assisted content was critically evaluated and verified by the author, who takes full responsibility for the scientific content and conclusions.
General scalar-gauge variation of the 2EMT before gauge fixing
We start from the general scalar quadratic source T μ ν ( 2 , e f f ) T_{\mu\nu}^{(2,\mathrm{eff})} T μν ( 2 , eff ) underlying Eqs. (106) and (A4) of the uniform-density companion paper Ref. [13 ] . We apply the arbitrary scalar transformation ξ μ = ( ξ 0 , ∂ i ζ ) \xi^\mu=(\xi^0,\partial^i\zeta) ξ μ = ( ξ 0 , ∂ i ζ ) directly to the perturbation variables in configuration space, substitute the transformed variables into the general quadratic source before gauge fixing, and subtract the original expression. No gauge condition or residual-gauge restriction is imposed. We use V 1 ≡ V ϕ ( ϕ 0 ) V_1\equiv V_\phi(\phi_0) V 1 ≡ V ϕ ( ϕ 0 ) and V 2 ≡ V ϕ ϕ ( ϕ 0 ) V_2\equiv V_{\phi\phi}(\phi_0) V 2 ≡ V ϕϕ ( ϕ 0 ) .
In simplifying the resulting expressions, contributions that are odd under reversal of the integration momentum are omitted, since they vanish under symmetric momentum integration. In the present calculation, the total-spatial-derivative contributions that are removed in this way belong to this odd sector. The expressions displayed below retain nonvanishing even-in-k k k contributions. Since an odd contribution cannot cancel a nonzero even contribution identically, these surviving terms are sufficient to establish the nonvanishing general gauge variation discussed in Sec. III A. We use the symbol ≡ \equiv ≡ below to denote equality after omission of these k k k -odd contributions. The spatial off-diagonal terms require an additional isotropic angular integration and are discussed separately in Sec. III A. The resulting 00 and diagonal spatial variations are (with no sum over i i i )
Δ ξ δ 00 = 2 H [ 2 H ∇ ζ ⋅ ∇ E ′ + ( 2 H ∇ ζ + ∇ ξ 0 ) ⋅ ( ∇ Q ′ − ∇ ξ 0 ′ ) ] − 4 H ∇ ξ 0 ⋅ ( H ′ ∇ ζ + H ∇ ζ ′ ) + 2 H ∇ Q ⋅ ( 2 H ′ ∇ ζ + 2 H ∇ ζ ′ + ∇ ξ 0 ′ ) + 4 H ∇ E ⋅ ( H ′ ∇ ξ 0 + H ∇ ξ 0 ′ ) + ∇ 2 ζ [ 2 H ( ∇ 2 ζ − ∇ 2 E ) − ( H ′ + 2 H 2 ) ζ ′ ] + ∇ 2 E ′ [ − 2 H ∇ 2 ζ + ( H ′ + 2 H 2 ) ζ ′ ] + ( H ′ + 2 H 2 ) E ′ ∇ 2 ζ − ( H ′ + 2 H 2 ) ( ξ 0 ′ ) 2 + 2 ξ 0 ξ 0 ′ [ − H H ′ + 4 H 3 + 4 π G a 2 V 1 ϕ 0 ′ ] + ( ξ 0 ) 2 [ − 3 H ′ 2 + 8 H 2 H ′ + 4 H 4 + a 2 V 2 ( H 2 − H ′ ) + 4 π G a 2 V 1 2 + 32 π G a 2 V 1 H ϕ 0 ′ ] + 2 Ψ [ − ∇ 4 ζ − 2 H ∇ 2 ζ ′ + 2 H ∇ 2 ξ 0 + ∇ 2 ξ 0 ′ + a 2 V 1 ϕ 0 ′ ∇ 2 ξ 0 + ξ 0 ( H ′ + 8 H 2 − a 2 V 1 H ϕ 0 ′ ) ] − ξ 0 ( − 4 H H ′ + 4 H 3 + a 2 V 2 H + a 2 V 1 ( H ′ + 6 H 2 ) ϕ 0 ′ + a 4 V 1 2 H ϕ 0 ′ 2 + 8 π G a 2 V 1 ϕ 0 ′ ) + 2 Ψ ′ [ 2 H ∇ 2 ζ − ∇ 2 ζ ′ + ∇ 2 ξ 0 − ξ 0 ( 3 H ′ + a 2 V 2 + 7 a 2 V 1 H ϕ 0 ′ + a 4 V 1 2 ϕ 0 ′ 2 ) − a 2 V 1 ϕ 0 ′ ξ 0 ′ ] + 2 Q ′ [ ( H ′ + 2 H 2 ) ξ 0 ′ − ξ 0 ( − H H ′ + 4 H 3 + 4 π G a 2 V 1 ϕ 0 ′ ) ] \begin{aligned}
\Delta_\xi\delta_{00}={}&2\mathcal{H}\left[2\mathcal{H}\nabla\zeta\cdot\nabla E'+(2\mathcal{H}\nabla\zeta+\nabla\xi^0)\cdot(\nabla Q'-\nabla\xi^{0\prime})\right]-4\mathcal{H}\nabla\xi^0\cdot(\mathcal{H}'\nabla\zeta+\mathcal{H}\nabla\zeta')\\
&+2\mathcal{H}\nabla Q\cdot(2\mathcal{H}'\nabla\zeta+2\mathcal{H}\nabla\zeta'+\nabla\xi^{0\prime})+4\mathcal{H}\nabla E\cdot(\mathcal{H}'\nabla\xi^0+\mathcal{H}\nabla\xi^{0\prime})\\
&+\nabla^2\zeta\left[2\mathcal{H}(\nabla^2\zeta-\nabla^2E)-(\mathcal{H}'+2\mathcal{H}^2)\zeta'\right]+\nabla^2E'\left[-2\mathcal{H}\nabla^2\zeta+(\mathcal{H}'+2\mathcal{H}^2)\zeta'\right]+(\mathcal{H}'+2\mathcal{H}^2)E'\nabla^2\zeta\\
&-(\mathcal{H}'+2\mathcal{H}^2)(\xi^{0\prime})^2+2\xi^0\xi^{0\prime}\left[-\mathcal{H}\mathcal{H}'+4\mathcal{H}^3+4\pi Ga^2V_1\phi_0'\right]\\
&+(\xi^0)^2\left[-3\mathcal{H}'^2+8\mathcal{H}^2\mathcal{H}'+4\mathcal{H}^4+a^2V_2(\mathcal{H}^2-\mathcal{H}')+4\pi Ga^2V_1^2+32\pi Ga^2V_1\mathcal{H}\phi_0'\right]\\
&+2\Psi\left[-\nabla^4\zeta-2\mathcal{H}\nabla^2\zeta'+2\mathcal{H}\nabla^2\xi^0+\nabla^2\xi^{0\prime}+\frac{a^2V_1}{\phi_0'}\nabla^2\xi^0+\xi^0\left(\mathcal{H}'+8\mathcal{H}^2-\frac{a^2V_1\mathcal{H}}{\phi_0'}\right)\right]\\
&-\xi^0\left(-4\mathcal{H}\mathcal{H}'+4\mathcal{H}^3+a^2V_2\mathcal{H}+\frac{a^2V_1(\mathcal{H}'+6\mathcal{H}^2)}{\phi_0'}+\frac{a^4V_1^2\mathcal{H}}{\phi_0'^2}+8\pi Ga^2V_1\phi_0'\right)\\
&+2\Psi'\left[2\mathcal{H}\nabla^2\zeta-\nabla^2\zeta'+\nabla^2\xi^0-\xi^0\left(3\mathcal{H}'+a^2V_2+\frac{7a^2V_1\mathcal{H}}{\phi_0'}+\frac{a^4V_1^2}{\phi_0'^2}\right)-\frac{a^2V_1}{\phi_0'}\xi^{0\prime}\right]\\
&+2Q'\left[(\mathcal{H}'+2\mathcal{H}^2)\xi^{0\prime}-\xi^0(-\mathcal{H}\mathcal{H}'+4\mathcal{H}^3+4\pi Ga^2V_1\phi_0')\right]
\end{aligned} Δ ξ δ 00 = 2 H [ 2 H ∇ ζ ⋅ ∇ E ′ + ( 2 H ∇ ζ + ∇ ξ 0 ) ⋅ ( ∇ Q ′ − ∇ ξ 0′ ) ] − 4 H ∇ ξ 0 ⋅ ( H ′ ∇ ζ + H ∇ ζ ′ ) + 2 H ∇ Q ⋅ ( 2 H ′ ∇ ζ + 2 H ∇ ζ ′ + ∇ ξ 0′ ) + 4 H ∇ E ⋅ ( H ′ ∇ ξ 0 + H ∇ ξ 0′ ) + ∇ 2 ζ [ 2 H ( ∇ 2 ζ − ∇ 2 E ) − ( H ′ + 2 H 2 ) ζ ′ ] + ∇ 2 E ′ [ − 2 H ∇ 2 ζ + ( H ′ + 2 H 2 ) ζ ′ ] + ( H ′ + 2 H 2 ) E ′ ∇ 2 ζ − ( H ′ + 2 H 2 ) ( ξ 0′ ) 2 + 2 ξ 0 ξ 0′ [ − H H ′ + 4 H 3 + 4 π G a 2 V 1 ϕ 0 ′ ] + ( ξ 0 ) 2 [ − 3 H ′2 + 8 H 2 H ′ + 4 H 4 + a 2 V 2 ( H 2 − H ′ ) + 4 π G a 2 V 1 2 + 32 π G a 2 V 1 H ϕ 0 ′ ] + 2Ψ [ − ∇ 4 ζ − 2 H ∇ 2 ζ ′ + 2 H ∇ 2 ξ 0 + ∇ 2 ξ 0′ + ϕ 0 ′ a 2 V 1 ∇ 2 ξ 0 + ξ 0 ( H ′ + 8 H 2 − ϕ 0 ′ a 2 V 1 H ) ] − ξ 0 ( − 4 H H ′ + 4 H 3 + a 2 V 2 H + ϕ 0 ′ a 2 V 1 ( H ′ + 6 H 2 ) + ϕ 0 ′2 a 4 V 1 2 H + 8 π G a 2 V 1 ϕ 0 ′ ) + 2 Ψ ′ [ 2 H ∇ 2 ζ − ∇ 2 ζ ′ + ∇ 2 ξ 0 − ξ 0 ( 3 H ′ + a 2 V 2 + ϕ 0 ′ 7 a 2 V 1 H + ϕ 0 ′2 a 4 V 1 2 ) − ϕ 0 ′ a 2 V 1 ξ 0′ ] + 2 Q ′ [ ( H ′ + 2 H 2 ) ξ 0′ − ξ 0 ( − H H ′ + 4 H 3 + 4 π G a 2 V 1 ϕ 0 ′ ) ]
− 2 Q [ ξ 0 ′ ′ ( − H H ′ + 4 H 3 + 4 π G a 2 V 1 ϕ 0 ′ ) + ξ 0 ( − 3 H ′ 2 + 8 H 2 H ′ + 4 H 4 + a 2 V 2 ( H 2 − H ′ ) + 4 π G a 4 V 1 2 + 32 π G a 2 V 1 H ϕ 0 ′ ) ] . (105) -2Q\left[\xi^{0\prime\prime}\left(-\mathcal{H}\mathcal{H}^\prime+4\mathcal{H}^3+4\pi G a^2V_1\phi_0^\prime\right)+\xi^0\left(-3\mathcal{H}^{\prime2}+8\mathcal{H}^2\mathcal{H}^\prime+4\mathcal{H}^4+a^2V_2(\mathcal{H}^2-\mathcal{H}^\prime)+4\pi G a^4V_1^2+32\pi G a^2V_1\mathcal{H}\phi_0^\prime\right)\right]. \tag*{(105)} − 2 Q [ ξ 0′′ ( − H H ′ + 4 H 3 + 4 π G a 2 V 1 ϕ 0 ′ ) + ξ 0 ( − 3 H ′2 + 8 H 2 H ′ + 4 H 4 + a 2 V 2 ( H 2 − H ′ ) + 4 π G a 4 V 1 2 + 32 π G a 2 V 1 H ϕ 0 ′ ) ] . (105)
Δ E S i i ≔ − 2 3 ( ∇ 2 ζ ′ ) 2 + 2 3 [ ( ∇ 2 E − ∇ 2 ζ ) ∇ 2 ζ ′ ′ + ∇ 2 ζ ∇ 2 E ′ ′ ] + 2 H ( ∇ 2 ζ ′ ′ − ∇ 2 E ′ ′ ) ζ ′ − E ′ [ 2 H ∇ 2 ζ ′ ′ + ( H ′ + 2 H 2 ) ∇ 2 ζ ′ ] + 1 3 ∇ 2 ζ ′ [ 4 H ( ∇ 2 E − ∇ 2 ζ ) + 3 ( H ′ + 2 H 2 ) ζ ′ ] + 1 3 ∇ 2 E ′ [ 4 H ∇ 2 ζ + 4 ∇ 2 ζ ′ + 8 H ξ 0 − 3 ( H ′ + 2 H 2 ) ζ ′ + 8 ( H ′ + H 2 ) ξ 0 ] + ( 2 H 2 − H ′ ) ( ξ 0 ) 2 − 8 3 ξ 0 ′ [ H ∇ 2 ζ ′ + ( H ′ + H 2 ) ( ∇ 2 ζ − ∇ 2 E ) + 3 ξ 0 ( 3 4 H H ′ + H 3 + 3 π G a 2 V 1 ϕ 0 ′ ) ] − 4 3 ξ 0 [ 2 ( H ′ + H 2 ) ∇ 2 ζ ′ + H ( ∇ 2 ζ ′ ′ − ∇ 2 E ′ ′ ) + 4 ( H 3 + 2 π G a 2 V 1 ϕ 0 ′ ) ( ∇ 2 ζ − ∇ 2 E ) ] + ( ξ 0 ) 2 [ − 3 H ′ 2 − 12 H 4 − a 2 V 2 ( H 2 − H ′ ) + 4 π G a 4 V 1 2 − 32 π G a 2 V 1 H ϕ 0 ′ ] + 2 3 ∇ ξ 0 ⋅ ( 2 H ∇ ξ 0 + ∇ ξ 0 ′ ) + 2 3 ∇ ξ 0 ⋅ ∇ ξ 0 ′ ′ + [ 4 3 H ( ∇ 2 E − ∇ 2 ζ ) + 2 H ξ 0 ′ − 2 H ′ ξ 0 ] ζ ′ ′ + Ψ ′ [ − 2 3 ∇ 2 ζ ′ − 2 ξ 0 ′ ′ + 8 3 ( 2 H + a 2 V 1 ϕ 0 ′ ) ∇ 2 ζ + 6 a 2 V 1 ϕ 0 ′ ξ 0 ′ + 2 ξ 0 ( a 2 V 2 − H ′ + a 2 V 1 H ϕ 0 ′ − a 4 V 1 2 ϕ 0 ′ 2 ) ] + Ψ [ − 4 3 ∇ 4 ζ + 2 3 ∇ 2 ζ ′ ′ + 4 3 H ∇ 2 ζ ′ − 2 ∇ 2 ξ 0 ′ + 8 3 ( H ′ + 2 H 2 + a 2 V 1 H ϕ 0 ′ ) ∇ 2 ζ + 2 a 2 V 1 ϕ 0 ′ ∇ 2 ξ 0 − 2 ξ 0 ( 3 H ′ + 4 H 2 − 3 a 2 V 1 H ϕ 0 ′ ) − 8 H ξ 0 ′ ′ + 2 ξ 0 ( − 4 H 3 + a 2 V 2 H + a 2 V 1 ( 2 H 2 − H ′ ) ϕ 0 ′ − a 4 V 1 2 H ϕ 0 ′ 2 − 8 π G a 2 V 1 ϕ 0 ′ ) ] − 4 3 ∇ Q ′ ⋅ ( H ∇ ξ 0 + ∇ ξ 0 ′ ) − 2 3 ∇ ξ 0 ⋅ ∇ Q ′ ′ − 2 3 ∇ Q ⋅ ( 2 H ∇ ξ 0 ′ + ∇ ξ 0 ′ ′ ) + ( 4 3 H ∇ 2 ζ − 2 H ξ 0 ′ + 2 H ′ ξ 0 ) Q ′ ′ + Q ′ [ 8 3 H ∇ 2 ζ ′ + 8 3 ( H ′ + H 2 ) ∇ 2 ζ − 2 ( 2 H 2 − H ′ ) ξ 0 ′ − 2 H ξ 0 ′ ′ + 2 ξ 0 ( 3 H H ′ + 4 H 3 + 12 π G a 2 V 1 ϕ 0 ′ ) ] + Q [ 8 3 ( H ′ + H 2 ) ∇ 2 ζ ′ + 4 3 H [ ( 4 H 2 + 8 π G a 2 V 1 ϕ 0 ′ ) ∇ 2 ζ + ∇ 2 ζ ′ ′ ] + 2 ξ 0 ′ ( 3 H H ′ + 4 H 3 + 12 π G a 2 V 1 ϕ 0 ′ ) + 2 H ξ 0 ′ ′ + 2 ξ 0 ( 3 H ′ 2 + 12 H 4 + a 2 V 2 ( H 2 − H ′ ) − 4 π G a 4 V 1 2 + 32 π G a 2 V 1 H ϕ 0 ′ ) ] . (106) \begin{aligned}
\Delta_{\mathcal{E}}S_{ii} \coloneqq{}&-\frac{2}{3}(\nabla^2\zeta^\prime)^2+\frac{2}{3}\left[(\nabla^2E-\nabla^2\zeta)\nabla^2\zeta^{\prime\prime}+\nabla^2\zeta\nabla^2E^{\prime\prime}\right]+2\mathcal{H}(\nabla^2\zeta^{\prime\prime}-\nabla^2E^{\prime\prime})\zeta^\prime\\
&-E^\prime\left[2\mathcal{H}\nabla^2\zeta^{\prime\prime}+(\mathcal{H}^\prime+2\mathcal{H}^2)\nabla^2\zeta^\prime\right]+\frac{1}{3}\nabla^2\zeta^\prime\left[4\mathcal{H}(\nabla^2E-\nabla^2\zeta)+3(\mathcal{H}^\prime+2\mathcal{H}^2)\zeta^\prime\right]\\
&+\frac{1}{3}\nabla^2E^\prime\left[4\mathcal{H}\nabla^2\zeta+4\nabla^2\zeta^\prime+8\mathcal{H}\xi^0-3(\mathcal{H}^\prime+2\mathcal{H}^2)\zeta^\prime+8(\mathcal{H}^\prime+\mathcal{H}^2)\xi^0\right]+(2\mathcal{H}^2-\mathcal{H}^\prime)(\xi^0)^2\\
&-\frac{8}{3}\xi^{0\prime}\left[\mathcal{H}\nabla^2\zeta^\prime+(\mathcal{H}^\prime+\mathcal{H}^2)(\nabla^2\zeta-\nabla^2E)+3\xi^0\left(\frac{3}{4}\mathcal{H}\mathcal{H}^\prime+\mathcal{H}^3+3\pi G a^2V_1\phi_0^\prime\right)\right]\\
&-\frac{4}{3}\xi^0\left[2(\mathcal{H}^\prime+\mathcal{H}^2)\nabla^2\zeta^\prime+\mathcal{H}(\nabla^2\zeta^{\prime\prime}-\nabla^2E^{\prime\prime})+4(\mathcal{H}^3+2\pi G a^2V_1\phi_0^\prime)(\nabla^2\zeta-\nabla^2E)\right]\\
&+(\xi^0)^2\left[-3\mathcal{H}^{\prime2}-12\mathcal{H}^4-a^2V_2(\mathcal{H}^2-\mathcal{H}^\prime)+4\pi G a^4V_1^2-32\pi G a^2V_1\mathcal{H}\phi_0^\prime\right]\\
&+\frac{2}{3}\nabla\xi^0\cdot(2\mathcal{H}\nabla\xi^0+\nabla\xi^{0\prime})+\frac{2}{3}\nabla\xi^0\cdot\nabla\xi^{0\prime\prime}+\left[\frac{4}{3}\mathcal{H}(\nabla^2E-\nabla^2\zeta)+2\mathcal{H}\xi^{0\prime}-2\mathcal{H}^\prime\xi^0\right]\zeta^{\prime\prime}\\
&+\Psi^\prime\left[-\frac{2}{3}\nabla^2\zeta^\prime-2\xi^{0\prime\prime}+\frac{8}{3}\left(2\mathcal{H}+\frac{a^2V_1}{\phi_0^\prime}\right)\nabla^2\zeta+\frac{6a^2V_1}{\phi_0^\prime}\xi^{0\prime}+2\xi^0\left(a^2V_2-\mathcal{H}^\prime+\frac{a^2V_1\mathcal{H}}{\phi_0^\prime}-\frac{a^4V_1^2}{\phi_0^{\prime2}}\right)\right]\\
&+\Psi\left[-\frac{4}{3}\nabla^4\zeta+\frac{2}{3}\nabla^2\zeta^{\prime\prime}+\frac{4}{3}\mathcal{H}\nabla^2\zeta^\prime-2\nabla^2\xi^{0\prime}+\frac{8}{3}\left(\mathcal{H}^\prime+2\mathcal{H}^2+\frac{a^2V_1\mathcal{H}}{\phi_0^\prime}\right)\nabla^2\zeta+\frac{2a^2V_1}{\phi_0^\prime}\nabla^2\xi^0\right.\\
&\left.\qquad-2\xi^0\left(3\mathcal{H}^\prime+4\mathcal{H}^2-\frac{3a^2V_1\mathcal{H}}{\phi_0^\prime}\right)-8\mathcal{H}\xi^{0\prime\prime}+2\xi^0\left(-4\mathcal{H}^3+a^2V_2\mathcal{H}+\frac{a^2V_1(2\mathcal{H}^2-\mathcal{H}^\prime)}{\phi_0^\prime}-\frac{a^4V_1^2\mathcal{H}}{\phi_0^{\prime2}}-8\pi G a^2V_1\phi_0^\prime\right)\right]\\
&-\frac{4}{3}\nabla Q^\prime\cdot(\mathcal{H}\nabla\xi^0+\nabla\xi^{0\prime})-\frac{2}{3}\nabla\xi^0\cdot\nabla Q^{\prime\prime}-\frac{2}{3}\nabla Q\cdot(2\mathcal{H}\nabla\xi^{0\prime}+\nabla\xi^{0\prime\prime})\\
&+\left(\frac{4}{3}\mathcal{H}\nabla^2\zeta-2\mathcal{H}\xi^{0\prime}+2\mathcal{H}^\prime\xi^0\right)Q^{\prime\prime}\\
&+Q^\prime\left[\frac{8}{3}\mathcal{H}\nabla^2\zeta^\prime+\frac{8}{3}(\mathcal{H}^\prime+\mathcal{H}^2)\nabla^2\zeta-2(2\mathcal{H}^2-\mathcal{H}^\prime)\xi^{0\prime}-2\mathcal{H}\xi^{0\prime\prime}+2\xi^0(3\mathcal{H}\mathcal{H}^\prime+4\mathcal{H}^3+12\pi G a^2V_1\phi_0^\prime)\right]\\
&+Q\left[\frac{8}{3}(\mathcal{H}^\prime+\mathcal{H}^2)\nabla^2\zeta^\prime+\frac{4}{3}\mathcal{H}\left[(4\mathcal{H}^2+8\pi G a^2V_1\phi_0^\prime)\nabla^2\zeta+\nabla^2\zeta^{\prime\prime}\right]+2\xi^{0\prime}(3\mathcal{H}\mathcal{H}^\prime+4\mathcal{H}^3+12\pi G a^2V_1\phi_0^\prime)+2\mathcal{H}\xi^{0\prime\prime}\right.\\
&\left.\qquad+2\xi^0\left(3\mathcal{H}^{\prime2}+12\mathcal{H}^4+a^2V_2(\mathcal{H}^2-\mathcal{H}^\prime)-4\pi G a^4V_1^2+32\pi G a^2V_1\mathcal{H}\phi_0^\prime\right)\right]. \tag*{(106)}
\end{aligned} Δ E S ii : = − 3 2 ( ∇ 2 ζ ′ ) 2 + 3 2 [ ( ∇ 2 E − ∇ 2 ζ ) ∇ 2 ζ ′′ + ∇ 2 ζ ∇ 2 E ′′ ] + 2 H ( ∇ 2 ζ ′′ − ∇ 2 E ′′ ) ζ ′ − E ′ [ 2 H ∇ 2 ζ ′′ + ( H ′ + 2 H 2 ) ∇ 2 ζ ′ ] + 3 1 ∇ 2 ζ ′ [ 4 H ( ∇ 2 E − ∇ 2 ζ ) + 3 ( H ′ + 2 H 2 ) ζ ′ ] + 3 1 ∇ 2 E ′ [ 4 H ∇ 2 ζ + 4 ∇ 2 ζ ′ + 8 H ξ 0 − 3 ( H ′ + 2 H 2 ) ζ ′ + 8 ( H ′ + H 2 ) ξ 0 ] + ( 2 H 2 − H ′ ) ( ξ 0 ) 2 − 3 8 ξ 0′ [ H ∇ 2 ζ ′ + ( H ′ + H 2 ) ( ∇ 2 ζ − ∇ 2 E ) + 3 ξ 0 ( 4 3 H H ′ + H 3 + 3 π G a 2 V 1 ϕ 0 ′ ) ] − 3 4 ξ 0 [ 2 ( H ′ + H 2 ) ∇ 2 ζ ′ + H ( ∇ 2 ζ ′′ − ∇ 2 E ′′ ) + 4 ( H 3 + 2 π G a 2 V 1 ϕ 0 ′ ) ( ∇ 2 ζ − ∇ 2 E ) ] + ( ξ 0 ) 2 [ − 3 H ′2 − 12 H 4 − a 2 V 2 ( H 2 − H ′ ) + 4 π G a 4 V 1 2 − 32 π G a 2 V 1 H ϕ 0 ′ ] + 3 2 ∇ ξ 0 ⋅ ( 2 H ∇ ξ 0 + ∇ ξ 0′ ) + 3 2 ∇ ξ 0 ⋅ ∇ ξ 0′′ + [ 3 4 H ( ∇ 2 E − ∇ 2 ζ ) + 2 H ξ 0′ − 2 H ′ ξ 0 ] ζ ′′ + Ψ ′ [ − 3 2 ∇ 2 ζ ′ − 2 ξ 0′′ + 3 8 ( 2 H + ϕ 0 ′ a 2 V 1 ) ∇ 2 ζ + ϕ 0 ′ 6 a 2 V 1 ξ 0′ + 2 ξ 0 ( a 2 V 2 − H ′ + ϕ 0 ′ a 2 V 1 H − ϕ 0 ′2 a 4 V 1 2 ) ] + Ψ [ − 3 4 ∇ 4 ζ + 3 2 ∇ 2 ζ ′′ + 3 4 H ∇ 2 ζ ′ − 2 ∇ 2 ξ 0′ + 3 8 ( H ′ + 2 H 2 + ϕ 0 ′ a 2 V 1 H ) ∇ 2 ζ + ϕ 0 ′ 2 a 2 V 1 ∇ 2 ξ 0 − 2 ξ 0 ( 3 H ′ + 4 H 2 − ϕ 0 ′ 3 a 2 V 1 H ) − 8 H ξ 0′′ + 2 ξ 0 ( − 4 H 3 + a 2 V 2 H + ϕ 0 ′ a 2 V 1 ( 2 H 2 − H ′ ) − ϕ 0 ′2 a 4 V 1 2 H − 8 π G a 2 V 1 ϕ 0 ′ ) ] − 3 4 ∇ Q ′ ⋅ ( H ∇ ξ 0 + ∇ ξ 0′ ) − 3 2 ∇ ξ 0 ⋅ ∇ Q ′′ − 3 2 ∇ Q ⋅ ( 2 H ∇ ξ 0′ + ∇ ξ 0′′ ) + ( 3 4 H ∇ 2 ζ − 2 H ξ 0′ + 2 H ′ ξ 0 ) Q ′′ + Q ′ [ 3 8 H ∇ 2 ζ ′ + 3 8 ( H ′ + H 2 ) ∇ 2 ζ − 2 ( 2 H 2 − H ′ ) ξ 0′ − 2 H ξ 0′′ + 2 ξ 0 ( 3 H H ′ + 4 H 3 + 12 π G a 2 V 1 ϕ 0 ′ ) ] + Q [ 3 8 ( H ′ + H 2 ) ∇ 2 ζ ′ + 3 4 H [ ( 4 H 2 + 8 π G a 2 V 1 ϕ 0 ′ ) ∇ 2 ζ + ∇ 2 ζ ′′ ] + 2 ξ 0′ ( 3 H H ′ + 4 H 3 + 12 π G a 2 V 1 ϕ 0 ′ ) + 2 H ξ 0′′ + 2 ξ 0 ( 3 H ′2 + 12 H 4 + a 2 V 2 ( H 2 − H ′ ) − 4 π G a 4 V 1 2 + 32 π G a 2 V 1 H ϕ 0 ′ ) ] . (106)
For completeness, we record the uniform-expansion expressions obtained from the general Fourier-space scalar 2EMT, Eqs. (A2) and (A4) of the uniform-density companion paper Ref. [13 ] , after imposing the two UE gauge conditions but before eliminating Q Q Q with the Fourier-space constraint. These formulas retain the derivative structure inside the Fourier-paired quadratic brackets. We use V 0 ≡ V ( ϕ 0 ) V_0\equiv V(\phi_0) V 0 ≡ V ( ϕ 0 ) , V 1 ≡ V ϕ ( ϕ 0 ) V_1\equiv V_\phi(\phi_0) V 1 ≡ V ϕ ( ϕ 0 ) , and V 2 ≡ V ϕ ϕ ( ϕ 0 ) V_2\equiv V_{\phi\phi}(\phi_0) V 2 ≡ V ϕϕ ( ϕ 0 ) .
τ ^ 00 ( 2 ) ( k ) = 1 H 2 − H ′ { ⟨ ( ∇ Ψ ′ ) 2 ⟩ + ⟨ ( Δ Ψ ) 2 ⟩ + ( 8 H + 2 a 2 V 1 ϕ 0 ′ ) ⟨ ∇ Ψ ⋅ ∇ Ψ ′ ⟩ (107) \widehat{\tau}_{00}^{(2)}(k)=\frac{1}{\mathcal{H}^2-\mathcal{H}^\prime}\left\{\left\langle(\nabla\Psi^\prime)^2\right\rangle+\left\langle(\Delta\Psi)^2\right\rangle+\left(8\mathcal{H}+\frac{2a^2V_1}{\phi_0^\prime}\right)\left\langle\nabla\Psi\cdot\nabla\Psi^\prime\right\rangle
\right. \tag*{(107)} τ 00 ( 2 ) ( k ) = H 2 − H ′ 1 { ⟨ ( ∇ Ψ ′ ) 2 ⟩ + ⟨ ( ΔΨ ) 2 ⟩ + ( 8 H + ϕ 0 ′ 2 a 2 V 1 ) ⟨ ∇Ψ ⋅ ∇ Ψ ′ ⟩ (107)
+ 2 a 2 V 1 ( H ′ + 5 H 2 ) ϕ 0 ′ + 10 a 2 V 1 ( H 2 − H ′ ) ϕ 0 ′ ⟨ Ψ Ψ ′ ⟩ + [ − 9 ( H 2 − H ′ ) + 4 a 4 G π V 1 2 H 2 − H ′ + a 2 V 2 + 9 H 2 + 6 a 2 V 1 H ϕ 0 ′ ] ⟨ ( Ψ ′ ) 2 ⟩ + ( 4 H ′ − 16 H 2 + 2 a 2 V 1 H ϕ 0 ′ ) ⟨ Ψ Q ′ ⟩ + ( 6 H + 2 a 2 V 1 ϕ 0 ′ ) ⟨ Ψ ′ Q ′ ⟩ + [ − 26 H H ′ + 8 a 2 V 1 H ′ ϕ 0 ′ + 48 a 2 G π V 0 H + 2 a 2 V 2 H + 8 H 3 + 8 a 4 G π V 1 2 H H 2 − H ′ + 6 a 2 V 1 H 2 ϕ 0 ′ + 64 a 2 G π V 1 ϕ 0 ′ ] ⟨ Ψ Q ⟩ + ( 12 H ′ + 2 a 2 V 2 − 6 H 2 + 8 a 4 G π V 1 2 H 2 − H ′ + 14 a 2 V 1 H ϕ 0 ′ ) ⟨ Ψ ′ Q ⟩ − ( H ′ + 2 H 2 ) ⟨ ( Q ′ ) 2 ⟩ + ( − 8 H H ′ + 14 H 3 + 8 a 2 G π V 1 ϕ 0 ′ ) ⟨ Q Q ′ ⟩ + ( − 3 H ′ 2 − a 2 V 2 H ′ + 8 H 2 H ′ + a 2 V 2 H 2 + 4 H 4 + 32 a 2 G π V 1 H ϕ 0 ′ + 4 a 4 G π V 1 2 ) ⟨ Q 2 ⟩ . (108) \begin{aligned}
&+\frac{2a^{2}V_{1}(\mathcal{H}'+5\mathcal{H}^{2})}{\phi'_0}+\frac{10a^{2}V_{1}(\mathcal{H}^{2}-\mathcal{H}')}{\phi'_0}\langle\Psi\Psi'\rangle\\
&+\left[-9(\mathcal{H}^{2}-\mathcal{H}')+\frac{4a^{4}G\pi V_{1}^{2}}{\mathcal{H}^{2}-\mathcal{H}'}+a^{2}V_{2}+9\mathcal{H}^{2}+\frac{6a^{2}V_{1}\mathcal{H}}{\phi'_0}\right]\langle(\Psi')^{2}\rangle\\
&+\left(4\mathcal{H}'-16\mathcal{H}^{2}+\frac{2a^{2}V_{1}\mathcal{H}}{\phi'_0}\right)\langle\Psi Q'\rangle+\left(6\mathcal{H}+\frac{2a^{2}V_{1}}{\phi'_0}\right)\langle\Psi'Q'\rangle\\
&+\left[-26\mathcal{H}\mathcal{H}'+\frac{8a^{2}V_{1}\mathcal{H}'}{\phi'_0}+48a^{2}G\pi V_{0}\mathcal{H}+2a^{2}V_{2}\mathcal{H}+8\mathcal{H}^{3}+\frac{8a^{4}G\pi V_{1}^{2}\mathcal{H}}{\mathcal{H}^{2}-\mathcal{H}'}\right.\\
&\left.\qquad+\frac{6a^{2}V_{1}\mathcal{H}^{2}}{\phi'_0}+64a^{2}G\pi V_{1}\phi'_0\right]\langle\Psi Q\rangle\\
&+\left(12\mathcal{H}'+2a^{2}V_{2}-6\mathcal{H}^{2}+\frac{8a^{4}G\pi V_{1}^{2}}{\mathcal{H}^{2}-\mathcal{H}'}+\frac{14a^{2}V_{1}\mathcal{H}}{\phi'_0}\right)\langle\Psi'Q\rangle\\
&-(\mathcal{H}'+2\mathcal{H}^{2})\langle(Q')^{2}\rangle+(-8\mathcal{H}\mathcal{H}'+14\mathcal{H}^{3}+8a^{2}G\pi V_{1}\phi'_0)\langle QQ'\rangle\\
&+\left(-3\mathcal{H}'^{2}-a^{2}V_{2}\mathcal{H}'+8\mathcal{H}^{2}\mathcal{H}'+a^{2}V_{2}\mathcal{H}^{2}+4\mathcal{H}^{4}+32a^{2}G\pi V_{1}\mathcal{H}\phi'_0+4a^{4}G\pi V_{1}^{2}\right)\langle Q^{2}\rangle. \tag*{(108)}
\end{aligned} + ϕ 0 ′ 2 a 2 V 1 ( H ′ + 5 H 2 ) + ϕ 0 ′ 10 a 2 V 1 ( H 2 − H ′ ) ⟨ Ψ Ψ ′ ⟩ + [ − 9 ( H 2 − H ′ ) + H 2 − H ′ 4 a 4 G π V 1 2 + a 2 V 2 + 9 H 2 + ϕ 0 ′ 6 a 2 V 1 H ] ⟨( Ψ ′ ) 2 ⟩ + ( 4 H ′ − 16 H 2 + ϕ 0 ′ 2 a 2 V 1 H ) ⟨ Ψ Q ′ ⟩ + ( 6 H + ϕ 0 ′ 2 a 2 V 1 ) ⟨ Ψ ′ Q ′ ⟩ + [ − 26 H H ′ + ϕ 0 ′ 8 a 2 V 1 H ′ + 48 a 2 G π V 0 H + 2 a 2 V 2 H + 8 H 3 + H 2 − H ′ 8 a 4 G π V 1 2 H + ϕ 0 ′ 6 a 2 V 1 H 2 + 64 a 2 G π V 1 ϕ 0 ′ ] ⟨ Ψ Q ⟩ + ( 12 H ′ + 2 a 2 V 2 − 6 H 2 + H 2 − H ′ 8 a 4 G π V 1 2 + ϕ 0 ′ 14 a 2 V 1 H ) ⟨ Ψ ′ Q ⟩ − ( H ′ + 2 H 2 ) ⟨( Q ′ ) 2 ⟩ + ( − 8 H H ′ + 14 H 3 + 8 a 2 G π V 1 ϕ 0 ′ ) ⟨ Q Q ′ ⟩ + ( − 3 H ′ 2 − a 2 V 2 H ′ + 8 H 2 H ′ + a 2 V 2 H 2 + 4 H 4 + 32 a 2 G π V 1 H ϕ 0 ′ + 4 a 4 G π V 1 2 ) ⟨ Q 2 ⟩ . (108)
τ ^ i i ( 2 ) ( k ) = 1 H ′ − H 2 { 1 3 ⟨ ( ∇ Ψ ′ ) 2 ⟩ − ⟨ ( Δ Ψ ) 2 ⟩ − ( 16 3 H + 2 a 2 V 1 ϕ 0 ′ ) ⟨ ∇ Ψ ⋅ ∇ Ψ ′ ⟩ + [ 1 3 ( 5 H ′ − 22 H 2 ) − 2 a 2 V 1 H ϕ 0 ′ ] ⟨ ( ∇ Ψ ) 2 ⟩ + [ 7 ( H ′ ) 2 + 4 H 2 H ′ + a 2 V 2 H 2 − 20 H 4 − 2 a 2 V 1 H ( 2 H ′ + H 2 ) ϕ 0 ′ + 4 a 4 G π V 1 2 H 2 H ′ − H 2 ] ⟨ Ψ 2 ⟩ + [ 14 H H ′ − 4 a 2 V 1 ( H ′ + 2 H 2 ) ϕ 0 ′ + 2 a 2 V 2 H − 32 H 3 + 8 a 4 G π V 1 2 H H ′ − H 2 ] ⟨ Ψ Ψ ′ ⟩ + [ H ′ + a 2 V 2 − 10 H 2 − 6 a 2 V 1 H ϕ 0 ′ + 4 a 4 G π V 1 2 H ′ − H 2 ] ⟨ ( Ψ ′ ) 2 ⟩ + ( 8 H ′ + 12 H 2 − 2 a 2 V 1 H ϕ 0 ′ ) ⟨ Ψ Q ′ ⟩ + ( 6 H − 2 a 2 V 1 ϕ 0 ′ ) ⟨ Ψ ′ Q ′ ⟩ + [ 8 a 2 V 1 H ′ ϕ 0 ′ − 2 a 2 V 2 H + 8 H 3 − 10 a 2 V 1 H 2 ϕ 0 ′ + 16 a 2 G π V 1 ϕ 0 ′ − 8 a 4 G π V 1 2 H H ′ − H 2 ] ⟨ Ψ Q ⟩ + ( 2 H ′ − 2 a 2 V 2 − 8 a 4 G π V 1 2 H ′ − H 2 − 2 a 2 V 1 H ϕ 0 ′ ) ⟨ Ψ ′ Q ⟩ + 6 H ⟨ Ψ Q ′ ′ ⟩ + H ′ ⟨ ( Q ′ ) 2 ⟩ + ( − 12 H H ′ + 16 a 2 G π V 0 H − 8 H 3 − 8 a 2 G π V 1 ϕ 0 ′ ) ⟨ Q Q ′ ⟩ + [ − 3 ( H ′ ) 2 + a 2 V 2 H ′ − 32 a 2 G π V 1 H ϕ 0 ′ − a 2 V 2 H 2 − 12 H 4 + 4 a 4 G π V 1 2 ] ⟨ Q 2 ⟩ − 2 H 2 ⟨ Q Q ′ ′ ⟩ + 2 H ⟨ Q ′ Q ′ ′ ⟩ } . (109) \begin{aligned}
\hat{\tau}^{(2)}_{ii}(k)=\frac{1}{\mathcal{H}'-\mathcal{H}^{2}}\Bigg\{&
\frac{1}{3}\langle(\nabla\Psi')^{2}\rangle-\langle(\Delta\Psi)^{2}\rangle-\left(\frac{16}{3}\mathcal{H}+\frac{2a^{2}V_{1}}{\phi'_0}\right)\langle\nabla\Psi\cdot\nabla\Psi'\rangle\\
&+\left[\frac{1}{3}(5\mathcal{H}'-22\mathcal{H}^{2})-\frac{2a^{2}V_{1}\mathcal{H}}{\phi'_0}\right]\langle(\nabla\Psi)^{2}\rangle\\
&+\left[7(\mathcal{H}')^{2}+4\mathcal{H}^{2}\mathcal{H}'+a^{2}V_{2}\mathcal{H}^{2}-20\mathcal{H}^{4}-\frac{2a^{2}V_{1}\mathcal{H}(2\mathcal{H}'+\mathcal{H}^{2})}{\phi'_0}+\frac{4a^{4}G\pi V_{1}^{2}\mathcal{H}^{2}}{\mathcal{H}'-\mathcal{H}^{2}}\right]\langle\Psi^{2}\rangle\\
&+\left[14\mathcal{H}\mathcal{H}'-\frac{4a^{2}V_{1}(\mathcal{H}'+2\mathcal{H}^{2})}{\phi'_0}+2a^{2}V_{2}\mathcal{H}-32\mathcal{H}^{3}+\frac{8a^{4}G\pi V_{1}^{2}\mathcal{H}}{\mathcal{H}'-\mathcal{H}^{2}}\right]\langle\Psi\Psi'\rangle\\
&+\left[\mathcal{H}'+a^{2}V_{2}-10\mathcal{H}^{2}-\frac{6a^{2}V_{1}\mathcal{H}}{\phi'_0}+\frac{4a^{4}G\pi V_{1}^{2}}{\mathcal{H}'-\mathcal{H}^{2}}\right]\langle(\Psi')^{2}\rangle\\
&+(8\mathcal{H}'+12\mathcal{H}^{2}-\frac{2a^{2}V_{1}\mathcal{H}}{\phi'_0})\langle\Psi Q'\rangle+(6\mathcal{H}-\frac{2a^{2}V_{1}}{\phi'_0})\langle\Psi'Q'\rangle\\
&+\left[\frac{8a^{2}V_{1}\mathcal{H}'}{\phi'_0}-2a^{2}V_{2}\mathcal{H}+8\mathcal{H}^{3}-\frac{10a^{2}V_{1}\mathcal{H}^{2}}{\phi'_0}+16a^{2}G\pi V_{1}\phi'_0-\frac{8a^{4}G\pi V_{1}^{2}\mathcal{H}}{\mathcal{H}'-\mathcal{H}^{2}}\right]\langle\Psi Q\rangle\\
&+\left(2\mathcal{H}'-2a^{2}V_{2}-\frac{8a^{4}G\pi V_{1}^{2}}{\mathcal{H}'-\mathcal{H}^{2}}-\frac{2a^{2}V_{1}\mathcal{H}}{\phi'_0}\right)\langle\Psi'Q\rangle+6\mathcal{H}\langle\Psi Q''\rangle\\
&+\mathcal{H}'\langle(Q')^{2}\rangle+(-12\mathcal{H}\mathcal{H}'+16a^{2}G\pi V_{0}\mathcal{H}-8\mathcal{H}^{3}-8a^{2}G\pi V_{1}\phi'_0)\langle QQ'\rangle\\
&+\left[-3(\mathcal{H}')^{2}+a^{2}V_{2}\mathcal{H}'-32a^{2}G\pi V_{1}\mathcal{H}\phi'_0-a^{2}V_{2}\mathcal{H}^{2}-12\mathcal{H}^{4}+4a^{4}G\pi V_{1}^{2}\right]\langle Q^{2}\rangle\\
&-2\mathcal{H}^{2}\langle QQ''\rangle+2\mathcal{H}\langle Q'Q''\rangle\Bigg\}. \tag*{(109)}
\end{aligned} τ ^ ii ( 2 ) ( k ) = H ′ − H 2 1 { 3 1 ⟨( ∇ Ψ ′ ) 2 ⟩ − ⟨( ΔΨ ) 2 ⟩ − ( 3 16 H + ϕ 0 ′ 2 a 2 V 1 ) ⟨ ∇Ψ ⋅ ∇ Ψ ′ ⟩ + [ 3 1 ( 5 H ′ − 22 H 2 ) − ϕ 0 ′ 2 a 2 V 1 H ] ⟨( ∇Ψ ) 2 ⟩ + [ 7 ( H ′ ) 2 + 4 H 2 H ′ + a 2 V 2 H 2 − 20 H 4 − ϕ 0 ′ 2 a 2 V 1 H ( 2 H ′ + H 2 ) + H ′ − H 2 4 a 4 G π V 1 2 H 2 ] ⟨ Ψ 2 ⟩ + [ 14 H H ′ − ϕ 0 ′ 4 a 2 V 1 ( H ′ + 2 H 2 ) + 2 a 2 V 2 H − 32 H 3 + H ′ − H 2 8 a 4 G π V 1 2 H ] ⟨ Ψ Ψ ′ ⟩ + [ H ′ + a 2 V 2 − 10 H 2 − ϕ 0 ′ 6 a 2 V 1 H + H ′ − H 2 4 a 4 G π V 1 2 ] ⟨( Ψ ′ ) 2 ⟩ + ( 8 H ′ + 12 H 2 − ϕ 0 ′ 2 a 2 V 1 H ) ⟨ Ψ Q ′ ⟩ + ( 6 H − ϕ 0 ′ 2 a 2 V 1 ) ⟨ Ψ ′ Q ′ ⟩ + [ ϕ 0 ′ 8 a 2 V 1 H ′ − 2 a 2 V 2 H + 8 H 3 − ϕ 0 ′ 10 a 2 V 1 H 2 + 16 a 2 G π V 1 ϕ 0 ′ − H ′ − H 2 8 a 4 G π V 1 2 H ] ⟨ Ψ Q ⟩ + ( 2 H ′ − 2 a 2 V 2 − H ′ − H 2 8 a 4 G π V 1 2 − ϕ 0 ′ 2 a 2 V 1 H ) ⟨ Ψ ′ Q ⟩ + 6 H ⟨ Ψ Q ′′ ⟩ + H ′ ⟨( Q ′ ) 2 ⟩ + ( − 12 H H ′ + 16 a 2 G π V 0 H − 8 H 3 − 8 a 2 G π V 1 ϕ 0 ′ ) ⟨ Q Q ′ ⟩ + [ − 3 ( H ′ ) 2 + a 2 V 2 H ′ − 32 a 2 G π V 1 H ϕ 0 ′ − a 2 V 2 H 2 − 12 H 4 + 4 a 4 G π V 1 2 ] ⟨ Q 2 ⟩ − 2 H 2 ⟨ Q Q ′′ ⟩ + 2 H ⟨ Q ′ Q ′′ ⟩ } . (109)
References