Next-to-Minimal Bouncing Cosmology

• NMBC is a class of bounce models that replace the initial singularity with a contracting-to-expanding transition, inheriting perturbative predictions from slow-roll inflation.
• It employs a non-minimal conformal transformation of the inflationary action to achieve a stable attractor and evade the challenges of ghost instabilities and the no-go theorem.
• An extended model variant introduces an early contraction phase to produce a broken power-law stochastic gravitational-wave background, opening new observational windows.

Searching arXiv for the cited NMBC papers and related context. Next-to-Minimal Bouncing Cosmology (NMBC) denotes a class of non-singular early-Universe bounce models built to reproduce the perturbative successes of slow-roll inflation while replacing the initial singularity with a contracting-to-expanding transition. In the literature summarized here, the term is used primarily for a conformally constructed non-minimal bounce in which an inflationary scalar-tensor theory is mapped into a bounce frame with a non-minimally coupled scalar field, yielding a stable attractor, a non-singular bounce, and inherited predictions for $r$ and $f_{NL}$ (Nandi, 2020, Nandi et al., 2022, Nandi, 2020). A later and distinct usage applies the same acronym to a multi-phase bounce with an additional early contraction phase, introduced to obtain an analytic broken power-law stochastic gravitational-wave background (SGWB) (Li, 17 Jul 2025). The shared acronym therefore refers to related but not identical constructions.

1. Conceptual definition and scope

In the conformal construction, NMBC is obtained from a minimal or non-minimal slow-roll inflationary action by a conformal transformation that reinterprets the inflationary background as a bouncing one. The central motivation is to address three standard obstacles in bounce model building: a stable attractor solution, a non-singular bounce, and evasion of the no-go theorem stating that one cannot simultaneously maintain acceptable tensor-to-scalar ratio and non-Gaussian scalar spectrum in many non-inflationary single-field contracting models (Nandi, 2020).

The basic claim of this construction is that the bounce inherits its perturbation sector from a viable inflationary parent theory. This means that curvature perturbations and tensor perturbations are preserved by the conformal map, so the bounce frame reproduces the observational predictions of the parent slow-roll model rather than generating a new and potentially problematic perturbation sector (Nandi, 2020, Nandi et al., 2022).

A separate 2025 usage of NMBC extends a four-phase “minimal” bounce by adding an extra early contraction phase, with the principal purpose of modifying the SGWB into a broken power law and enlarging the observational window (Li, 17 Jul 2025). This suggests that “NMBC” has become a broader label rather than a uniquely fixed model name.

2. Non-minimal conformal construction

The conformal NMBC is formulated in a Jordan-frame bounce metric $g^b_{\mu\nu}$, with action

$$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$

where

$$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$

(Nandi, 2020).

The corresponding Einstein-frame metric is introduced by

$g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$

under which the action becomes

$$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$

One may also define a canonically normalized Einstein-frame field via

$d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$

(Nandi, 2020).

A related formulation in a non-minimal inflationary Jordan frame writes the total action as

$$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$

with an extra barotropic fluid characterized by $w_m=P_m/\rho_m$ (Nandi et al., 2022). In that treatment, the bounce frame is generated by a conformal rescaling

$f_{NL}$0

leading to transformed couplings

$f_{NL}$1

$f_{NL}$2

(Nandi et al., 2022).

The conformal factor is chosen so that the bounce-frame scale factor contracts as a power law in conformal time,

$f_{NL}$3

or, in the compact summary of the original construction,

$f_{NL}$4

(Nandi, 2020, Nandi, 2020). This engineered contraction is the defining step that turns the inflationary ancestor into a bounce model.

3. Background dynamics and bounce conditions

For a spatially flat FLRW background, the bounce-frame Friedmann equations in the conformal NMBC are

$f_{NL}$5

$f_{NL}$6

together with

$f_{NL}$7

(Nandi, 2020).

With an additional barotropic fluid, the background equations become

$f_{NL}$8

$f_{NL}$9

$g^b_{\mu\nu}$0

$g^b_{\mu\nu}$1

(Nandi, 2020).

A non-singular bounce requires

$g^b_{\mu\nu}$2

(Nandi, 2020). In the conformal construction, these conditions are met because as $g^b_{\mu\nu}$3 approaches the minimum of $g^b_{\mu\nu}$4, one finds $g^b_{\mu\nu}$5, so that $g^b_{\mu\nu}$6 crosses zero without $g^b_{\mu\nu}$7 or $g^b_{\mu\nu}$8 (Nandi, 2020). In the alternative notation of the non-minimal inflation construction,

$g^b_{\mu\nu}$9

so the contracting phase arises when $$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$0, the expanding phase when $$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$1, and the bounce occurs at $$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$2 (Nandi et al., 2022).

The model violates the effective null energy condition at the bounce, but the construction is stated to avoid ghost pathology because $$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$3 remains positive outside the contracting phase and the kinetic prefactor does not flip sign violently (Nandi, 2020). In the perturbative treatment summarized below, the absence of ghosts and gradient instabilities is instead established through the quadratic action for $$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$4 (Nandi et al., 2022).

4. Attractor structure, barotropic fluids, and BKL avoidance

A central requirement for a viable contracting cosmology is the existence of a stable attractor that suppresses sensitivity to initial conditions and evades Belinsky-Khalatnikov-Lifshitz (BKL) instability. In the conformal NMBC, the contracting scale factor is taken as

$$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$5

for which the slow-roll-like parameter in the bounce frame satisfies

$$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$6

(Nandi, 2020). For $$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$7 and $$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$8, numerical phase-space trajectories for different initial $$S_b = \frac12 \int d^4x \sqrt{-g_b}\,\Bigl[ F(\phi)\,R_b - \omega(\phi)\,g_b^{\mu\nu}\partial_\mu\phi\,\partial_\nu\phi - 2\,V_b(\phi) \Bigr],$$9 and $$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$0 rapidly converge to $$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$1, demonstrating stability and absence of the BKL instability in that setting (Nandi, 2020).

The fluid-extended dynamical-system analysis introduces

$$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$2

with $$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$3 and

$$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$4

(Nandi, 2020). The Friedmann constraint is

$$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$5

The contracting bounce-attractor fixed point is

$$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$6

with Jacobian eigenvalues

$$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$7

(Nandi, 2020). Because the contracting phase has $$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$8 and $$F(\phi)=f(\phi)^2,\qquad \omega(\phi)=f(\phi)^2\Bigl[1-6\,(f_{,\phi}/f)^2\Bigr],\qquad V_b(\phi)=f(\phi)^4\,V_I(\phi)$$9, stability requires $g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$0 and $g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$1. This yields

$g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$2

(Nandi, 2020).

For anisotropic stress, modeled as a stiff fluid with $g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$3, the BKL-avoidance condition becomes

$g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$4

For typical slow-roll models with $g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$5, this reduces to

$g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$6

(Nandi, 2020). The 2022 treatment reaches the parallel conclusion that, in the bounce frame, stability requires $g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$7 and, for $g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$8, $g^I_{\mu\nu}=f(\phi)^2\,g^b_{\mu\nu},$9, so anisotropic energy $$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$0 never dominates the contracting scalar background (Nandi et al., 2022). Both analyses identify ekpyrotic behavior, corresponding to $$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$1, as the strongest attractor (Nandi, 2020).

5. Perturbations, $$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$2, $$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$3, and the no-go theorem

The most distinctive feature of the conformal NMBC is the claim that perturbations are inherited exactly from the parent inflationary model. In the original construction, one has

$$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$4

and the tensor modes are likewise unchanged under the conformal transformation (Nandi, 2020). The 2022 formulation states that the quadratic curvature action in the inflationary frame is

$$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$5

with the corresponding bounce-frame quantity satisfying $$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$6, so that $$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$7 to all orders and tensor modes are also unchanged (Nandi et al., 2022).

As a result, the usual single-field slow-roll expressions carry over: $$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$8 with

$$S_I = \frac12 \int d^4x \sqrt{-g_I}\,[\,R_I-(\partial\phi)^2-2V_I(\phi)\,].$$9

(Nandi, 2020). For the Starobinsky inflationary potential, the explicit values quoted are $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$0 and $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$1, and these are carried over into the bounce frame unmodified (Nandi, 2020).

This directly addresses the observational no-go theorem: in many non-inflationary single-field contracting models one cannot obtain both small $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$2 and small $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$3, but here the perturbation sector is not generated by a conventional contracting mechanism. Instead, it is inherited from a slow-roll inflationary model by a conformal map, with no new cubic interactions that would blow up $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$4 and no extra tensor sources that would increase $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$5 (Nandi, 2020). A plausible implication is that the construction shifts the burden of phenomenological viability from bounce-specific perturbation generation to the consistency of the conformal embedding.

The perturbative stability statement is correspondingly strong in the 2022 analysis: because $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$6 throughout, including at the bounce, the kinetic term has the correct sign and the gradient term has coefficient $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$7, implying $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$8. The same conclusion is stated for the tensor action, and the bounce is therefore described as free of ghost and gradient instabilities at linear and higher orders (Nandi et al., 2022).

6. Exit from the bounce and reheating

Immediately after the bounce, $d\chi/d\phi=\sqrt{\omega(\phi)}/f(\phi)$9 approaches the minimum of $$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$0 and begins to oscillate with frequency $$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$1, or more generally $$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$2 (Nandi, 2020). In this regime,

$$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$3

so the bounce-frame action reduces to the minimal Einstein-frame action (Nandi, 2020). Couplings of $$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$4 to matter fields then reheat the Universe “in the usual way” (Nandi, 2020).

The reheating time scale is quoted as

$$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$5

where $$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$6 is the decay rate of $$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$7 into matter (Nandi, 2020). For chaotic $$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$8, one typically finds

$$S_J=\frac12\int d^4x\sqrt{-g}\,[\,f^2(\phi)R-\omega(\phi)g^{\mu\nu}\partial_\mu\phi\partial_\nu\phi-2V(\phi)\,]+S_m[g_{\mu\nu},\Psi_m],$$9

while for Starobinsky-type potentials the reheating scale is set by the coupling to gauge sectors and can range from

$w_m=P_m/\rho_m$0

(Nandi, 2020).

This exit mechanism is important because many bounce models require a separate post-bounce completion. Here the reheating phase is not introduced as an independent sector but emerges when the non-minimal structure relaxes back to the minimal one near the minimum of the potential (Nandi, 2020, Nandi et al., 2022).

7. Alternative NMBC usage: multi-phase SGWB phenomenology

A later paper uses “Next-to-Minimal Bouncing Cosmology” for a different construction: a five-phase bounce obtained by adding an extra early-contraction phase to the four-phase “minimal” bounce (Li, 17 Jul 2025). The phase structure is defined in conformal time by

$w_m=P_m/\rho_m$1

with phases 0 through 4 given by early contraction, pre-bounce contraction, bounce contraction, bounce expansion, and post-bounce radiation (Li, 17 Jul 2025).

In this framework, the new early-contraction phase introduces a pivot frequency

$w_m=P_m/\rho_m$2

and the bounce scale is characterized by

$w_m=P_m/\rho_m$3

(Li, 17 Jul 2025). The SGWB is derived by a matrix-representation method using transformation matrices $w_m=P_m/\rho_m$4, boundary-matching matrices $w_m=P_m/\rho_m$5 and $w_m=P_m/\rho_m$6, and an overall amplitude matrix $w_m=P_m/\rho_m$7 or $w_m=P_m/\rho_m$8, depending on whether the mode exits in phase 0 or phase 1 (Li, 17 Jul 2025).

The resulting present-day SGWB is expressed as

$w_m=P_m/\rho_m$9

and the essential phenomenological feature is a broken power law: for $f_{NL}$00 the spectrum reproduces the minimal-bounce branch with tilt $f_{NL}$01, while for $f_{NL}$02 it acquires a new tilt

$f_{NL}$03

(Li, 17 Jul 2025).

The same paper states that all such NMBC models satisfying the current $f_{NL}$04 bound

$f_{NL}$05

automatically satisfy

$f_{NL}$06

thereby avoiding the trans-Planckian problem (Li, 17 Jul 2025). It also lists representative observational windows spanning CMB, PTA, mHz, Hz–kHz, and MHz–GHz bands (Li, 17 Jul 2025).

This alternative nomenclature is potentially a source of confusion. The conformal scalar-tensor bounce and the five-phase SGWB construction share the acronym “NMBC,” but they address different model-building questions: the former centers on non-minimal coupling, stability, and inherited perturbations (Nandi, 2020, Nandi et al., 2022, Nandi, 2020), whereas the latter centers on phase structure and analytic gravitational-wave spectra (Li, 17 Jul 2025).