Quantum Equivalence Principle

• Quantum Equivalence Principle is a reformulation of Einstein’s classical principle that includes quantum matter, superpositions, and operator-based mass-energy descriptions.
• It features key formulations such as operator mass-energy equality, gauge-phase interference, and operational tests that translate free-fall into observable quantum effects.
• Experimental advances, notably in cold-atom interferometry, have validated cubic-phase predictions and highlighted the role of quantum coherence and entanglement in gravitational settings.

Quantum Equivalence Principle denotes a set of formulations that extend Einstein’s equivalence principle from classical trajectories and classical reference frames to quantum matter, quantum superpositions, and, in some approaches, quantum gravitational fields. In the literature, the term covers several distinct but related ideas: operator equality of rest, inertial, and gravitational mass-energy; gauge-equivalent descriptions of a matter wave in a uniform gravitational field and in a uniformly accelerated frame; and quantum-reference-frame constructions in which a locally inertial description exists even when the metric is in superposition (Zych et al., 2015, Dobkowski et al., 20 Feb 2025, Giacomini et al., 2020).

1. Classical antecedent and the need for reformulation

In its modern form, the Einstein Equivalence Principle states that “in any and every local inertial (classical) reference frame, anywhere and anytime in the universe, all nongravitational laws of physics take their special-relativistic form” (Giacomini et al., 2020). Classical discussions usually separate this into the Weak Equivalence Principle, Local Lorentz Invariance, and Local Position Invariance. In a local inertial frame constructed with Riemann normal coordinates, the metric satisfies

$\tilde g_{\mu\nu}(\xi)=\eta_{\mu\nu}+O(\xi^2),$

and free-fall geodesics reduce locally to straight-line motion (Giacomini et al., 2020). In the weak-field language used in quantum tests, the classical EEP is often summarized by the equalities between rest, inertial, and gravitational mass-energy contributions (Zych et al., 2015).

Quantum theory forces a reformulation because the central objects of the classical statement cease to be c-numbers. Internal energies become operators on an internal Hilbert space, probes can be delocalised in space or time, and in some proposals the gravitational field itself may be placed in superposition (Zych et al., 2015, Giacomini et al., 2020). Classical tests constrain only diagonal matrix elements in an energy basis, whereas a genuinely quantum statement must also constrain off-diagonal elements that govern superposition, coherence, and entanglement (Zych et al., 2015).

A further reason for reformulation is operational. In quantum mechanics a sharp trajectory is not generally available, so the classical slogan “all bodies fall alike” must be recast in terms of wavefunctions, observables, or interference phases rather than worldlines alone. This motivates several non-identical formulations rather than a single canonical one.

2. Principal formulations

A major line of work promotes rest, inertial, and gravitational masses to operators,

$$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$

and takes the quantum equivalence principle to be the operator identity

$$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$

In a weak, static gravitational field, this enters the composite-particle Hamiltonian

$$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$

If the internal-energy operators fail to commute, internal and external degrees of freedom generically entangle, producing effects invisible to classical EEP tests (Zych et al., 2015).

A second formulation is wave-mechanical and gauge-theoretic. For a uniform field, the Schrödinger dynamics in the Newtonian frame with

$L_N=\tfrac12 m\dot z^2-mgz$

and in the upward-accelerated Einsteinian frame with

$L_E=\tfrac12 m(\dot z+gt)^2$

differ only by a total derivative. In quantum theory this induces the gauge phase

$$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$

with

$$\phi_{\rm gauge}(z,t)=\frac{m}{\hbar}\Bigl[-\tfrac16 g^2 t^3-gzt\Bigr].$$

For a single packet at a fixed point this phase is locally unobservable, but it becomes observable as a relative phase when different branches of a matter wave experience different accelerations (Dobkowski et al., 20 Feb 2025).

Other formulations are explicitly operational. One proposal states that free fall in a uniform field translates the position-probability distribution of an arbitrary quantum state into that of free evolution, up to a mass-independent shift of its mean; another states that any two particles with the same initial velocity wave function behave identically in free fall, irrespective of their masses (Anastopoulos et al., 2017). A Heisenberg-picture formulation instead starts from

$$H=\tfrac12\,g^{\mu\nu}(\hat x)\,\hat p_\mu \hat p_\nu$$

and derives exact operator geodesic equations,

$$\frac{d\hat x^\mu}{d\tau}=\hat v^\mu,\qquad \frac{d\hat p_\mu}{d\tau}+\Gamma^\sigma_{\mu\lambda}(\hat x)\,\hat p_\sigma\,\hat v^\lambda=0,$$

so that universality is phrased as a statement about quantum observables rather than classical trajectories (Kong, 2023).

A fully relativistic extension generalizes the operator formulation to arbitrary background spacetimes and to massive bosons and fermions, while noting that the principle is trivially satisfied for massless particles. In that framework, a violation implies modified Lorentz transformations in flat spacetime and particle-dependent effective metrics in curved spacetime (Das et al., 2023).

Formulation	Central statement	Representative work
Operator mass-energy equality	$$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$0	(Zych et al., 2015, Das et al., 2023)
Gauge-phase formulation	Local equivalence holds up to $$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$1; relative phase is interferometrically observable	(Dobkowski et al., 20 Feb 2025)
Operational free-fall criteria	Position-probability and velocity-wave-function versions of quantum WEP	(Anastopoulos et al., 2017)
Observable/geodesic formulation	Heisenberg equations yield operator geodesic motion	(Kong, 2023)

3. Quantum reference frames and superposed geometries

A third major direction replaces classical local inertial frames by quantum reference frames. In this approach, macroscopically distinguishable metrics $$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$2 are assigned orthogonal quantum states $$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$3, and one considers superpositions over sectors $$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$4. For a reference particle $$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$5, a controlled unitary

$$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$6

defines a quantum locally inertial frame. In that frame,

$$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$7

so each amplitude sees a locally Minkowski metric at the quantum-reference origin (Giacomini et al., 2020).

This construction is closely related to the “quantum diffeomorphism” program. One introduces extended states that are superpositions of manifolds $$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$8 with fields $$\hat M_r := m_r\,\hat I_{\rm int}+\frac{\hat H_{\rm rest}}{c^2},\qquad \hat M_i := m_i\,\hat I_{\rm int}+\frac{\hat H_{\rm inertial}}{c^2},\qquad \hat M_g := m_g\,\hat I_{\rm int}+\frac{\hat H_{\rm grav}}{c^2},$$9, together with quantum coordinate systems built from identifications of points across branches. A quantum coordinate transformation updates both the state and the branchwise identifications, and can be chosen so that the causal structure is locally definite in the vicinity of a point (Hardy, 2019). In this language, the quantum equivalence principle is the existence of a quantum coordinate system in which local causal structure is sharp.

The same idea has been used to argue that a superposition of massive bodies and their associated metrics need not violate Einstein’s principle. Instead of demanding a single classical local inertial chart, one allows the frame transformation itself to be a unitary quantum map to a quantum reference frame. This yields a locally Minkowskian description in each branch and removes the need to invoke gravity-induced spontaneous collapse solely on equivalence-principle grounds (Giacomini et al., 2021).

Clock-based versions sharpen the operational content. For entangled clocks in a quantum superposition of positions in Earth’s field, the proper-time evolution along each branch is represented by an operator, and the validity of the generalized EEP is stated to be equivalent to the possibility of transforming to the perspective of each clock as a quantum reference frame. Violation would make dynamical evolution in the frame of each clock impossible and would modify laboratory-frame detection probabilities (Cepollaro et al., 2021).

4. Direct observation with matter waves

The most direct experimental realization to date is a cold-atom interferometer designed to observe the free-fall gauge phase itself. The key theoretical point is that a matter wave can be split into two branches that are simultaneously described with respect to two differently accelerating frames: one wave packet remains static in the laboratory frame while the other is in free fall. The accumulated relative phase is cubic in the interrogation time. For the idealized sequence, the relative phase between the ballistic and reference paths is proportional to $$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$0; including finite kick durations $$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$1 and inter-pulse delays $$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$2, the analytic expression becomes

$$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$3

which reduces to the cubic law in the short-pulse limit (Dobkowski et al., 20 Feb 2025).

Experimentally, a Bose–Einstein condensate of $$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$4 with $$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$5 atoms was produced $$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$6 below an atom-chip surface. After trap release, a $$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$7-kick cooling pulse collimated the cloud to $$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$8 and imparted a small upward launch. A microwave $$\hat M_r=\hat M_i=\hat M_g \quad\Longleftrightarrow\quad m_r=m_i=m_g,\qquad \hat H_{\rm rest}=\hat H_{\rm inertial}=\hat H_{\rm grav}.$$9 pulse created $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$0, where $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$1 is magnetically sensitive and $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$2 is first-order insensitive. A first Stern–Gerlach kick accelerated the $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$3 component upward; a subsequent microwave $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$4 pulse exchanged internal states so that the moving packet became truly ballistic while the other packet was held by a gradient that exactly cancels $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$5. A second swap and identical kick recombined the paths, and a final $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$6 pulse mapped the relative phase onto internal-state populations (Dobkowski et al., 20 Feb 2025).

Repeating the sequence for $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$7 from $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$8 up to $$\hat H_{\rm test}^Q = m_r c^2 + \hat H_{\rm rest} +\frac{\hat P^2}{2m_i} +m_g\,\phi(\hat Q) -\hat H_{\rm inertial}\frac{\hat P^2}{2m_i^2c^2} +\hat H_{\rm grav}\frac{\phi(\hat Q)}{c^2}.$$9, the experiment observed $L_N=\tfrac12 m\dot z^2-mgz$0 full oscillations in the population of $L_N=\tfrac12 m\dot z^2-mgz$1, corresponding to a total phase of $L_N=\tfrac12 m\dot z^2-mgz$2. The fringe visibility started at $L_N=\tfrac12 m\dot z^2-mgz$3 and fell to $L_N=\tfrac12 m\dot z^2-mgz$4 at the longest times, limited by imperfect overlap, wave-packet shape evolution, gradient curvature, and “Humpty–Dumpty” irreversibility. The shot-to-shot phase uncertainty was $L_N=\tfrac12 m\dot z^2-mgz$5 (SEM), and the systematic uncertainty, dominated by current-ratio stability, was $L_N=\tfrac12 m\dot z^2-mgz$6. A high-resolution numerical simulation incorporating $L_N=\tfrac12 m\dot z^2-mgz$7D wave-packet evolution, curved potentials, second-order Zeeman effects, atom–atom interactions, and realistic pulse shapes reproduced the observed phase within these uncertainties; the analytic cubic-phase model also agreed well (Dobkowski et al., 20 Feb 2025).

This experiment was presented as the first direct observation of the free-fall gauge phase for matter waves and as a confirmation of the equivalence principle in the quantum domain beyond classical tests and the original COW neutron interferometer. Its significance is not only metrological. The symmetry between the Newtonian and Einsteinian frames is exact for de Broglie waves, and no transformation of electromagnetic fields is required, in contrast to light-pulse interferometers (Dobkowski et al., 20 Feb 2025).

5. Violations, limits, and conceptual disputes

Not all discussions of quantum equivalence principle take the form of an exact extension. In low-energy quantum gravity treated as an effective field theory, loop effects generate non-local terms such as

$L_N=\tfrac12 m\dot z^2-mgz$8

and quantum corrections perturb the classical geodesic picture. Nevertheless, the same analysis emphasizes that general coordinate invariance survives quantization and continues to organize the theory. In that sense, the equivalence principle survives in a broadened symmetry form even when point-particle geodesics and sharply local inertial frames become fuzzy (Bjerrum-Bohr et al., 2015).

Scalar–tensor theories provide an explicit mechanism for small quantum violations. Ward identities derived from diffeomorphism invariance and a broken Weyl/shift symmetry guarantee the weak equivalence principle to all orders in matter coupling constants, so matter loops alone preserve universality. Once the scalar or graviton itself runs in the loop, however, the same identities predict weak-equivalence-principle violations suppressed by at least three powers of the gravitational couplings: $L_N=\tfrac12 m\dot z^2-mgz$9 The result is conceptually important because it shows that universality is not protected by an exact symmetry in that class of models (Armendariz-Picon et al., 2011).

There is also disagreement over which classical principle should be generalized. One overview argues that only the Strong Equivalence Principle is fundamental enough to survive quantization, because other variants already hold only approximately at the classical level. In that account, the Quantum SEP is associated with minimal coupling and local special-relativistic form, whereas Quantum WEP, understood as geodesic motion of quantum test particles, is generically violated for kink excitations in superposed geometries (Paunkovic et al., 2022). This does not negate other formulations, but it changes the object that is being generalized.

The role of mass superselection is another fault line. One nonrelativistic treatment uses the extended Galileo group, promotes mass to an operator, and concludes that Einstein’s principle can be maintained without invoking Bargmann’s mass superselection rule (Hernandez-Coronado et al., 2013). By contrast, a photonic quantum simulator based on polarization-entangled photons in curved and birefringent optical waveguides emulates coherent superpositions of effective mass states and treats the resulting Hong–Ou–Mandel antibunching as a signature of effective quantum-EP violation (Longhi, 2017). The literature therefore contains both non-violation and effective-violation narratives, depending on the underlying kinematics and on what is taken as physically admissible.

6. Experimental programs and open directions

Several experimental programs probe operator-level or branch-dependent versions of the principle. A simple proposal for a harmonically trapped spin-$L_E=\tfrac12 m(\dot z+gt)^2$0 atom in gravity and a magnetic field promotes mass to an operator and predicts that, after a sudden switch-off of the magnetic field, violations of Local Lorentz Invariance, Local Position Invariance, or quantum WEP appear as otherwise forbidden transitions between oscillator and spin states. The same work also proposes a free-fall release of the whole apparatus as a direct test of the classical WEP parameter $L_E=\tfrac12 m(\dot z+gt)^2$1 (Orlando et al., 2015). Entangled-clock interferometry aims at the generalized EEP by comparing branch-dependent proper times and looking for residual visibility loss or phase shifts beyond the standard overlap $L_E=\tfrac12 m(\dot z+gt)^2$2 (Cepollaro et al., 2021).

Gravitational-wave phenomenology extends the operator-mass formulation to macroscopic detectors. In that framework, QEP violations deform the standard formulas for the wave amplitude and chirp through the ratios $L_E=\tfrac12 m(\dot z+gt)^2$3 and $L_E=\tfrac12 m(\dot z+gt)^2$4, and bounds can be extracted from resolved LIGO/Virgo events. Using GW170817, GW190521, and GW190814, the reported limits are roughly at the $L_E=\tfrac12 m(\dot z+gt)^2$5 confidence level (Das et al., 16 Jun 2025).

Event	WEP bound $L_E=\tfrac12 m(\dot z+gt)^2$6	LLI / LPI bounds
GW170817	$L_E=\tfrac12 m(\dot z+gt)^2$7 to $L_E=\tfrac12 m(\dot z+gt)^2$8	LLI: $L_E=\tfrac12 m(\dot z+gt)^2$9 to $$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$0; LPI: $$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$1 to $$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$2
GW190521	$$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$3 to $$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$4	LLI: $$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$5 to $$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$6; LPI: $$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$7 to $$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$8
GW190814	$$\psi_N(z,t)=\exp[i\,\phi_{\rm gauge}(z,t)]\,\psi_E\!\left(z+\tfrac12 gt^2,t\right),$$9 to $$\phi_{\rm gauge}(z,t)=\frac{m}{\hbar}\Bigl[-\tfrac16 g^2 t^3-gzt\Bigr].$$0	LLI: $$\phi_{\rm gauge}(z,t)=\frac{m}{\hbar}\Bigl[-\tfrac16 g^2 t^3-gzt\Bigr].$$1 to $$\phi_{\rm gauge}(z,t)=\frac{m}{\hbar}\Bigl[-\tfrac16 g^2 t^3-gzt\Bigr].$$2; LPI: $$\phi_{\rm gauge}(z,t)=\frac{m}{\hbar}\Bigl[-\tfrac16 g^2 t^3-gzt\Bigr].$$3 to $$\phi_{\rm gauge}(z,t)=\frac{m}{\hbar}\Bigl[-\tfrac16 g^2 t^3-gzt\Bigr].$$4

The same body of work identifies several forward directions. The matter-wave interferometer of the direct observation experiment was proposed as a platform for incorporating internal “clock” degrees of freedom, for precision gravimetry, for searches for short-range forces and dark-sector couplings, and for extensions to larger masses such as levitated nanoparticles or nanodiamonds (Dobkowski et al., 20 Feb 2025). Another line of analysis, working in linearized quantum gravity and in the monopole approximation, argues that the equivalence principle remains valid even though the gravitational field is quantum, by mapping the problem to uniformly accelerating Unruh–DeWitt detectors (Zhang, 2024). A conceptually different proposal defines the quantum equivalence principle at the level of second-order moments: the dynamic part of quantum fluctuations is frame dependent and removable, while the geometric part is universal, mass independent, and tied to coarse-graining and diffeomorphism anomaly; this is then used to motivate an induced Einstein–Hilbert action in a non-linear sigma-model of a material quantum reference frame (Luo, 2024).

Taken together, these developments show that “Quantum Equivalence Principle” is not a single statement but a research program. Across its different formulations, the recurring theme is that classical universality must be re-expressed in operator language, interference observables, or quantum-reference-frame transformations. The resulting questions are no longer confined to whether free fall is universal for a point particle; they concern whether local inertiality, clock universality, and frame covariance survive when the probe, the frame, or even the geometry itself is quantum.