Berry-Phase-Induced Chiral Work Difference

• The paper demonstrates that Berry-phase-induced chiral work difference isolates an orientation-sensitive geometric contribution from the metric part of finite-time work in cyclic quantum protocols.
• It employs a dissipative adiabatic perturbation expansion to separate Berry curvature effects from conventional energy dissipation, clarifying the unitary–dissipative crossover.
• The study connects quantum geometric tensors to thermodynamic metrics and offers experimental diagnostics for driven qubits, chiral magnets, and other chiral systems.

Berry-phase-induced chiral work difference denotes an orientation-sensitive geometric contribution to work that appears when a cyclic protocol is traversed with opposite senses on the same loop in parameter space. In its most explicit formulation, the effect is defined for slowly driven open quantum systems as the work difference between clockwise and counterclockwise cycles, and it isolates the Berry-phase or Berry-curvature contribution from the orientation-insensitive metric part of finite-time dissipation. In that setting it survives decoherence and interpolates between an interferometric thermodynamic Aharonov–Bohm effect in the unitary regime and a fringe-free dissipative signal in the open-system regime (Fei et al., 13 May 2026). Related literature uses closely allied constructions for chirality-dependent power, energy, or work asymmetries generated by Berry geometry in chiral kinetic theory, chiral magnets, nonlinear edge modes, and chiral insulators (Mueller et al., 2017, Freimuth et al., 2013, Beauvillain et al., 2024, Kubler et al., 2013).

1. Definition in cyclic thermodynamic driving

The thermodynamic setting is a cyclic protocol of control parameters $$\lambda(t)=(\lambda^1(t),\ldots,\lambda^i(t),\ldots)$$ executed either clockwise or counterclockwise on a closed loop $\Sigma$ in parameter space. The chirality of the protocol is the sense of traversal, and the chiral work difference is defined by

$$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$

Work and heat follow Alicki’s splitting,

$$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$

so that the total work over a cycle is

$$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$

In the open-system framework developed for this effect, the instantaneous eigenenergies are constant and non-degenerate, and all work stems from the geometric rotation of the instantaneous eigenbasis $\{|n(\lambda)\rangle\}$ (Fei et al., 13 May 2026).

Using the adiabatic-frame kinematics $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$ with $A_{mn}(t)=i\langle m|\partial_t n\rangle$, the work becomes

$$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$

where $\omega_{mn}=(E_m-E_n)/\hbar$. This form already shows that the effect is controlled by off-diagonal density-matrix elements in the instantaneous eigenbasis and by the Berry connection itself, rather than by a change of the spectrum (Fei et al., 13 May 2026).

A defining conceptual point is that chirality here refers to protocol orientation, not to handed quasiparticles or structural enantiomorphs. Other subfields use the same phrase for handedness-dependent energetic asymmetries, but the cyclic open-system definition is specifically an orientation-sensitive finite-time work difference.

2. Dissipative adiabatic perturbation expansion and the central formula

Under weak coupling to a thermal bath, Born–Markov dynamics, neglected Lamb shift, and the regime $\Sigma$0, the density matrix in the instantaneous eigenbasis obeys coupled equations for populations and coherences. The small adiabatic parameter is

$\Sigma$1

with $\Sigma$2 the driving period. Detailed balance ensures that populations vary only at order $\Sigma$3, so the leading-order structure is controlled by coherences (Fei et al., 13 May 2026).

The dissipative adiabatic perturbation expansion splits the coherences as

$\Sigma$4

where $\Sigma$5 is population-driven and $\Sigma$6 is multilevel feedback. The work correspondingly decomposes as

$\Sigma$7

Here $\Sigma$8 is the orientation-insensitive metric contribution, while $\Sigma$9 is the orientation-sensitive Berry contribution (Fei et al., 13 May 2026).

The geometric phase difference along the loop is

$$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$0

which can be written by Stokes’ theorem as a surface integral of the Berry-curvature difference,

$$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$1

Orientation reversal sends $$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$2, $$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$3, and $$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$4. As a result, the metric part is invariant, whereas the Berry-phase part changes sign (Fei et al., 13 May 2026).

The resulting chiral work difference is

$$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$5

with

$$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$6

For cyclic driving generated by a time-independent Hermitian $$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$7, $$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$8, the expression simplifies to

$$\Delta W_{\mathrm{chiral}} \equiv W_{\circlearrowright}-W_{\circlearrowleft}.$$9

In the special case of constant $$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$0, the dissipative contribution takes the surface-integral form

$$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$1

making the geometric control by Berry curvature explicit (Fei et al., 13 May 2026).

3. Unitary–dissipative crossover and the two-level realization

Two limiting regimes organize the phenomenon. In the unitary regime, $$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$2, one has

$$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$3

This is an interferometric thermodynamic Aharonov–Bohm effect: the signal depends on the product of a geometric factor $$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$4 and a dynamical factor $$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$5. In the dissipative regime, $$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$6, the coherent oscillations are exponentially suppressed and

$$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$7

The interference fringes disappear, but a finite orientation-sensitive work signal persists in the non-equilibrium periodic steady state (Fei et al., 13 May 2026).

A notable scaling statement is that the symmetric dissipation changes its scaling across the crossover, becoming the standard $$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$8 term in the dissipative limit, whereas the chiral signal remains $$\delta W=\mathrm{Tr}(\rho\, dH), \qquad \delta Q=\mathrm{Tr}(H\, d\rho),$$9 throughout. This persistence is central to the claim that the Berry contribution survives decoherence (Fei et al., 13 May 2026).

The explicit model used to illustrate the theory is a driven qubit with Hamiltonian

$$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$0

For this uniformly precessing two-level system, the Berry phase difference over one precession is

$$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$1

and the exact chiral work difference is

$$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$2

It vanishes at $$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$3 and at $$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$4; the latter is attributed to a chiral symmetry $$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$5 (Fei et al., 13 May 2026).

The same work also assesses experimental feasibility. For liquid-state NMR, typical parameters quoted are $$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$6, $$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$7, and effective spin temperature $$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$8. In the unitary regime with $$W=\oint dt\, \mathrm{Tr}(\rho_t \dot H_t)=\oint d\lambda_i\, \mathrm{Tr}(\rho\,\partial_{\lambda_i}H).$$9, the single-spin chiral work is estimated as $\{|n(\lambda)\rangle\}$0; ensembles of $\{|n(\lambda)\rangle\}$1 spins raise this to kHz. In the dissipative regime with $\{|n(\lambda)\rangle\}$2, $\{|n(\lambda)\rangle\}$3 per spin, while ensembles of $\{|n(\lambda)\rangle\}$4 spins amplify it to detectable levels (Fei et al., 13 May 2026).

4. Relation to quantum geometry and thermodynamic metrics

The open-system formulation places the chiral work difference within geometric thermodynamics. The orientation-insensitive part defines a thermodynamic Riemannian metric on control space,

$\{|n(\lambda)\rangle\}$5

This metric is tied to the symmetric real part of the quantum geometric tensor,

$\{|n(\lambda)\rangle\}$6

while the orientation-sensitive work is governed by the curvature $\{|n(\lambda)\rangle\}$7, the imaginary part (Fei et al., 13 May 2026).

The corresponding thermodynamic length is

$\{|n(\lambda)\rangle\}$8

and the metric contribution satisfies

$\{|n(\lambda)\rangle\}$9

with equality for constant thermodynamic speed. Subtracting clockwise and counterclockwise cycles removes this metric friction and leaves the Berry contribution. In that sense, $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$0 is the anti-symmetric partner of thermodynamic length: it is the part of finite-time work that changes sign under orientation reversal (Fei et al., 13 May 2026).

The formalism also clarifies gauge invariance. The physically relevant object is $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$1, gauge invariant modulo $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$2. In the unitary regime, $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$3 is strictly gauge invariant; in the dissipative regime, the surface-integral expression through curvature differences makes the geometric content manifestly gauge invariant. The validity conditions are likewise explicit: $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$4, $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$5, weak system–bath coupling, time-independent rates, and an initial instantaneous thermal state (Fei et al., 13 May 2026).

A related but conceptually distinct literature emphasizes that not every Berry-geometric chiral observable is already a work observable. In chiral-symmetric lattice systems, the Berry connection determines the mean chiral displacement of delocalized wavefunctions,

$\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$6

and in the narrow-wavepacket limit $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$7. That work explicitly states that the Berry connection does not, by itself, generate work; additional couplings are required to convert the displacement into an energy-transfer observable (Colandrea et al., 2024).

5. Related realizations in condensed-matter and transport settings

Outside open-system thermodynamics, the same phrase or closely related constructions appear in several branches of condensed-matter theory. In chiral insulators FeSi, RuSi, and OsSi, the electronic Berry phase $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$8 is nonzero and flips sign with crystal handedness. With spin–orbit coupling and a small magnetic field along $\partial_t|n\rangle=-i\sum_m A_{mn}|m\rangle$9, a magnetoelectric polarization $A_{mn}(t)=i\langle m|\partial_t n\rangle$0 appears, and opposite enantiomorphs yield opposite surface potential and work-function shifts. For identical terminations and surface normal parallel to $A_{mn}(t)=i\langle m|\partial_t n\rangle$1, the chiral work-function difference is estimated as

$A_{mn}(t)=i\langle m|\partial_t n\rangle$2

subject to attenuation by screening. The same work emphasizes that these systems are not topological insulators in the TME sense and that the numerical closeness of $A_{mn}(t)=i\langle m|\partial_t n\rangle$3 to $A_{mn}(t)=i\langle m|\partial_t n\rangle$4 is not evidence of quantized axion response (Kubler et al., 2013).

In chiral magnets, Berry-phase theories of the Dzyaloshinskii–Moriya interaction and phase-space Berry curvature produce an explicitly chirality-odd energetic bias. For cubic helimagnets the DM energy density can be written as

$A_{mn}(t)=i\langle m|\partial_t n\rangle$5

and reversing helicity changes its sign. The corresponding chiral work difference for opposite-helicity deformations is

$A_{mn}(t)=i\langle m|\partial_t n\rangle$6

For MnSi, an ab initio value $A_{mn}(t)=i\langle m|\partial_t n\rangle$7 per 8-atom cell was reported, in good agreement with experiment, and integrating the first-order free-energy correction over the magnetic unit cell of the skyrmion lattice yielded a free-energy reduction of $A_{mn}(t)=i\langle m|\partial_t n\rangle$8 per skyrmion and layer (Freimuth et al., 2013). A parallel Berry-phase formalism for DMI and spin-orbit torque interprets DMI as a spiralization and identifies the work asymmetry between opposite chiralities with the sign reversal of the linear-in-gradient DMI term (Freimuth et al., 2013).

In chiral kinetic theory, the phrase denotes a power difference rather than a cyclic thermodynamic work. The Berry Fermi-gas formulation defines

$A_{mn}(t)=i\langle m|\partial_t n\rangle$9

and for uniform fields $$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$0 gives

$$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$1

The same framework relates axial pumping to

$$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$2

with $$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$3 restored (Chen, 2016). A world-line QFT derivation reaches the same conceptual conclusion while stressing that Berry phase and chiral anomaly have distinct topological origins: the anomaly comes from the imaginary part of the fermion determinant, whereas Berry’s phase emerges from the real part in a non-relativistic adiabatic limit (Mueller et al., 2017).

A further variant occurs in bosonized nonlinear chiral edge modes. There the phase shift of periodic wave profiles splits into dynamical and Berry terms,

$$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$4

and for travelling waves the Berry contribution is explicitly chiral through $$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$5. The same work then connects this phase shift to a chiral work difference by coupling a cyclic drive to the $$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$6 zero mode, giving

$$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$7

for the compared propagation directions (Beauvillain et al., 2024).

6. Conceptual boundaries, counterexamples, and experimental diagnostics

The literature draws several boundaries around the concept. First, Berry geometry can produce handed observables without producing work. The strain-driven intermediate phase of a $$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$8 superconductor under in-plane strain acquires a Berry phase of $$W(t)=-\hbar \sum_{m\neq n}\omega_{mn}\int_0^t \mathrm{Im}[A_{nm}(t')\rho_{mn}(t')]\,dt',$$9 when the strain loop encloses the Dirac cone in the thermodynamic phase diagram, and that phase appears in the winding of the superfluid stiffness tensor and in the geometry of vortices and the upper critical field. Yet the quasistatic equilibrium cycle has

$\omega_{mn}=(E_m-E_n)/\hbar$0

because the minimized free energy is a state function on the branch followed by the system (Sun et al., 24 Nov 2025). This is a geometric response, but not a net thermodynamic work output.

Second, a claimed Berry-phase-induced chiral work channel can fail in materials whose transport is topologically trivial. In NbP, the dominant Shubnikov–de Haas frequency $\omega_{mn}=(E_m-E_n)/\hbar$1 gives a Landau-fan intercept $\omega_{mn}=(E_m-E_n)/\hbar$2, interpreted as a trivial Berry phase for that pocket, and no negative longitudinal magnetoresistance was observed up to $\omega_{mn}=(E_m-E_n)/\hbar$3 at $\omega_{mn}=(E_m-E_n)/\hbar$4. The study therefore argues against anomaly-driven chiral pumping under the measured conditions (Sudesh et al., 2016).

Third, surface or transport asymmetries should not be conflated with quantized topological responses unless the relevant topological criteria are met. In the chiral insulators FeSi, RuSi, and OsSi, the first Chern invariant was found to be zero within numerical accuracy, and the authors explicitly caution that the numerical proximity of the magnetoelectric slope to a topological magnetoelectric value is not evidence of a quantized axion response (Kubler et al., 2013).

Experimentally, the diagnostic depends on the realization. The open-system thermodynamic effect is accessed by measuring

$\omega_{mn}=(E_m-E_n)/\hbar$5

for opposite orientations of the same control loop; the predicted signatures are interference fringes in the unitary regime and a fringe-free residual signal in the dissipative regime (Fei et al., 13 May 2026). Chiral-insulator realizations suggest Kelvin probe force microscopy and ultraviolet photoelectron spectroscopy to resolve opposite work-function shifts of opposite enantiomorphs under $\omega_{mn}=(E_m-E_n)/\hbar$6 (Kubler et al., 2013). In strain-tuned superconductors, the relevant observables are the winding of $\omega_{mn}=(E_m-E_n)/\hbar$7, the inversion of the vortex semimajor axis, and the exchange of the two $\omega_{mn}=(E_m-E_n)/\hbar$8 maxima after a non-contractible strain loop (Sun et al., 24 Nov 2025).

Taken together, these results delimit a broad but internally structured research area. In the strict thermodynamic sense, Berry-phase-induced chiral work difference is an anti-symmetric finite-time work functional of loop orientation in open, slowly driven quantum systems (Fei et al., 13 May 2026). In allied senses, it names Berry-geometric handed energy or power asymmetries in chiral magnets, chiral kinetic theory, edge hydrodynamics, and chiral insulators. The common element is not a single microscopic mechanism, but a recurring rule: Berry connection, Berry phase, or Berry curvature contributes a term that is odd under orientation reversal, propagation reversal, or exchange of handedness, thereby generating a measurable chiral energetic asymmetry.