🔬 condensed matter

Jarzynski equality for counterwork under reversed memory-filtered driving

This paper introduces a counterwork functional derived from a sign-inverting memory-filtered protocol that satisfies a reciprocal Jarzynski equality with the original work, establishing that the sum of their average values is non-negative and that robust strategies under incomplete information should prioritize endpoint reversal while minimizing dissipated counterwork.

Original authors: Pierre Nazé

Published 2026-06-29

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Pierre Nazé

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are pushing a heavy box up a hill. This is your original work. You start at the bottom (Point A), push the box to the top (Point B), and you get tired. In physics, we measure how much energy you spent and how much "potential" the box gained at the top.

Now, imagine a second person, let's call them the "Counter-Operator." Their job is to take that box from the top (Point B) and push it back down to the bottom (Point A).

This paper is about a very specific, clever way the Counter-Operator does their job, and the surprising math that links your effort to theirs.

The Magic "Memory Filter"

Usually, if you push a box up a hill, the ground just sits there. But in this paper, the Counter-Operator uses a special tool: a Memory Filter.

Think of this filter like a smart, active mirror. Instead of just reacting to the box, the mirror inverts the instructions.

When you pushed forward (up), the mirror tells the Counter-Operator to push backward (down).
When you pushed fast, the mirror might tell them to push slowly, or vice versa, depending on the "kernel" (the settings of the mirror).

The paper introduces a specific setting for this mirror: a "Sign-Inverting Kernel." This is just a fancy way of saying the mirror is programmed to flip the direction of your movement perfectly. If you went from A to B, the mirror ensures the Counter-Operator goes from B back to A.

The "Counterwork"

The energy the Counter-Operator spends is called Counterwork.

If the box is heavy and you had to push hard to get it up, the Counter-Operator might be able to let gravity help them, effectively extracting energy (getting work out of the system) as they push it down.
However, if the hill is tricky or the box is stuck, the Counter-Operator might actually have to push harder to get it down, spending their own energy.

The paper proves a fascinating rule: The average energy you spend and the average energy the Counter-Operator spends are locked together.

The "Magic Math" (Jarzynski's Equality)

The paper uses a famous physics rule called Jarzynski's Equality. Think of this as a law of conservation for "effort averages."

The paper shows that if you look at the exponential average of your effort and the exponential average of the Counter-Operator's effort, they multiply to equal 1.

Simple translation: If your average effort was huge (making the box go up a steep hill), the Counter-Operator's average effort (bringing it down) will be mathematically linked in a way that balances the equation.
The Catch: This doesn't mean that for every single trip, your effort is exactly the opposite of theirs. Sometimes you might slip, and they might slip differently. But on average, across many, many trips, the math holds up perfectly.

The "No Free Lunch" Rule

The paper uses a mathematical trick (Jensen's inequality) to show a very practical limit: You can't get something for nothing.

If the Counter-Operator manages to get energy out of the system (negative work, like a generator), it is only possible because you put a lot of energy in during the first step.

Analogy: Imagine you pay \ $100 to climb a mountain. On the way down, the Counter-Operator might be able to harvest \$ 80 of that energy. But they can't harvest \ $100 unless you paid \$ 100.
The paper proves that the sum of your average effort and their average effort must always be zero or positive. You can't have a scenario where you do nothing, and they get free energy, or where they get free energy without you paying the price first.

The "Blind" Counter-Operator

The paper also discusses a scenario where the Counter-Operator doesn't know the details of the hill (the "free-energy landscape"). They just know they have to get the box from Point B back to Point A.

The Strategy: Since they don't know if the trip down will be easy (releasing energy) or hard (requiring energy), their best strategy isn't to guess. Instead, they should just focus on reversing the path perfectly and moving as smoothly as possible to avoid wasting energy on friction or mistakes.
The "Memory Filter" ensures they go from B to A, but it doesn't guarantee they will make money. That depends entirely on the shape of the hill you climbed in the first place.

Summary

In short, this paper describes a system where a second process (the Counter-Operator) is mathematically forced to reverse the path of a first process (the Original Work) using a special "inverting mirror."

The Link: The statistical averages of the energy spent in both directions are reciprocally related.
The Limit: You cannot extract net energy from the universe by doing this. If the second person gets energy out, the first person must have put it in.
The Goal: The best way to run this "reverse" process is to simply ensure the path is perfectly reversed and as smooth as possible, rather than trying to force a specific energy outcome.

The paper concludes that while this "sign-inverting" mechanism is a clever theoretical tool, it obeys the same strict laws of thermodynamics as everything else: Energy is conserved, and you can't cheat the system.

Technical Summary: Jarzynski Equality for Counterwork under Reversed Memory-Filtered Driving

Problem Statement
Fluctuation relations, particularly Jarzynski's equality, provide exact constraints on the statistics of work performed on systems driven far from equilibrium. While the standard equality relates the exponential average of work to the equilibrium free-energy difference for a given protocol $\lambda(t)$ , this work addresses a complementary construction: the thermodynamic status of a "counterwork" process. The central problem is to define and analyze a counterwork functional generated by an effective protocol $\Lambda(t)$ that reverses the endpoints of an imposed protocol $\lambda(t)$ . Specifically, the paper investigates how to construct such a reversed protocol using a memory-filtered driving mechanism that is not a passive bath response but an active component of the driving apparatus, and how this affects the application of Jarzynski's equality and the Second Law.

Methodology
The authors introduce a formalism where an effective protocol $\Lambda(t)$ is generated from an imposed protocol $\lambda(t)$ via a memory-filtered velocity:
$\dot{\Lambda}(t) = \int_0^\tau K(t, t') \dot{\lambda}(t') dt'$
Crucially, the kernel $K(t, t')$ is interpreted as an active protocol-memory kernel within the external driving apparatus, allowing for sign-changing properties (inversion) rather than representing a passive thermal response.

The "counterwork" $C$ is defined as the ordinary Hamiltonian work associated with the effective protocol $\Lambda(t)$ acting on the Hamiltonian family $H(\Gamma, \Lambda)$ :
$C = \int_0^\tau \frac{\partial H}{\partial \Lambda}(\Gamma_t, \Lambda(t)) \dot{\Lambda}(t) dt$
The analysis assumes the system is initially prepared in the canonical equilibrium state corresponding to the start of the respective protocol ( $\lambda(0)$ for the original work $W$ , and $\Lambda(0)$ for the counterwork $C$ ). The authors derive the conditions under which $\Lambda(t)$ reverses the endpoints of $\lambda(t)$ (i.e., $\Lambda(0) = \lambda(\tau)$ and $\Lambda(\tau) = \lambda(0)$ ) and analyze the resulting statistical relations.

Key Contributions and Results

Reciprocal Fluctuation Relations:
When the effective protocol reverses the endpoints of the original protocol, the free-energy difference associated with the counterwork satisfies $\Delta F_C = -\Delta F_W$ . Consequently, the Jarzynski equality for the counterwork yields:
$\langle e^{-\beta C} \rangle = e^{-\beta \Delta F_C} = e^{\beta \Delta F_W} = \frac{1}{\langle e^{-\beta W} \rangle}$
This establishes that the exponential average of the counterwork is the reciprocal of the exponential average of the original work. The authors emphasize that this relation holds for the exponential averages of two distinct processes starting from different canonical preparations, not for the trajectory-by-trajectory equality $C = -W$ .
Kernel Normalization for Reversed Displacement:
The paper derives the specific condition required for the memory kernel to guarantee a reversed displacement ( $\delta \Lambda = -\delta \lambda$ ) for arbitrary protocol velocities. This condition is a sign-reversing normalization:
$\int_0^\tau K(t, t') dt = -1 \quad \text{for } 0 \le t' \le \tau$
An example is provided using a constant kernel $K(t, t') = -1/\tau$ , which transforms a linear forward protocol into a linear backward effective protocol.
Bounds on Average Work and Counterwork:
By applying Jensen's inequality to the two fluctuation relations, the authors derive a bound on the product of the exponentials of the average work and counterwork:
$e^{-\beta \langle C \rangle} e^{-\beta \langle W \rangle} \le 1 \implies \langle C \rangle + \langle W \rangle \ge 0$
This result implies that negative average counterwork (work extraction) is thermodynamically possible but only if it is compensated by a sufficiently large positive average work in the original operation. The total average work of the combined forward and counter stages cannot be negative.
Second Law Consistency:
The construction is shown to be consistent with the Second Law. The sign-inverting kernel changes the effective driving protocol but does not violate the thermodynamic bound applied to the process generated by $H(\Gamma, \Lambda(t))$ . The dissipated counterwork $C_{diss} = \langle C \rangle - \Delta F_C$ remains non-negative. The sum of dissipated work and dissipated counterwork satisfies $W_{diss} + C_{diss} \ge 0$ , with equality only in the quasistatic limit.
Counteroperation under Incomplete Information:
The paper analyzes the scenario where the counteroperator does not know the sign of the free-energy difference $\Delta F_W$ in advance. Since $\Delta F_C = -\Delta F_W$ , the counteroperation may be either work-extracting or work-consuming depending on the free-energy landscape. The authors argue that the robust thermodynamic strategy is not to prescribe the sign of the counterwork, but to enforce endpoint reversal while minimizing the dissipated counterwork (approaching a quasistatic transformation).

Significance and Claims
The paper claims to provide a rigorous thermodynamic framework for "counterwork" generated by active, memory-filtered driving. Its primary significance lies in clarifying that while a sign-inverting kernel can reverse the thermodynamic endpoints and generate a reciprocal fluctuation relation, it does not allow for the bypassing of thermodynamic bounds. The sign-inverting kernel acts as a controlled reversal mechanism that maps an imposed protocol to an effective one with reversed displacement, but the energetic cost or gain is strictly determined by the underlying free-energy landscape and the Second Law.

The authors explicitly state that a complete energetic description of the full cycle, including heat exchange during the necessary equilibration steps between the forward and counter processes, is left for future work. The current contribution is limited to establishing the counterwork functional, the kernel conditions for reversal, and the resulting fluctuation relations and inequalities.