Coherent-state path integrals in quantum thermodynamics

This paper clarifies subtle aspects of coherent-state path integrals in quantum thermodynamics, demonstrating that when properly handled in the continuum limit, they yield results consistent with the canonical Hamiltonian approach across various paradigmatic bosonic and fermionic systems.

The Gist

The Big Picture: Two Ways to Count the Same Thing

Imagine you are trying to calculate the total "happiness" (or energy cost) of a massive crowd of people at a concert. In physics, this "happiness" is called the Partition Function, and it tells us everything about how a system behaves at a certain temperature.

There are two main ways physicists try to calculate this:

1. The "Operator" Way (Hamiltonian): This is like counting the crowd by looking at every single person individually, checking their name tag, and adding up their specific energy levels. It's precise, logical, and based on hard rules (quantum mechanics).
2. The "Path Integral" Way: This is like looking at the crowd as a flowing river. Instead of counting individuals, you look at the waves, the currents, and the overall flow. You treat the particles as if they are waves moving through time.

The Problem:
For decades, students and scientists have noticed that when they use the "Path Integral" method (the river), they sometimes get a slightly different answer than the "Operator" method (the individual count). It's like if you counted the crowd by waves and got 1,000 people, but counted them individually and got 1,002.

The Paper's Mission:
Salasnich and Vianello wrote this paper to say: "Don't panic! The river method is just as accurate as the individual count, but you have to be very careful about how you measure the waves." They show that if you fix a few subtle technical details (which they call "subtleties"), both methods give the exact same result.

---

The Main Characters: Bosons and Fermions

To understand the paper, you need to know the two types of "particles" in the quantum world:

• Bosons (The Socialites): Think of these as people who love to dance in a group. They can all stand on the same spot at the same time (like photons in a laser or atoms in a Bose-Einstein Condensate).
• Fermions (The Introverts): Think of these as people who need personal space. No two can ever be in the exact same spot at the same time (like electrons in an atom). This is the "Pauli Exclusion Principle."

The paper shows how to calculate the "happiness" of both types of crowds using the "river" (path integral) method.

---

The Secret Sauce: "Time-Ordering" and the "Convergence Factor"

This is the most technical part of the paper, but here is the analogy:

Imagine you are recording a video of a ball bouncing.

• The Operator Method is like taking a high-speed photo of the ball at every single instant.
• The Path Integral Method is like drawing a smooth line connecting all the dots.

The Trap:
When you draw that smooth line, you have to decide: Does the ball exist at time $t$, or at time $t$ plus a tiny, invisible fraction of a second?

In quantum mechanics, the order matters. If you measure the position of a particle before you measure its momentum, you get a different result than if you do it the other way around. This is called Time-Ordering.

The Mistake:
Many textbooks treat the "smooth line" as if it's perfectly continuous and ignore that tiny fraction of a second. They assume the ball is at time $t$ for both measurements.

• The Result: This tiny assumption error creates a "ghost" energy. It's like adding a phantom tax to your bill. For simple systems, it's a small error. For complex systems (like superconductors), it can lead to completely wrong predictions about how the material behaves.

The Fix (The "Convergence Factor"):
The authors say: "You must add a tiny, invisible 'convergence factor' to your math."
Think of this factor as a microscopic ruler that forces you to acknowledge that the "future" measurement happens a tiny bit after the "past" measurement.

• When you add this ruler, the "ghost energy" disappears.
• Suddenly, the "River" calculation matches the "Individual Count" calculation perfectly.

---

The Examples: From Simple to Complex

The paper walks through several examples to prove their point, like a teacher solving math problems on a blackboard:

1. The Simple Oscillator (The Swing):

   • Scenario: A single particle swinging back and forth.
   • Lesson: Even here, if you ignore the "time-ordering" rule, you get the wrong answer. The authors show how to fix the math so the "River" matches the "Swing."

2. The Single-Site Models (The One-Room Apartment):

   • Scenario: Imagine a tiny apartment where particles live. They can interact (bump into each other).
   • The Twist: When particles interact, the math gets messy. The authors use a trick called the Hubbard-Stratonovich Transformation.
   • The Analogy: Imagine two people arguing in a room. Instead of tracking their shouting match directly, you introduce a "messenger" (a new field) who carries the message between them.
   • The Catch: If the messenger is "noisy" (random), you have to be careful how you average the noise. The authors show that if you treat the noise correctly (using a rule called Itô calculus, which is like a specific way to handle random walks), the math works out perfectly.

3. The Weakly Interacting Bose Gas (The Dance Floor):

   • Scenario: A huge crowd of Bosons dancing.
   • The Goal: Calculate the energy of the whole crowd.
   • The Result: The authors show that if you use the "River" method with the correct "Time-Ordering" ruler, you get the exact same energy as the "Individual Count" method. This is crucial because the "River" method is often easier to use for huge crowds.

4. The BCS Superconductor (The Electron Pairing):

   • Scenario: Electrons (Fermions) pairing up to dance together, allowing electricity to flow with zero resistance.
   • The Danger: This is the most complex example. If you make the "Time-Ordering" mistake here, you don't just get a small error; you get the wrong number of electrons in the system.
   • The Fix: By applying their "convergence factor" rule, the authors show that the path integral correctly predicts the "Gap" (the energy needed to break the pairs) and the number of electrons, matching the standard theory perfectly.

---

Why Should You Care?

You might ask, "I'm not a physicist, why does this matter?"

1. It's About Trust: Science relies on different methods agreeing with each other. If two ways of calculating the same thing give different answers, we don't know which one is right. This paper says, "Both are right, but you have to be careful."
2. It Helps with New Tech: We are building quantum computers and new superconducting materials. To design these, we need to simulate how particles behave. If the simulation software uses the "Path Integral" method but ignores these subtle rules, the design might fail.
3. It's a Guide for Students: The authors wrote this specifically to help students who are confused by the "ghost errors" in their textbooks. They are saying, "You aren't crazy; the textbooks just skipped a step."

The Takeaway

The paper is a "user manual" for a powerful tool in physics. It tells us that the Path Integral method is a beautiful, powerful way to understand the quantum world, but it requires a specific "calibration" (the convergence factor) to ensure it doesn't drift off course.

In short: If you want to calculate the energy of a quantum system using waves instead of particles, make sure you respect the tiny difference between "now" and "a tiny bit later." If you do, the math will work perfectly, and the "River" will flow exactly where the "Individual Count" says it should.

Technical Summary

1. Problem Statement

The central problem addressed in this work is the subtle discrepancy that often arises between thermodynamic results derived from the canonical Hamiltonian approach (operator formalism) and those derived from the coherent-state path integral formalism when applied to quantum many-particle systems at finite temperature.

While the path integral method is a cornerstone of modern theoretical physics, standard textbook treatments often gloss over critical technical details regarding:

• The continuum limit of discretized path integrals.
• The correct handling of variable transformations (e.g., to number-phase representations).
• The regularization of Matsubara-frequency summations.
• The treatment of functional determinants, particularly when stochastic fields (like Hubbard-Stratonovich fields) are involved.

Neglecting these subtleties leads to incorrect results, such as missing interaction terms or spurious zero-point energy contributions, even in simple systems like harmonic oscillators or interacting Bose gases. The authors aim to demonstrate that these discrepancies are not fundamental failures of the path integral method but rather artifacts of improper mathematical handling in the continuum limit.

2. Methodology

The authors employ a rigorous, pedagogical approach comparing the canonical operator method with the coherent-state path integral method across a hierarchy of systems, increasing in complexity:

• Discretization and Continuum Limits: They begin by constructing the path integral on a discrete imaginary-time lattice ($M$ slices) and carefully analyzing the limit $M \to \infty$. They emphasize that the order of operators in the Hamiltonian (normal ordering) dictates the specific form of the action and the boundary conditions (periodic for bosons, anti-periodic for fermions).
• Hubbard-Stratonovich (HS) Transformation: For interacting systems, they utilize the HS transformation to decouple interaction terms. A key methodological innovation is the treatment of the HS field not as a smooth function but as a stochastic field with white-noise correlations. This necessitates the use of Itô calculus rules (specifically the Itô substitution rule) when taking the continuum limit, which introduces necessary correction terms (e.g., shifts in the chemical potential).
• Matsubara Frequency Summation: The paper extensively analyzes the summation of Matsubara frequencies in frequency space.
  • They highlight the necessity of convergence factors ($e^{i\omega_n 0^+}$) arising from the implicit time-ordering of the path integral.
  • They demonstrate that naive summation of symmetric frequency pairs (e.g., $\omega_n$ and $-\omega_n$) without these factors leads to divergent or incorrect results.
  • They show how to correctly regularize these sums using contour integration and the residue theorem, ensuring the results match the canonical approach.
• Matrix Form Actions: For systems requiring matrix actions (like Bogoliubov-de Gennes or BCS theories), they show that different components of the matrix may require different convergence factors due to time-ordering. They propose a method to rewrite the action with a uniform convergence factor by adding specific counterterms to the action, which cancel spurious energy shifts.

3. Key Contributions and Results

The paper validates the path integral formalism by reproducing exact canonical results for several paradigmatic systems:

A. Harmonic Oscillators (Bosonic and Fermionic)

• Result: The path integral yields the exact partition function $Z = (1 - e^{-\beta \epsilon})^{-1}$ (bosons) and $Z = 1 + e^{-\beta \epsilon}$ (fermions).
• Insight: This serves as a baseline, proving that the functional determinant of the operator $(\partial_\tau + \epsilon)$ correctly reproduces the trace of the Boltzmann operator when boundary conditions and convergence factors are handled correctly.

B. Single-Site Bose-Hubbard Model

• Problem: Previous attempts using a continuous number-phase transformation or naive HS transformation yielded an incorrect interaction term ($N^2$ instead of $N(N-1)$).
• Resolution: The authors show that the error stems from assuming the HS field is smooth. By treating the HS field as a stochastic variable with correlation $\langle \phi(\tau)\phi(\tau') \rangle \propto \delta(\tau-\tau')$, they derive an Itô correction.
• Result: This correction shifts the chemical potential by $g/2$, restoring the exact result $Z = \sum e^{-\beta[-\mu N + \frac{g}{2}N(N-1)]}$.

C. Single-Site Hubbard Model (Fermionic)

• Result: Using a two-component HS field, the authors show that because the fields are not auto-correlated, they behave as smooth functions. No Itô correction is needed, and the path integral yields the exact partition function $Z = 1 + 2e^{-\beta \epsilon} + e^{-\beta(2\epsilon+g)}$.

D. Weakly-Interacting Bose Gas

• Context: Application of the Bogoliubov approximation to a uniform gas with finite-range interactions.
• Methodology: The authors perform the path integral over fluctuation fields in the presence of a condensate. They carefully handle the matrix determinant of the inverse propagator.
• Result: They derive the grand potential $\Omega_G$ exactly matching the Hamiltonian result.
  • Crucial Finding: A naive summation of Matsubara frequencies (often found in literature) yields a spurious zero-point energy term ($\frac{1}{2}\sum E_k$) that differs from the canonical result. The authors show that this discrepancy is resolved only by correctly applying time-ordering convergence factors, which cancel the spurious term without needing ad-hoc regularization.

E. BCS Superconductor

• Context: Application to a spin-singlet superconductor with finite-range interactions.
• Methodology: Using the Nambu spinor formalism and HS transformation for the gap parameter $\Delta$.
• Result: The path integral yields the exact BCS grand potential and the correct gap equation and number equation.
• Significance: The authors demonstrate that for BCS theory, a naive evaluation of the determinant leads to an incorrect number equation (particle density). The correct treatment of time-ordering and convergence factors is essential to recover the self-consistent gap equation and the correct thermodynamic relations.

4. Significance

• Resolution of Discrepancies: The paper resolves long-standing confusion regarding why path integrals sometimes yield "wrong" results compared to operator methods. It establishes that the path integral is fully consistent with the canonical formalism if and only if the discretization, continuum limits, and time-ordering are treated with mathematical rigor.
• Pedagogical Value: It provides a unified framework for students and researchers to avoid common pitfalls, such as ignoring the stochastic nature of HS fields or mishandling Matsubara summations.
• Technical Rigor: The introduction of the Itô substitution rule for HS fields and the explicit construction of counterterms to unify time-ordering in matrix actions offers a robust mathematical toolkit for future calculations in quantum thermodynamics.
• Generalizability: The methods presented are applicable to a wide range of quantum many-body problems, including ultracold atomic gases, superconductors, and systems with finite-range interactions, which are often treated only at a mean-field level in standard texts.

In conclusion, the authors successfully demonstrate that the coherent-state path integral is a powerful and exact tool for equilibrium thermodynamics, provided that the subtle technical aspects of the continuum limit and frequency summation are handled with the same care as in the canonical operator approach.