Collision models: Markovian and Non-Markovian impurity models

arXiv:2610.01814 · quant-ph · Submitted 2026-10-01 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: Today's paper: "Collision models: Markovian and Non-Markovian impurity models".

Mira: The gist: If a p-impurity model is non-Markovian by the BLP measure, then the corresponding (p + 1)-impurity model is also non-Markovian,

Kai: First, who's behind it and why it matters.

Title and authors: Kai: So we’re looking at this paper titled "Collision models: Markovian and Non-Markovian impurity models," and it’s essentially setting up a way to study open quantum systems using these collision models.

Mira: Exactly. It introduces a new class of these models, which the authors call Markovian impurity models, which they say are motivated by situations where information backflow is periodically blocked.

Kai: Right, so the core idea here is that if you start with a p-impurity model and it’s non-Markovian—meaning it has memory—then adding one more impurity, going from a p to a p plus one model, keeps it non-Markovian too.

Lev: That implies there has to be some critical value of p where the process switches from being Markovian to being non-Markovian at all >.

Kai: And the authors use this concept of a critical p, and they look at how that number changes depending on the strength of the system-ancilla and ancilla-ancilla interactions.

Mira: They found that generally increasing either interaction strength lowers that critical p, which means you can tune your system to stay Markovian for longer if you want.

Kai: And there’s a specific regime where even when you don't have any impurities at all, if the ancilla-ancilla interactions are weak enough, the dynamics stay Markovian.

Lev: That’s important because it tells us that sometimes environmental memory isn't necessary for the system to behave in a simple way.

Kai: The paper uses the Breuer–Laine–Piilo measure of non-Markovianity, which is based on the idea that Markovian dynamics can't increase how distinguishable two states are.

Mira: They define this measure by summing up all the increases in trace distance over time, and if a process is Markovian, that total sum has to be zero.

Kai: The p-impurity model itself has a specific structure where the map changes at fixed time intervals, specifically at times t equal to p times N.

Lev: They also mentioned another model called the non-Markovian impurity model, which they say is essentially the opposite of the Markovian one they focused on.

Kai: So what does this mean for us in practice? It suggests that in certain physical scenarios, you can systematically tune how much memory your system has by changing that impurity parameter p.

Mira: The numerical results show something interesting about the relationship between those interaction strengths and this critical p, and they found it’s quite nontrivial.

Kai: They specifically looked at how increasing the system-ancilla interaction strength, denoted as thetaXY, affects that critical value.

Lev: And they also looked at the ancilla-ancilla interaction strength, thetaYY, and how that changes things when you increase thetaXY.

Kai: The simulations showed that for high enough values of those interaction strengths, even with a p of two impurities, the process is still non-Markovian <ref:2610.01814#pg2>.

Mira: They also found a functional relationship for that value of thetaYY where the asymptote occurs as thetaXY increases; it generally goes down.

Lev: This suggests that if you want to push the limits of what your system can do before it becomes memory-dependent, you need to be careful about those interaction parameters.

Kai: So, to wrap up this paper on "Collision models: Markovian and Non-Markovian impurity models," they’ve shown a systematic way to find the point where non-Markovianity kicks in based on how many impurities you have.

Mira: The main implication is that there's a crossover point, that critical p, which dictates when the dynamics shift from being Markovian to having memory effects.

Lev: For someone working on quantum error correction or simulating hardware, this gives them a concrete parameter—that critical p—to target when designing experiments.

Kai: And they pointed out a limitation of their study: they only focused on what happens when the incoming ancillas are initially uncorrelated with everything, which might not cover all physical situations.

Mira: That’s fair, they acknowledged that the model is restricted to certain assumptions about the initial state of those ancillas.

Lev: But for now, this framework lets us predict when environmental memory will start to matter in our quantum dynamics.

The paper's summary: Kai: So, we're looking at how this paper summarizes its main finding on these collision models—basically showing that if you start with a p-impurity model and it has memory, then adding more impurities makes it even more non-Markovian.

Mira: Right, so the core takeaway is this new idea of a critical p value. They’re saying there’s a specific point in the impurity frequency where the whole system switches from behaving like something simple, Markovian dynamics, to something with real memory effects.

Kai: That means for anyone building quantum hardware or simulating these open systems, they can actually tune it. If you want your device to stay predictable and simple for a certain range of impurity settings, you just need to stay below that critical p threshold.

Lev: From an error-correction standpoint, that’s interesting because if we know the critical p, we know exactly where the dynamics start getting complicated enough that standard Markovian assumptions break down. It gives us a boundary condition for what we can expect from our models.

Mira: And they showed numerically how this critical p shifts based on your interaction strengths—both the system-ancilla interactions and those between the ancillas themselves. Increasing those interaction strengths generally makes that critical p smaller, which means you get Markovian behavior over a wider range of parameters.

Kai: That’s a useful design tip, I guess. So, even if you're trying to keep things simple and Markovian, you can still manipulate your settings to push the non-Markovian boundary further out by tweaking those couplings.

Lev: Exactly. And they also pointed out a specific scenario where things get interesting even without any impurities present at all—if the ancilla-ancilla interactions are weak enough, the dynamics stay Markovian regardless of the impurity count.

Mira: That's a big deal for understanding when environmental memory is actually needed versus when it’s just an artifact of how you set up your initial conditions. It suggests that sometimes you can get away with simpler models if those internal couplings are tuned just right.

Kai: So, to summarize the main point, this paper gives us a systematic way to find where non-Markovianity kicks in based on the number of impurities you introduce and how strongly everything is coupled together.

Lev: And that leads us to thinking about what happens when we look at those specific numerical results—how p relates to those interaction parameters—because that's where the real constraints on implementing this theory come into play.

The paper's improvements: Tom: So, we’re looking at how this paper suggests we can actually improve these collision models—they’re proposing ways to make them more useful for real physics, and they're focusing on what happens if you add more complexity to the system structure.

Kai: They suggest that by moving beyond just a single impurity frequency, we can build systems where the dynamics are even richer because you get this relationship between p and those interaction parameters being really nontrivial.

Mira: That’s interesting because it means we aren't stuck with just one way to model these impurities; there’s a whole landscape of possibilities based on how those coupling strengths interact. It opens up new avenues for modeling systems that have more complex environmental interactions.

Lev: For error correction, if the dynamics are this flexible, we might be able to design error-correcting codes that can better handle these non-Markovian effects without having to rely on overly simplified assumptions about the environment.

Kai: They also suggest looking at how increasing those interaction strengths affects that critical p value specifically, which is a way for experimentalists to optimize their setup. So you can deliberately tune your hardware interactions to push the boundary of what you can achieve before you hit a memory barrier.

Mira: And they flag this limitation—the paper only looks at Markovian impurity models initially—but they suggest that by exploring these higher-order structures, we're setting the stage for understanding even more complex, realistic open quantum systems.

Lev: It changes how we think about system design because instead of just trying to eliminate memory effects entirely, you can now design for a specific level of memory based on these calculated critical points. That’s a practical shift for building scalable quantum processors.

Kai: So the big improvement here is moving from just identifying a single threshold to understanding how that threshold evolves when you change the internal dynamics of the collision itself.

Mira: And this leads right into what they call that non-Markovian impurity model—which, as we touched on before, is essentially the opposite of their main focus and shows us another side of these memory effects.

Conclusion: Kai: So, to wrap up this discussion on "Collision models: Markovian and Non-Markovian impurity models," the main point is that we’ve found a systematic way to map out when these open quantum systems stop behaving simply and start showing real memory effects based on how many impurities are present.

Mira: Exactly, they showed that if you have a p-impurity model, you can predict whether it will be non-Markovian by looking at the Breuer–Laine–Piilo measure of non-Markovianity.

Kai: It’s about tuning parameters like those interaction strengths to find a critical p where the system transitions from Markovian to something more complex.

Lev: For error correction, that means we have a metric for when standard assumptions fail, which is helpful for designing more robust codes.

Mira: The implication is that this framework allows us to predict non-Markovian behavior in certain physical scenarios and even suggest ways to tune those systems away from it by adjusting couplings.

Kai: It’s about finding that crossover point, and they showed numerically how increasing the interaction strength generally makes you want a smaller critical p.

Lev: That means if you're building something, you can optimize your hardware parameters to keep the dynamics simple for a longer time.

Mira: And even when there are no impurities at all, if those ancilla interactions are weak enough, the system stays Markovian even without any of those added noise sources.

Kai: So we’ve established that there’s a boundary condition where memory becomes essential in these collision models.

Lev: It gives us a concrete target for what kind of dynamics we should expect in experiments before we worry about non-Markovian corrections.

Mira: This work sets the stage for deeper exploration into how these structures evolve when you introduce even more complexity, which is where things get really interesting next.

Jacob Werner

Department of Physics, The University of Tokyo

quant-ph

Submitted: 2026-10-01

Updated: 2026-10-01

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 66/100

The gist: The gist: If a p-impurity model is non-Markovian by the BLP measure, then the corresponding (p + 1)-impurity model is also non-Markovian, implying the existence of a critical p at which the process

Key concepts

Collision Models (CM): CMs are a framework for studying open quantum systems where the environment is modeled as a sequence of ancillas that sequentially interact with the system. Each interaction, called a collision, represents a discrete time step in the dynamics.
Markovian Impurity Model: This model involves inserting 'impurities' at fixed time intervals within a Collision Model. It is characterized by specific unitary operations and is defined as a repeated completely positive and trace-preserving (CPTP) map, often related to partial swap operations.
BLP Measure of Non-Markovianity: This measure quantifies non-Markovian behavior by summing all increases in the trace distance between two states over time. If the process is Markovian, this sum must be zero, meaning dynamics cannot increase the distinguishability between states.
Critical p: This refers to a specific impurity frequency where the system's dynamics switch from being Markovian to non-Markovian. The paper shows that if one model is non-Markovian at frequency 'p', the (p+1)-impurity model will also be non-Markovian, implying this transition point exists.

Terminology

Non-Markovianity — 19× in this paper · explained in 3 episodes
Trace — 17× in this paper · explained in 2 episodes
Trace Distance — 11× in this paper · explained in 3 episodes
Core — 2× in this paper · explained in 2 episodes
Using the Breuer-Laine-Piilo — 1× in this paper

Summary

The gist: If a p-impurity model is non-Markovian by the BLP measure, then the corresponding (p + 1)-impurity model is also non-Markovian, implying the existence of a critical p at which the process transitions from Markovian to non-Markovian

Collision Models and Non-Markovianity

Collision models provide a flexible framework for studying open quantum systems and introducing non-Markovianity. In a CM, the environment is modelled as a series of ancillas that sequentially interact with the system, where each interaction is called a collision and designates a discrete time step The primary contribution of this paper is the introduction and analysis of the Markovian impurity model, which may arise in certain physical settings where information backflow is periodically blocked.

Markovian Collision Models

In a Markovian CM, each incoming ancilla (Y) is initially uncorrelated with both the system (X) and other ancillas. The state of the system at time t, following the tth collision, is given by a specific equation involving the unitary governing the system-ancilla interaction. These models are characterized as repeated completely positive and trace preserving (CPTP) maps. The interactions are often described by partial swap operations parameterized by theta, of the form U(θ) = cos(θ)I + isin(θ) S.

Non-Markovian Collision Models and Measures

The paper employs the Breuer–Laine–Piilo (BLP) measure of non-Markovianity to define non-Markovian processes. This measure is based on the principle that Markovian dynamics cannot increase the distinguishability between two states, which is quantified by the trace distance D(ρ1, ρ2). The NBLP is obtained by summing all increases in the trace distance over time. If a process is Markovian, the trace distance between two states must decrease so that NBLP = 0.

Markovian Impurity Model

The p-impurity model is defined by inserting “impurities” at fixed time intervals. For every t ∈ pN the map is given by a specific form involving TrYt and UXY. For all other time steps, the map is given by a different form involving Y Y interactions. A 1-impurity model reduces to the Markovian CM of Eq. (1).

Critical Impurity Frequency

The paper aims to show that if a p-impurity model is non-Markovian (NBLP > 0), then the corresponding (p + 1)-impurity model is also non-Markovian. This implies the existence of a critical p at which the process transitions from Markovian to non-Markovian. Numerical investigations show that increasing either interaction strength generally lowers the critical p, while sufficiently weak ancilla–ancilla interactions lead to a regime in which the dynamics remain Markovian even in the absence of impurities.

Numerical Results and Critical p

Numerical simulations determine if and when the model transitions to being non-Markovian, where NBLP is calculated using the σx eigenstates as the initial state pair. The relationship between p′(θXY, θY Y) is shown to be nontrivial. For high enough θXY and θY Y, even p = 2 results in NBLP > 0. The functional relationship of the value of θY Y at which the asymptote occurs, θ′Y Y (θXY), is shown to be such that as θXY is increased, θ′Y Y generally decreases.

Non-Markovian Impurity Model

The non-Markovian p-impurity model involves a map that is a CPTP map except for t ∈ pZ + + 1. The non-Markovian impurity model is essentially the opposite of the Markovian impurity model given by Eqs. (10)–(11). This model is considerably less interesting and only mentioned here for completeness.

Conclusion

The overarching goal of this paper is to introduce a novel class of non-Markovian CMs that may arise in certain physical scenarios. We showed that if a p-impurity model is non-Markovian by the BLP measure, then the corresponding (p + 1)-impurity model is also non-Markovian. This suggests a crossover p at which the process becomes non-Markovian, and we call this the critical p. We presented numerical results that show that in general increasing θXY or θY Y reduces, or leaves unchanged, the critical p. We also showed that upon decreasing θY Y, an asymptote is approached beyond which the process remains Markovian even for p = ∞. We then briefly mentioned the non-Markovian impurity model. Although several interesting features of Markovian impurity models have been identified, a systematic exploration of the model lies beyond this paper’s scope and remains an open direction for future work.

References

[1] F. Ciccarello, S. Lorenzo, V. Giovannetti, and G. M. Palma, “Quantum collision models: Open system dynamics from repeated interactions,” Physics Reports, vol. 954, p. 1–70, Apr. 2022

[3] H.-P. Breuer, E.-M. Laine, and J. Piilo, “Measure for the Degree of Non-Markovian Behavior of Quantum Processes in Open Systems,” Physical Review Letters, vol. 103, Nov. 2009

[4] G. Lindblad, “Completely Positive Maps and Entropy Inequalities,” Communications in Mathematical Physics, vol. 40, pp. 147–151, 1975

[5] H. T. Şenyaşa, ff. Kesgin, G. Karpat, and B. Çakmak, “Entropy Production in Non-Markovian Collision Models: Information Backflow vs. System-Environment Correlations,” Entropy, vol. 24, p. 824, June 2022

[6] R. McElvogue, A. K. Mitchell, G. T. Landi, and S. Campbell, “Nonequilibrium steady states in multibath quantum collision models,” Physical Review A, vol. 113, Feb. 2026

[7] The paper is structured as follows

--- Page 5 ---

Figure 3: p′ is plotted against θY Y for different values of θXY, with ρY = ρth(0.3)

Figure 4: p′(θXY, θY Y) in Fig. 4 for clarity

Figure 5: The maximum value of θY Y for which p = ∞ still satisfies NBLP = 0 is plotted against θXY

--- Page 6 ---

Figure 3: p′ is plotted against θY Y for different values of θXY, with ρY = ρth(0.3)

Figure 4: p′(θXY, θY Y) in Fig.

Improvements for AI systems

Improved AI systems can accurately model dynamics in open quantum systems with memory by implementing collision models that incorporate periodic blocking of information backflow, which allows for a systematic variation of non-Markovianity via an impurity parameter p.
The improved system can determine the existence of a critical p at which the process transitions from Markovian to non-Markovian, providing a quantitative measure for when environmental memory becomes important in quantum dynamics.
The AI can predict whether a given physical scenario described by these models will exhibit non-Markovian behavior by calculating the Breuer–Laine–Piilo (BLP) measure of non-Markovianity, which is based on the principle that Markovian dynamics cannot increase the distinguishability between two states.
The system can identify regimes where dynamics remain Markovian even in the absence of impurities, specifically when sufficiently weak ancilla–ancilla interactions lead to a regime in which the dynamics remain Markovian even in the absence of impurities.
The AI can assess the impact of interaction strengths on non-Markovianity by showing that increasing either interaction strength generally lowers the critical p, allowing for optimized system design based on these parameters.

Abstract

Collision models provide a flexible framework for studying open quantum systems and introducing non-Markovianity. We present a novel class of collision models, which we term Markovian impurity models, motivated by physical scenarios in which information backflow is periodically blocked. Using the Breuer-Laine-Piilo (BLP) measure of non-Markovianity, we show that if a p-impurity model is non-Markovian, then the corresponding (p+1) -impurity model is also non-Markovian, implying the existence of a critical p at which the process transitions from Markovian to non-Markovian. Numerical simulations are used to investigate how this critical parameter depends on the system-ancilla and ancilla-ancilla interaction strengths. We find that, altough exceptions exist, increasing either interaction strength generally lowers the critical p, while sufficiently weak ancilla-ancilla interactions lead to a regime in which the dynamics remain Markovian even in the absence of impurities. Finally, we briefly introduce the complementary non-Markovian impurity model.