On deforming and breaking integrability

summary

The gist

In this paper we study nearest-neighbour deformations of integrable models by expanding in the deformation parameter to identify four possible types of deformations and discuss their effect on…

In short

The study examines nearest-neighbor deformations of integrable models by expanding in a deformation parameter to classify four types of integrability breaking scenarios. Using boost operator formalism, it analyzes spectral statistics to observe chaos onset, finding that the critical coupling scales intermediately between strong and weak integrability breaking models.

Key concepts

Classification of Integrability Breaking Scenarios
The authors categorize deformations into four types based on how they affect integrability. These include simple breaking/preservation, exactly integrable perturbations, those requiring all orders of the deformation parameter for integrability, and those integrable only to a certain order.
Boost Operator Formalism
This mathematical tool is used to classify deformations by imposing the condition for integrability: [Q2(ϵ), Q3(ϵ)] = 0. This method allows researchers to study deformations directly by finding linear equations for first-order terms, revealing how models break integrability at different orders.
Spectral Statistics and Chaos Onset
The paper compares level spacing distributions between integrable (Poisson) and chaotic systems (Wigner's surmise). It shows that the onset of chaos occurs at different deformation strengths for various models, providing a way to distinguish the types of integrability breaking.

Terminology used across episodes

This episode discusses

The paper

On deforming and breaking integrability · Read on arXiv

Ysla F. Adansa, Marius de Leeuwa, Tristan McLoughlina

School of Mathematics & Hamilton Mathematics Institute · Universidade Estadual Paulista Instituto de Física Teórica

In this paper we study nearest-neighbour deformations of integrable models. After expanding in the deformation parameter, we identify four possible types of deformations. First there are deformations that simply break or preserve integrability. Then we find two different subtle cases. The first case is where the deformation is only integrable if all orders of the deformation parameter are taken into account. An example of these are the long-range deformations that appear in holographic models. The second case is when the deformation is perturbatively integrable to some order in the deformation parameter but can not be extended to an integrable model. In this paper we work this out for the XXZ spin chain and discuss the level statistics of each of these cases. We find numerical evidence that the onset of chaos occurs differently in each of these models. For the perturbatively integrable models, we find that the deformation strength at which chaos appears demonstrates a volume-scaling intermediate between strong and weak integrability breaking models.

Transcript

Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.

Kai: I'm Kai.

Mira: And I'm Mira, here with Lev.

Kai: We're looking at the paper: "On deforming and breaking integrability".

Mira: Here's the gist of the paper. In this paper we study nearest-neighbour deformations of integrable models by expanding in the deformation parameter to identify four possible types of deformations and discuss their effect on spectral statistics The onset of chaos occurs differently in each case, and for perturbatively integrable models, the deformation strength at which chaos appears demonstrates a volume-scaling intermediate between strong and weak integrability breaking models Classification of Integrability Breaking Scenarios The authors identify four possible types of deformations after expanding in the deformation parameter First, there are deformations that simply break or preserve integrability Second, there are models for which the perturbation is exactly integrable The third case is where the deformation is only integrable if all orders of the deformation parameter are taken into account The fourth case occurs when the deformation is integrable to some order but cannot be extended to an integrable model beyond that For example, a specific deformation of the XXZ spin-chain is presented as an example of this fourth case Analysis using Boost Operator Formalism The study employs the boost operator formalism to classify deformations by imposing the necessary condition for integrability, namely

Q2(ϵ), Q3(ϵ): = 0 order by order in epsilon This approach allows for a direct study of deformations rather than relying on the indirect classification using the boost operator method The analysis involves finding linear equations for H(1) and G(1) at first order, which leads to quadratic constraints for H(1) and linear equations in H(2) and G(2) This method reveals models that are integrable to, for example, first order in epsilon but break integrability one order higher Spectral Statistics and Chaos Onset The paper numerically analyzes the spectral statistics of these deformed models to observe the onset of quantum chaos For an integrable model, level spacings are expected to follow a Poisson distribution P(s) = e-s For a chaotic system, the distribution is expected to closely follow Wigner’s surmise PGOE(s) = 1/2 πse− π/4 s squared The comparison between the two models shows that the onset of chaos occurs at larger values of epsilon for the quasi-integrable model HQInt compared to HdXYZ Volume Dependence and Scaling The volume dependence of the critical coupling, ϵc, is investigated to understand the scaling behavior near integrability breaking For gapless systems, numerical evidence suggests that the critical coupling scales as ∼ L-3 For gapped systems, a faster-than-power-law dependence on system size is proposed The quasi-integrable model HQInt exhibits a critical coupling scaling as ϵc ∼ L-b with an intermediate exponent between those expected for a generic gapless system b = 3 and the weakly deformed XXZ model b = 2 Eigenvector Entropy Measures The information entropy of the eigenstates is computed to quantify delocalization and ergodicity in the transition region For integrability breaking deformations, the entropy becomes a relatively smooth function of the energy as one would expect in a thermalised system In contrast, for integrable or quasi-integrable deformations, the entropy shows little sign of decreasing and does not become a smooth function of the energy The numerical studies confirm that the models with broken integrability, HdXYZ, show a relatively sharp onset of chaos with an exponential scaling of the critical coupling ϵc ∼ e-dL The quasi-integrable model HQInt exhibits an intermediate character between strong-breaking and weak-breaking results in terms of scaling It is suggested that the quasi-integrable model might be a partially delocalised perturbation in Fock space, generating an onset of chaos intermediate between a transition and crossover The paper suggests that it would be interesting to explore whether the quasi-integrable model thermalises slower than generic integrability breaking models on time-scales τ ∝ ϵ-b with b > 2 A Second Order G coefficients The second order G coefficients for Model A are explicitly displayed, involving terms like g∆(2)xy and gh(2)xy,+ For Model B, the second order constraints on the first order parameters are given by gδ(2)x,+ = 0 and gδ(2)y,+ = 0 Data Preparation To compare spectral statistics with RMT predictions, the authors employ "clipping" of the spectrum to remove states from the ends of the spectrum The clipping proportion affects both the maximum value of the Brody parameter, ωmax, and the critical coupling ϵc at which ω(ϵc) = 1/2 For reliable numerical results, a clipping proportion around 20% is chosen Operator Norms The operator norm, specifically the normalized Frobenius norm OF = s Tr(O†O) Tr(1), is used to quantify the size of deformations relative to the undeformed Hamiltonian The Frobenius norms for the deformations are found to be clearly linear in L, with HdefF = 2.28 + 0.21L for HdXYZ and HdefF = 5.80 + 0.48L for HQInt An effective deformation strength is defined as epsiloneff = H(ϵ = 1) − H(ϵ = 0)F / H(ϵ = 0)F References

1: M.

Kai: De Leeuw, A.

Mira: Pribytok, and P.

Kai: Ryan, “Classifying two-dimensional integrable spin chains,” J.

Paper summary: Kai: So we're looking at this paper today called "On deforming and breaking integrability." Basically, they take these perfect integrable models, like the XXZ spin chain, and start messing with them by adding small changes. They want to figure out what happens when you mess with them.

Mira: Exactly. The main idea here is that when you perturb an integrable system, there isn't just one way it breaks its special properties; there are actually four distinct kinds of deformations you can end up with depending on how deep you look into the math.

Lev: I'm interested in what those types actually mean for the hardware we build. Can we translate this kind of mathematical classification into something that would actually show up when you try to simulate this on a real quantum computer?

Kai: Right, Lev, so it’s about finding these different deformation types by expanding in a parameter called epsilon. The authors identify four scenarios: simple breaking or preserving integrability, models where the perturbation is exactly integrable, and then two more subtle cases involving higher orders of that deformation parameter.

Mira: That third case is interesting because it means you can only keep the system integrable if you account for every single term in that deformation series. Then there's the fourth case, which is when it’s integrable up to a certain order but then breaks down beyond that point.

Lev: If we're thinking about error correction, does this mean some models are fundamentally harder to fix than others? For instance, if the fourth case means integrability stops after a few terms, that implies the chaos starts much sooner than expected.

Kai: Well, looking at the spectral statistics—how the energy levels spread out—the paper shows that chaos starts off in these different cases in very different ways. The onset of chaos happens differently depending on which type of deformation you have.

Mira: And for these models that are only perturbatively integrable, like a quasi-integrable model, the strength of the deformation where chaos appears scales in a way that sits between what we call strong and weak integrability breaking models.

Lev: That intermediate scaling is something I can get behind. If it’s not purely weak or purely strong, it suggests there's some kind of crossover happening in how quickly things go chaotic.

Paper summary: Kai: They use a boost operator formalism to sort these deformations out, imposing the condition that the commutator of two specific operators is zero at every order of epsilon. This gives them a direct way to study them without relying on other methods.

Mira: And what they find through that formalism is that they can distinguish these four types clearly. They find that for instance, models integrable only up to first order in epsilon are integrable, but one order higher breaks it.

Lev: That means if we try to engineer a system close to integrability, we have a very precise way of knowing exactly where the integrability boundary is located in terms of the perturbation strength.

Kai: The paper then moves on to looking at how the energy levels behave numerically. For an integrable model, you expect level spacings to follow a Poisson distribution with that s power law.

Mira: But for a chaotic system, you want Wigner's surmise, which is that one/two pi s to the power of negative half pi over four times s squared <ref:2603.17018#pg1>. Comparing those two distributions shows that chaos starts at larger epsilon values for the quasi-integrable model HQInt compared to the HdXYZ model.

Lev: So, numerically, it takes a bigger nudge to kick a quasi-integrable system into chaos than a strongly breaking one does in this specific comparison. That’s something we need to keep in mind when designing our experiments.

Kai: Then they look at how that critical coupling, epsilon c, scales with the size of the system, which we call L. For gapless systems, they find it scales as L to the negative power of three.

Mira: That L to the negative power of three scaling is what you’d typically expect for a generic gapless system when you look at critical coupling near integrability breaking. But then they find that the quasi-integrable model HQInt has an intermediate exponent between that three and two, which is what we call the weakly deformed XXZ model.

Lev: An intermediate scaling exponent suggests something about the nature of the delocalization in Fock space, maybe a partially delocalized perturbation causing chaos in a transitional way. That’s a specific physical picture they are proposing.

Paper summary: Kai: They also computed eigenvector entropy measures to quantify how spread out the energy eigenstates are in this transition region. For integrability breaking deformations, that entropy becomes smooth as you move along the energy, which is what you'd expect in a thermalized system.

Mira: But for these integrable or quasi-integrable deformations, they don't see that smoothing of entropy; it shows little sign of decreasing and doesn't become a smooth function with the energy. That implies they aren't fully thermalizing in the same way.

Lev: If it isn't thermalizing smoothly, that has big implications for how we model relaxation dynamics on real hardware, where those long time scales matter a lot.

Kai: The paper also gives us some concrete measures of the deformation strength using operator norms. They find that the Frobenius norms for these deformations scale linearly with L. For instance, for HdXYZ, it's two point twenty-eight plus zero point twenty-one times L.

Mira: And for HQInt, it’s a bit larger, scaling as five point eight zero plus zero point forty eight times L. This shows how the size of the perturbation grows with the system size in these specific models.

Lev: That linear scaling of the deformation strength is good to know because it tells us that even for large systems, we can quantify exactly how much we are deviating from the perfect integrable limit.

Kai: So, to wrap up this paper "On deforming and breaking integrability," it seems they’ve mapped out a very detailed landscape of how integrability breaks. The main thing is that the quasi-integrable model HQInt acts as an intermediate case between strong and weak breaking in its scaling behavior.

Mira: And the authors suggest that this quasi-integrable model might be a partially delocalized perturbation in Fock space, leading to chaos onset that sits somewhere between a transition and a crossover.

Lev: It gives us some very specific scaling laws for the critical coupling, epsilon c, depending on whether you have strong or weak breaking behavior. That’s useful input for designing our error correction codes to predict where we might hit the chaotic regime.

Kai: This paper shows that understanding these different deformation scenarios isn't just a mathematical exercise; it directly affects how we predict when and how quantum systems will exhibit chaos in practice.

Conclusion: Kai: So we've looked at how these small changes to integrable models can actually break their special mathematical rules, and this paper, "On deforming and breaking integrability," looks at those possibilities pretty closely.

Mira: Yeah, it’s about taking systems that are perfectly predictable and seeing what happens when you introduce a little bit of noise or deformation into them. It identifies four different ways that the system can change its behavior under these perturbations.

Lev: From an engineering standpoint, knowing those four types is crucial because it tells us exactly what kind of instability we're dealing with when we try to build something close to that perfect integrable state.

Kai: The authors use this boost operator formalism to systematically sort those deformations, which helps them classify the models into these different categories. They also look at how the system’s energy levels spread out in numerical simulations to see if chaos starts in a predictable way.

Mira: And what they found is that for these quasi-integrable systems, the chaos doesn't start as suddenly as you might expect from a generic breaking model; it shows an intermediate behavior in its scaling.

Lev: That intermediate scaling between strong and weak breaking is interesting because it suggests the perturbation isn't just adding random noise; it’s doing something more complex, like causing some kind of partial delocalization in the system's state space.

Kai: Exactly, and they suggest that this quasi-integrable model might be a partially delocalized perturbation in Fock space, which generates a chaos onset that sits somewhere between a transition and a crossover.

Mira: That means if you’re trying to model these systems for real hardware, you can’t just assume one scaling law applies everywhere; the nature of the deformation dictates the physics.

Lev: If it is partially delocalized, that has huge implications for how we think about thermalization time scales; it suggests that models like HQInt might relax slower than generic breaking models on time scales proportional to epsilon to some power greater than two.

Kai: So, this paper really gives us a framework for understanding *why* chaos appears when you perturb these integrable systems and where that boundary lies in terms of the deformation strength.

Mira: It sets up a clear roadmap for researchers trying to map out the transition region between order and chaos in quantum matter.

More episodes

← Home