Non-local approximations for the Schr"odinger operator with a general point interaction via the non-relativistic limit
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematical Physics. The day in mathematical physics: rigorous models of physical systems.
Kai: We're looking at the paper: "Non-local approximations for the Schr"odinger operator with a general point interaction via the non-relativistic limit".
Mira: Here's the gist of the paper. The paper investigates non-local approximations for point interactions in one-dimensional relativistic quantum mechanics by analyzing the convergence of these approximations under a combined limit process involving both a spatial regularization parameter and the speed of light, revealing a uniform convergence to the Schrödinger operator with a general point interaction Relativistic and Non-Relativistic Limits The study connects the one-dimensional Dirac operator with a point interaction, denoted as DA,c, to the Schrödinger operator HA through the non-relativistic limit as c tends to infinity.
Kai: The paper establishes that for sufficiently large c, DA,c - c 2/2 converges in the norm resolvent sense to HA with the general point interaction.
Mira: Furthermore, it proves that the non-relativistic limit of the non-local approximations Dε A,c also converges to a Schrödinger operator with a rank-two non-local potential Hε A.
Kai: The main result demonstrates that the diagram relating these limits commutes, meaning the double limit exists and results in HA in the norm resolvent sense as ε → 0+.
Title and authors: Kai: So we're looking at this paper, "Non-local approximations for the Schrödinger operator with a general point interaction via the non-relativistic limit." It’s about taking something relativistic, like a Dirac operator with a point interaction, and seeing how it behaves when you take the speed of light to be huge.
Mira: Exactly. The title tells us this whole thing hinges on using that non-relativistic limit as a tool to understand the physics of these point interactions in one dimension. It’s about bridging the gap between relativistic quantum mechanics and something more familiar, like Schrödinger theory.
Lev: From my side, I'm curious how this actually translates into something you can build with real hardware. When we talk about these approximations, what does that mean for the error budget when we try to implement this on a system?
Kai: Well, the paper tackles a specific problem: approximating the Dirac operator with a point interaction using some non-local approximations, D epsilon A c. The key is seeing how these approximate operators converge to the standard Schrödinger operator with that general point interaction as you let both the regularization parameter epsilon go to zero and the speed of light c go to infinity.
Mira: That convergence is what they spend a lot of time proving, and it’s not just some vague idea; they are giving explicit rates for how fast things get close. They show that this double limit exists, meaning you can switch the order in which you take those two limits, and it still lands on the Schrödinger operator with the general point interaction.
Lev: Explicit rates are good, but I need to know what these numbers mean for computational cost when we try to actually run these things. Are we talking about polynomial scaling or something exponential?
Kai: They give us some very concrete rates. For a fixed spatial position z that’s outside the region where the Schrödinger operator has no states, the difference between their approximations falls off at a rate of O epsilon over two, uniformly with respect to c in a neighborhood of infinity. That means as you make your regularization parameter smaller, you get this kind of convergence speed.
Title and authors: Mira: And they also have another convergence rate tied to the speed of light for every fixed epsilon. The non-relativistic limit operator D epsilon A c minus half the square of the speed of light converges to a different operator H epsilon A at a rate of O one over c.
Lev: So you’ve got two different ways things are converging depending on which parameter you fix first, and they combine to give this overall uniform convergence rate of O epsilon over two plus one over c. That sounds like the overhead we'd have if we tried to implement these approximations in a real time simulation.
Kai: Right, and that’s what they prove by showing the diagram relating all these things commutes, which means the double limit exists and yields the Schrödinger operator with a general point interaction in terms of its resolvent. That’s pretty strong structural information.
Mira: And they also point out something interesting about renormalization. In the relativistic setting, approximating Dirac operators often requires adjusting coupling constants, but here, for these natural non-local approximations for point interactions, they converge without needing any renormalization at all.
Lev: That contrast is important for me. If we are building error-correction codes or anything that relies on precise couplings, knowing when you *don't* need that adjustment saves a lot of complexity and potential sources of error in the design phase.
Kai: They detail how the coupling matrix B epsilon differs from A, and they give an explicit formula for this potential renormalization in terms of H epsilon A. It shows exactly what the correction looks like when you move from one approximation scheme to another.
Mira: And they identify the limit operators very precisely as well. For a fixed epsilon, they show that D epsilon A c minus half the square of the speed of light converges to a rank two perturbation of the free Schrödinger operator H zero.
Lev: A rank two perturbation is specific, and knowing it’s a perturbation of H zero helps us understand the structure we’re aiming for. The paper also identifies what happens to these resolvents depending on some condition involving the matrix elements, specifically when one minus a two two n epsilon equals zero or not.
Title and authors: Kai: That condition determines whether the limiting object L epsilon z is an actual operator or just some kind of closed linear relation, which depends on whether that quantity equals zero. This distinction is crucial for understanding how we describe the resulting physics in the non-relativistic limit.
Mira: It’s a subtle point, but it tells us about the nature of the limiting object itself, not just its convergence speed. The work shows that this structure is sensitive to those specific matrix elements involving a two two n epsilon.
Lev: If we think about running this on real hardware, and we hit that condition where one minus a two two n epsilon equals zero, it suggests the math might break down or require a different approach than just standard operator theory.
Kai: Exactly. And they confirm the final big result: that the diagram relating D epsilon A c, DA c, and H epsilon A commutes in terms of resolvents as both epsilon goes to zero and c goes to infinity. This is a stronger statement than just saying the iterated limits commute; it means this convergence is robust across both parameters simultaneously.
Mira: And under that moment condition they mentioned earlier, for fixed z outside the range of states for H A, the Dirac operator with its scaling converges to HA at a rate of O one over c.
Lev: So we have these precise rates and structural guarantees about how these approximations behave as they approach the true limit. It gives us a solid mathematical foundation to even consider using them for modeling complex systems, provided we handle the finite size effects correctly.
Kai: Before we wrap up, I want to touch on what this means for practical quantum hardware development. If we are trying to model interactions where relativistic effects matter but our hardware is non-relativistic, this framework gives us a roadmap for which approximations are reliable and where they start breaking down based on those epsilon and c parameters.
Mira: And I’m interested in the distinction they make between renormalization in the relativistic setting versus its absence here. It suggests that for certain natural point interactions, we can bypass the usual headache of coupling constant adjustments that plagues other approximations.
Lev: That’s a big deal because those adjustments often introduce new errors and complexity into our error-correction schemes. If this framework avoids that renormalization, it simplifies the modeling significantly for us.
Title and authors: Kai: So to summarize this paper, "Non-local approximations for the Schrödinger operator with a general point interaction via the non-relativistic limit," they prove that these specific non-local approximations converge uniformly to the standard Schrödinger operator with a general point interaction as you take both limits together, giving us concrete convergence rates and structural information about the resulting limiting operators.
Mira: It’s important to remember that this convergence relies on specific assumptions, like the moment condition from part (i) mentioned in their work. That condition is what makes the rate estimates hold up uniformly.
Lev: And I'd add that the paper also points out limitations, specifically how it handles cases where the coupling matrix B epsilon converges to something else when a two two is zero, which leads to different outcomes for the interaction strength depending on those specific matrix elements.
Kai: So, this work provides an explicit renormalization scheme arising naturally from the non-relativistic limit—a contrast to other schemes that might require more manual adjustment. It shows that for certain coupling matrices, simpler approximations are enough without needing that renormalization we usually have to worry about.
Mira: Ultimately, the paper establishes a uniform convergence of these non-local approximations to the Schrödinger operator with a general point interaction, providing a rigorous way to connect these two different physical descriptions.
Lev: It’s useful for theoretical modeling because it gives us explicit error bounds in terms of epsilon and c. We need to keep an eye on those rates when designing simulations that try to emulate this physics.
Kai: Yeah, so the title "Non-local approximations for the Schrödinger operator with a general point interaction via the non-relativistic limit" describes a sophisticated way to use limits to tame complicated relativistic problems into something we can analyze more directly in quantum mechanics.
Mira: It’s a solid piece of work that gives us tools to understand how these two different physical pictures fit together when you zoom out from high speed and small interaction scales.
Lev: That's it for this paper. It lays out the convergence properties and the structural identification of the limit operators, which is exactly what we need to start thinking about implementing this in a way that respects those scaling behaviors.
The paper's summary: Kai: So, this paper takes something that's relativistic—a Dirac operator with a point interaction—and it uses the non-relativistic limit, where you take the speed of light to be incredibly large, to see how it connects smoothly to Schrödinger theory.
Mira: Exactly. The core idea is they use some non-local approximations for this Dirac operator and then they prove that both those approximations and the relativistic setup converge together when you let the regularization parameter go down and the speed of light go up.
Lev: What does that convergence look like in practice? Do we get a clean result, or is it messy?
Kai: It’s actually pretty clean. They show a uniform convergence to the standard Schrödinger operator with that general point interaction. This means no matter how you approach it from either direction—relativistic or non-relativistic—the final result is the same.
Mira: And they give us concrete numbers for how fast that happens, which is what I really like about this work. They prove a double limit exists, meaning you can switch the order of operations on those two limits and still get the Schrödinger operator in terms of its resolvent.
Lev: Concrete rates are good, but what do those rates tell us about the computational cost for running simulations?
Kai: They give us specific scaling behaviors. For instance, if you fix a spatial position that’s outside the region where the Schrödinger operator has states, the difference between their approximations falls off with a rate of one over two times epsilon, uniformly with respect to c in a neighborhood of infinity. That means as your regularization parameter gets smaller, you get this kind of convergence speed.
Mira: And they also have another rate tied directly to the speed of light for every fixed epsilon. The non-relativistic limit operator minus half the square of the speed of light converges to a different operator at a rate of one over c for any fixed epsilon.
Lev: So you’ve got two ways things are converging depending on which parameter you fix first, and they combine to give this overall uniform convergence rate of one over two times epsilon plus one over c. That sounds like the overhead we'd have if we tried to implement these approximations in a real time simulation.
Kai: Right, and that’s what they prove by showing the diagram relating all these things commutes, which means the double limit exists and yields the Schrödinger operator with a general point interaction in terms of its resolvent. That’s pretty strong structural information for how this physics connects.
Mira: And they also point out something interesting about renormalization. In the relativistic setting, approximating Dirac operators often requires adjusting coupling constants, but here, for these natural non-local approximations for point interactions, they converge without needing any renormalization at all.
The paper's summary: Lev: That contrast is important for me. If we are building error-correction codes or anything that relies on precise couplings, knowing when you *don't* need that adjustment saves a lot of complexity and potential sources of error in the design phase.
Kai: They detail how the coupling matrix B epsilon differs from A, and they give an explicit formula for this potential renormalization in terms of H epsilon A. It shows exactly what the correction looks like when you move from one approximation scheme to another.
Mira: And they identify the limit operators very precisely as well. For a fixed epsilon, they show that D epsilon A c minus half the square of the speed of light converges to a rank two perturbation of the free Schrödinger operator H zero.
Lev: A rank two perturbation is specific, and knowing it’s a perturbation of H zero helps us understand the structure we’re aiming for. The paper also identifies what happens to these resolvents depending on some condition involving the matrix elements, specifically when one minus a two two n epsilon equals zero or not.
Kai: That condition determines whether the limiting object L epsilon z is an actual operator or just some kind of closed linear relation, which depends on whether that quantity equals zero. This distinction is crucial for understanding how we describe the resulting physics in the non-relativistic limit.
Mira: It’s a subtle point, but it tells us about the nature of the limiting object itself, not just its convergence speed. The work shows that this structure is sensitive to those specific matrix elements involving a two two n epsilon.
Lev: If we think about running this on real hardware, and we hit that condition where one minus a two two n epsilon equals zero, it suggests the math might break down or require a different approach than just standard operator theory.
Kai: Exactly. And they confirm the final big result: that the diagram relating D epsilon A c, DA c, and H epsilon A commutes in terms of resolvents as both epsilon goes to zero and c goes to infinity. This is a stronger statement than just saying the iterated limits commute; it means this convergence is robust across both parameters simultaneously.
Mira: And under that moment condition they mentioned earlier, for fixed z outside the range of states for H A, the Dirac operator with its scaling converges to HA at a rate of one over c.
Lev: So we have these precise rates and structural guarantees about how these approximations behave as they approach the true limit. It gives us a solid mathematical foundation to even consider using them for modeling complex systems, provided we handle the finite size effects correctly.
The paper's summary: Kai: Before we wrap up, I want to touch on what this means for practical quantum hardware development. If we are trying to model interactions where relativistic effects matter but our hardware is non-relativistic, this framework gives us a roadmap for which approximations are reliable and where they start breaking down based on those epsilon and c parameters.
Mira: And I’m interested in the distinction they make between renormalization in the relativistic setting versus its absence here. It suggests that for certain natural point interactions, we can bypass the usual headache of coupling constant adjustments that plagues other approximations.
Lev: That’s a big deal because those adjustments often introduce new errors and complexity into our error-correction schemes. If this framework avoids that renormalization, it simplifies the modeling significantly for us.
Kai: So to summarize this paper, "Non-local approximations for the Schrödinger operator with a general point interaction via the non-relativistic limit," they prove that these specific non-local approximations converge uniformly to the standard Schrödinger operator with a general point interaction as you take both limits together, giving us concrete convergence rates and structural information about the resulting limiting operators.
Mira: It’s important to remember that this convergence relies on specific assumptions, like the moment condition from part i mentioned in their work. That condition is what makes the rate estimates hold up uniformly.
Lev: And I'd add that the paper also points out limitations, specifically how it handles cases where the coupling matrix B epsilon converges to something else when a two two is zero, which leads to different outcomes for the interaction strength depending on those specific matrix elements.
Kai: So, this work provides an explicit renormalization scheme arising naturally from the non-relativistic limit—a contrast to other schemes that might require more manual adjustment. It shows that for certain coupling matrices, simpler approximations are enough without needing that renormalization we usually have to worry about.
Mira: Ultimately, the paper establishes a uniform convergence of these non-local approximations to the Schrödinger operator with a general point interaction, providing a rigorous way to connect these two different physical descriptions.
Lev: It’s useful for theoretical modeling because it gives us explicit error bounds in terms of epsilon and c. We need to keep an eye on those rates when designing simulations that try to emulate this physics.
Kai: Yeah, so the title "Non-local approximations for the Schrödinger operator with a general point interaction via the non-relativistic limit" describes a sophisticated way to use limits to tame complicated relativistic problems into something we can analyze more directly in quantum mechanics.
Mira: It’s a solid piece of work that gives us tools to understand how these two different physical pictures fit together when you zoom out from high speed and small interaction scales.
Lev: That's it for this paper. It lays out the convergence properties and the structural identification of the limit operators, which is exactly what we need to start thinking about implementing this in a way that respects those scaling behaviors.
The paper's improvements: Tom: So, this part of the discussion is about what these authors suggest they should do next to make their work even better and more useful for other people.
Kai: They're essentially pointing out how they can get more detail on those convergence rates, specifically by tracking terms like theta epsilon and eta epsilon, so you can predict whether a convergence is one over two times epsilon or one over two times epsilon plus one over c in specific regimes.
Mira: That’s useful for me because it gives us a way to narrow down exactly when the O one over two times epsilon term dominates versus when the speed of light term takes precedence.
Lev: If we can predict that, I can start thinking about how to design our error-correction overhead. A clearer convergence prediction means we know where the computational cost really lands in terms of scaling.
Kai: Also, they suggest getting better at determining the renormalization behavior by analyzing that coupling matrix B epsilon as it approaches zero, specifically identifying when B epsilon converges to something like the determinant A a two two E one one versus when it results in that delta interaction star we talked about earlier.
Mira: That would be really helpful for me because understanding exactly how the coupling constants adjust is key to knowing if we need to add extra complexity or if we can keep things simple.
Lev: If they can pinpoint the exact moment where B epsilon starts behaving differently, it cuts down on guesswork when setting up our physical parameters.
Kai: They also propose using Lemma five point ten to identify the kernel of that limiting operator L epsilon z, which is explicitly given by equation five point three, allowing for direct computation of its structure based on v epsilon and f epsilon functions.
Mira: Being able to directly compute the structure of the kernel based on those functions is a big step because it moves us from just knowing *what* the limit is to knowing *exactly how* it looks mathematically.
Lev: That lets me know immediately if we're dealing with a standard operator or some kind of closed linear relation, which depends on whether one minus a two two n epsilon equals zero or not.
Kai: Right, and they’re pushing for stronger proofs on the commutation between the limits of those non-local approximations and the non-relativistic limit by leveraging theorem six point one to show that that whole diagram commutes uniformly in c and epsilon.
Mira: That uniform convergence rate estimate of one over theta epsilon plus eta epsilon plus one over c for the difference of resolvents is a strong statement, and it's what I need to see for the physics to be truly robust.
Lev: A stronger statement means we have better bounds on the error when we try to run this on real hardware, which is exactly where my focus is right now.
Kai: So in short, they are suggesting more detailed analysis of those convergence rates and structural elements so that other researchers can build on this foundation with more precise predictions.
Mira: It’s about tightening the assumptions underneath the results to see exactly what makes the convergence hold up under those specific conditions they mentioned earlier.
Lev: This work moves us from just having a general idea of how these approximations work to having a much more rigorous, detailed roadmap for implementation error budgeting.
Conclusion: Kai: So to wrap up, this paper, "Non-local approximations for the Schrödinger operator with a general point interaction via the non-relativistic limit," shows that these specific non-local methods converge uniformly to the standard Schrödinger operator with a general point interaction when you combine those two limits.
Mira: It’s important to remember that this convergence relies on those specific assumptions, like that moment condition from part i, which is what makes the rate estimates hold up uniformly across the whole range.
Lev: I just want to say that if we can get these explicit error bounds in terms of epsilon and c, it gives us a solid mathematical foundation for even trying to model this physics on real hardware.
Kai: And they also flagged limitations, particularly how the coupling matrix B epsilon behaves when a two two is zero, which leads to different outcomes for the interaction strength depending on those specific matrix elements.
Mira: That means we have to be careful about which approximations we choose if our physical situation forces that specific condition in the first place.
Lev: It’s useful for theoretical modeling because it gives us explicit error bounds in terms of epsilon and c. We need to keep an eye on those rates when designing simulations that try to emulate this physics.
Kai: Yeah, so the title "Non-local approximations for the Schrödinger operator with a general point interaction via the non-relativistic limit" describes a sophisticated way to use limits to tame complicated relativistic problems into something we can analyze more directly in quantum mechanics.
Mira: It’s a solid piece of work that gives us tools to understand how these two different physical pictures fit together when you zoom out from high speed and small interaction scales.
Lev: That's it for this paper. It lays out the convergence properties and the structural identification of the limit operators, which is exactly what we need to start thinking about implementing this in a way that respects those scaling behaviors.
Luka Heriban
Czech Technical University in Prague
math-ph, math.MP, math.SP
Submitted: 2026-11-03
Updated: 2026-11-03
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
Importance score: 76/100
The gist: The paper investigates non-local approximations for point interactions in one-dimensional relativistic quantum mechanics by analyzing the convergence of these approximations under a combined limit…
Key concepts
- Non-local Approximations
- These are mathematical methods used to approximate complex operators in quantum mechanics using non-local potentials. The paper specifically examines how these approximations behave when transitioning from a relativistic description (like the Dirac operator) to a simpler, non-relativistic Schrödinger description.
- Non-Relativistic Limit
- This is the process of taking the speed of light ($c$) to infinity in quantum mechanics. In this limit, relativistic equations simplify into standard non-relativistic Schrödinger equations. The paper analyzes how specific non-local approximations behave during this transition.
- Uniform Convergence
- Uniform convergence means that the error between an approximation and the true limiting operator shrinks at a predictable rate across all relevant parameters (like the regularization parameter $\epsilon$ and $c$). The paper establishes a combined convergence rate of $O(\epsilon_1/2 + c^{-1})$ for these limits.
- Renormalization
- Renormalization is the process of adjusting coupling constants in theoretical models to obtain physically meaningful results. The paper distinguishes between renormalization needed for relativistic approximations and its absence in certain non-relativistic point interaction schemes.
Terminology
Summary
The paper investigates non-local approximations for point interactions in one-dimensional relativistic quantum mechanics by analyzing the convergence of these approximations under a combined limit process involving both a spatial regularization parameter and the speed of light, revealing a uniform convergence to the Schrödinger operator with a general point interaction
Relativistic and Non-Relativistic Limits
The study connects the one-dimensional Dirac operator with a point interaction, denoted as DA,c, to the Schrödinger operator HA through the non-relativistic limit as c tends to infinity. The paper establishes that for sufficiently large c, DA,c - c 2/2 converges in the norm resolvent sense to HA with the general point interaction. p. 0 We consider the one-dimensional Dirac operator DA,c with a general, not necessarily self-adjoint, point interaction at the origin, whose strength is described by an arbitrary
complex 2×2 matrix A and suitably rescaled by the speed of light c, together with its non-local
approximations D
ε
A,c obtained by replacing the Dirac distribution by a rank-one projection onto
a scaled function vε. It is known that D
ε
A,c → DA,c as ε → 0+ and that DA,c − c
2
/2 converges,
as c → +∞, to the Schr¨odinger operator HA with the general point interaction, both in the
norm resolvent sense. ) For every sufficiently small fixed ε > 0, Dε
A,c − c
2/2 converges
in the norm resolvent sense, as c → +∞, to the Schr¨odinger operator with a rank-two
non-local potential
(1. 3) Hε
A = H0 +
X
2
j,k=1
(Bε)jk|u
ε
j
⟩⟨u
ε
k
|, uε
1 = vε, uε
2 = v
′
ε
, ) The diagram (1. 2) completed by Hε
A
commutes, the double limit of
(Dε
A,c−c
2/2−z)
−1 as (ε, c) → (0+, +∞) exists, and, in particular, Hε
A → HA in the norm
resolvent sense as ε → 0+.Where the paper says it
Where the paper says it
Where the paper says it
Convergence Rates and Uniformity
The authors provide explicit rates of convergence for both limits. Specifically, Theorem 4. p. 1 This allows us to prove our first main result. (i) (Theorem 4. (ii) (Theorems 5. ) For every sufficiently small fixed ε > 0, Dε
A,c − c
2/2 converges
in the norm resolvent sense, as c → +∞, to the Schr¨odinger operator with a rank-two
non-local potential
(1. If, in addition, R
R
|s|
1/2
|v(s)| ds < ∞, the rate of this convergence is O(ε
1/2
). Next, we compute the non-relativistic limit of Dε
A,c for every fixed ε > 0. We prove that the first convergence is uniform with respect to c in a
neighbourhood of +∞, and we provide explicit rates of convergence. Combining this with an
explicit computation of the non-relativistic limit of D
ε
A,c for fixed ε, we show that the two limits can be interchanged and that the corresponding double limit exists.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Renormalization and Operator Identification
A key finding is the distinction between renormalization in the relativistic setting and its absence in the non-relativistic one. While approximations by local, scaled potentials for Dirac operators require a renormalization of coupling constants, natural non-local approximations for point interactions (like those studied here) converge to DA without any renormalization. p. 1 Another natural question is how point interactions can be approximated by more regular potentials. For the Dirac operator, it is well known that approximations by local, scaled potentials
ε
−1h(x/ε) converge in the norm resolvent sense, but the coupling constant of the limit point
interaction differs in general from the one suggested by the formal limit; this effect is called the
renormalization of coupling constants [33, 18, 19, 35]. As a by-product we
obtain a family of non-local, rank-two approximations Hε
A of HA that converge to HA in the
norm resolvent sense. Contrary to the relativistic setting, the coupling constants of Hε
A are
renormalized; we give an explicit formula for this renormalization.Where the paper says it
Where the paper says it
Identification of Limit Operators
The paper identifies the limit operators precisely. For a fixed ε, Theorem 5. p. 0 We prove that the first convergence is uniform with respect to c in a
neighbourhood of +∞, and we provide explicit rates of convergence. Combining this with an
explicit computation of the non-relativistic limit of D
ε
A,c for fixed ε, we show that the two limits can be interchanged and that the corresponding double limit exists. (ii) (Theorems 5. ) For every sufficiently small fixed ε > 0, Dε
A,c − c
2/2 converges
in the norm resolvent sense, as c → +∞, to the Schr¨odinger operator with a rank-two
non-local potential
(1. We consider the one-dimensional Dirac operator DA,c with a general, not necessarily self-adjoint, point interaction at the origin, whose strength is described by an arbitrary
complex 2×2 matrix A and suitably rescaled by the speed of light c, together with its non-local
approximations D
ε
A,c obtained by replacing the Dirac distribution by a rank-one projection onto
a scaled function vε. It is known that D
ε
A,c → DA,c as ε → 0+ and that DA,c − c
2
/2 converges,
as c → +∞, to the Schr¨odinger operator HA with the general point interaction, both in the
norm resolvent sense. The rate of this convergence
is O(c
−1
). Using the theory of pseudoresolvents, we show moreover that the non-relativistic limit of Dε
A,c
is the resolvent of a closed operator also if v /∈ H1
(R), while for v ∈ H1
(R) and an exceptional
value of ε it is merely the resolvent of a linear relation (Propositions 5. NON-LOCAL APPROXIMATIONS FOR THE SCHRODINGER OPERATOR ¨
WITH A GENERAL POINT INTERACTION VIA THE
NON-RELATIVISTIC LIMIT
LUKA´S HERIBAN ˇ
Abstract. We consider the one-dimensional Dirac operator DA,c with a general, not necessarily self-adjoint, point interaction at the origin, whose strength is described by an arbitrary
complex 2×2 matrix A and suitably rescaled by the speed of light c, together with its non-local
approximations D
ε
A,c obtained by replacing the Dirac distribution by a rank-one projection onto
a scaled function vε.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Commutativity and Final Results
The final result confirms that the diagram relating Dε A,c, DA,c, and Hε A commutes in the norm resolvent sense as ε → 0+ and c → +∞. This commutativity is a stronger statement than the commutativity of iterated limits. p. 6 Consequently, [16, Prop. 6] yield the following statement, see also [16, Sect. The following lemma is the key observation of the paper. It shows that the resolvents of DA,c
and Dε
A,c at wc can be written in terms of the quantities (3.Where the paper says it
Where the paper says it
Renormalization of Coupling Constants
The analysis of the coupling constants reveals a specific renormalization pattern. For the naive approximations H0 + P j,k b jku εj⟩⟨u εk, if b22! p. 2 3 we describe the norm resolvent limits of the
“naive” approximations H0 +
P
j,k bjk|u
ε
j
⟩⟨u
ε
k
| with an ε-independent matrix B. For instance, if
b22 ̸= 0, then they converge to the δ-interaction of strength det B/b22; in particular, the naive
approximations of the δ
′
-interaction converge to the free Hamiltonian, which is in accordance
with [32]. 3 we describe the norm resolvent limits of the
“naive” approximations H0 +
P
j,k bjk|u
ε
j
⟩⟨u
ε
k
| with an ε-independent matrix B. For instance, if
b22 ̸= 0, then they converge to the δ-interaction of strength det B/b22; in particular, the naive
approximations of the δ
′
-interaction converge to the free Hamiltonian, which is in accordance
with [32]. As a by-product we
obtain a family of non-local, rank-two approximations Hε
A of HA that converge to HA in the
norm resolvent sense. Contrary to the relativistic setting, the coupling constants of Hε
A are
renormalized; we give an explicit formula for this renormalization. 2), the coupling matrix Bε in (1. 3) does differ from A; in fact, Bε does not converge to A
unless a22 = 0 and a12a21 = 0.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Conclusion on Approximation Schemes
The paper concludes by contrasting the explicit non-local approximations derived here with other schemes. It notes that while this approach provides an explicit renormalization, another scheme based on rank-one operators with disjoint supports avoids renormalization. p. 1 Another natural question is how point interactions can be approximated by more regular potentials. For the Dirac operator, it is well known that approximations by local, scaled potentials
ε
−1h(x/ε) converge in the norm resolvent sense, but the coupling constant of the limit point
interaction differs in general from the one suggested by the formal limit; this effect is called the
renormalization of coupling constants [33, 18, 19, 35]. Non-local approximations. Throughout the paper, the function v satisfies the following
assumption. Section 4 contains the
proof of the uniform convergence. In Section 5 we compute the non-relativistic limit of the
non-local approximations, discuss it from the point of view of pseudoresolvents, and identify the
limit operator. Section 4 contains the
proof of the uniform convergence. In Section 5 we compute the non-relativistic limit of the
non-local approximations, discuss it from the point of view of pseudoresolvents, and identify the
limit operator.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Along the upper branch of the diagram in Theorem 6.1, no renormalization of coupling constants occurs. p. 2 Interestingly, while no renormalization occurs along the upper branch of
(1. 2), the coupling matrix Bε in (1. NON-LOCAL APPROXIMATIONS FOR THE SCHRODINGER OPERATOR ¨
WITH A GENERAL POINT INTERACTION VIA THE
NON-RELATIVISTIC LIMIT
LUKA´S HERIBAN ˇ
Abstract. We consider the one-dimensional Dirac operator DA,c with a general, not necessarily self-adjoint, point interaction at the origin, whose strength is described by an arbitrary
complex 2×2 matrix A and suitably rescaled by the speed of light c, together with its non-local
approximations D
ε
A,c obtained by replacing the Dirac distribution by a rank-one projection onto
a scaled function vε. NON-LOCAL APPROXIMATIONS FOR THE SCHRODINGER OPERATOR ¨
WITH A GENERAL POINT INTERACTION VIA THE
NON-RELATIVISTIC LIMIT
LUKA´S HERIBAN ˇ
Abstract. We consider the one-dimensional Dirac operator DA,c with a general, not necessarily self-adjoint, point interaction at the origin, whose strength is described by an arbitrary
complex 2×2 matrix A and suitably rescaled by the speed of light c, together with its non-local
approximations D
ε
A,c obtained by replacing the Dirac distribution by a rank-one projection onto
a scaled function vε. 2), the coupling matrix Bε in (1. 3) does differ from A; in fact, Bε does not converge to A
unless a22 = 0 and a12a21 = 0. Alternatively, they can be introduced
starting from the formal expression
(1. 1) HA = −
d
2
dx
2
+ a11|δ⟩⟨δ| + a12|δ⟩⟨δ
′
| + a21|δ
′
⟩⟨δ| + a22|δ
′
⟩⟨δ
′
|, A = (aij )
2
i,j=1 ∈ C
2,2
,
and the theory of distributions acting on discontinuous test functions [23, 24, 3].Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Proposition on Naive Approximations
The authors show that the naive approximations of the δ'-interaction converge to the free Hamiltonian, which is in accordance with [32]. This suggests that for certain coupling matrices, simpler approximations suffice without renormalization. p. 1 On the other hand, Seba [ ˇ 33] observed for two
particular self-adjoint couplings, and it was later proved in full generality in [16, Thm. 1], that
the non-local approximations
Dε
A = D0 + A ⊗ |vε⟩⟨vε|, vε(x) = ε
−1
v(ε
−1x),
converge to the point interaction DA without any renormalization. On the other hand, Seba [ ˇ 33] observed for two
particular self-adjoint couplings, and it was later proved in full generality in [16, Thm. 1], that
the non-local approximations
Dε
A = D0 + A ⊗ |vε⟩⟨vε|, vε(x) = ε
−1
v(ε
−1x),
converge to the point interaction DA without any renormalization. See also [17] for similar
result in the higher dimension. For the Schr¨odinger operator, the situation is different: already
for the δ
′
-interaction, natural non-local approximations require a renormalization [32], and an
approximation scheme avoiding the renormalization was proposed in [4], see also [5]. See also [17] for similar
result in the higher dimension. For the Schr¨odinger operator, the situation is different: already
for the δ
′
-interaction, natural non-local approximations require a renormalization [32], and an
approximation scheme avoiding the renormalization was proposed in [4], see also [5]. As a by-product we
obtain a family of non-local, rank-two approximations Hε
A of HA that converge to HA in the
norm resolvent sense. Contrary to the relativistic setting, the coupling constants of Hε
A are
renormalized; we give an explicit formula for this renormalization.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The authors show that the naive approximations of the δ'-interaction converge to the free Hamiltonian, which is in accordance with [34]. This suggests that for certain coupling matrices, simpler approximations suffice without renormalization. The work provides a natural explicit renormalization scheme arising from the non-relativistic limit.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved ability to analyze convergence rates for non-local approximations by explicitly tracking terms like theta epsilon and eta epsilon, allowing prediction of whether a convergence is O epsilon 1/2 or O epsilon 1/2 + c minus 1 in specific regimes.
-
Enhanced capability to determine the renormalization behavior of coupling constants by analyzing the matrix Bepsilon as ε approaches zero, specifically identifying when Bepsilon converges to det A a22 E11 versus when it results in the delta interaction HA star.
-
Improved capacity for performing non-relativistic limit calculations by applying Lemma 5.10 to identify the kernel of the limiting operator L epsilon z, which is explicitly given by equation 5.3, allowing for direct computation of its structure based on v epsilon and F epsilon functions.
-
Stronger performance in proving commutation between limits of non-local approximations and the non-relativistic limit by leveraging Theorem 6.1 to show that the diagram (1.2) commutes uniformly in c and ε, yielding a unified convergence rate estimate O theta epsilon plus eta epsilon plus c minus 1 for the difference of resolvents.
-
Ability to distinguish between cases where the limiting object L epsilon z is an operator versus a closed linear relation by checking if Ker L epsilon z is zero, which depends on whether 1 minus a22 nepsilon equals zero or not, as detailed in Proposition 5.13.
Abstract
We consider the one-dimensional Dirac operator D A,c with a general, not necessarily self-adjoint, point interaction at the origin, whose strength is described by an arbitrary complex 2 times 2 matrix A and suitably rescaled by the speed of light c, together with its non-local approximations D A,c ε obtained by replacing the Dirac distribution by a rank-one projection onto a scaled function v ε. It is known that D A,c ε to D A,c as ε to 0+ and that D A,c-c 2/2 converges, as c to+ infinity, to the Schrödinger operator H A with the general point interaction, both in the norm resolvent sense. We prove that the first convergence is uniform with respect to c in a neighbourhood of + infinity, and we provide explicit rates of convergence. Combining this with an explicit computation of the non-relativistic limit of D A,c ε for fixed ε, we show that the two limits can be interchanged and that the corresponding double limit exists. As a by-product we obtain a family of non-local, rank-two approximations H A ε of H A that converge to H A in the norm resolvent sense. Contrary to the relativistic setting, the coupling constants of H A ε are renormalized; we give an explicit formula for this renormalization.
Related papers
- Existence of a fluid phase for the two-dimensional one-component plasma
- The Huang-Yang conjecture for the low-density Fermi gas
- Anderson localisation in infinite and finite resonator crystals with long-range interactions
- Free energy and spectral edge of the SYK model
- Anderson localization for 1-d quasi-periodic Schr"odinger operators with degenerate weights
- Irregular SLE(4) martingales and isomonodromic deformations