Improved Pion-Kaon Identification in Heavy-Ion Collisions with a Two-Dimensional Transformation

arXiv:2603.03353 · physics.ins-det, hep-ex, nucl-ex · Submitted 2026-02-27 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.

Kai: We're looking at the paper: "Improved Pion-Kaon Identification in Heavy-Ion Collisions with a Two-Dimensional Transformation".

Title and authors: Kai: So we've seen that the title of this paper is "Improved Pion-Kaon Identification in Heavy-Ion Collisions with a Two-Dimensional Transformation," and it points us directly toward the solution they found for those high momentum separation issues we talked about.

Mira: They are essentially proposing a geometric transformation, a shift and rotation, to map the m two and n sigma distributions onto a new plane where pions and kaons stand out more distinctly.

Lev: It's about redefining the coordinate system so that neither variable dominates the separation in this new space, which is smart because it addresses that scaling problem we discussed earlier.

Kai: They show how this transformation works mathematically by defining parameters like a scale factor, a translation parameter for n sigma times pi, and a rotation angle alpha.

Mira: And they use this framework to describe the resulting distributions using a two-dimensional Student’s t function with an extra flexibility parameter called nu for better fitting.

Lev: The optimization part is key here; they tune those transformation parameters so that the effective Gaussian widths of the particle peaks along both axes end up being comparable, and they align the pion and kaon distributions mostly along one axis.

Kai: So what this means for us is that instead of just doing a simple 1D fit to m two or n sigma, we're looking at a whole new shape in 2D space that separates the species better <ref:2603.03353#pg1>.

Mira: It moves us from relying on a single variable to using the joint information between m two and n sigma simultaneously for identification.

Lev: The method is mathematically defined by equations involving those scale factor, translation parameters, and rotation angle so we know exactly what the transformation looks like.

The paper's summary: Kai: To summarize what they did in "Improved Pion-Kaon Identification in Heavy-Ion Collisions with a Two-Dimensional Transformation," they introduced this two-dimensional shift and rotation method to exploit the correlation between m two and n sigma for better identification.

Mira: They showed that this technique effectively enhances particle identification performance by constructing a new variable space where pions and kaons are naturally more separated than before.

Lev: The entire procedure involves three sequential steps: scaling, shifting, and rotation, with the scaling step specifically designed to equalize the effective resolution of both variables in the transformed space.

Kai: They then optimized these transformation parameters so that the resulting distributions show clear separation between pions and kaons predominantly along one axis in this new setup.

Mira: The fitting strategy uses a two-dimensional Student’s t function, which includes an additional degree of freedom called nu to give it more flexibility when fitting the data.

Lev: This multi-particle fit function has thirteen free parameters that they optimize iteratively to find a stable and physically meaningful solution for the parameters describing the transformation.

Kai: The paper validates all this by applying these parameterized curves to the AMPT model output, smearing those m two and n sigma values on a track-by-track basis to simulate realistic, pT-dependent detector response effects.

Mira: This means they are controlling for how the detector resolution changes with transverse momentum while preserving the actual correlations between particles in the collision event.

The paper's improvements: Lev: What this actually improves is that they can accurately extract particle yields up to pT equals three point zero GeV per c with a purity exceeding ninety-eight percent, which is a significant extension beyond what was possible with conventional one-dimensional fitting methods.

Kai: That high fidelity at high pT, up to three point zero GeV per c, is the main win because that's where the overlap between m two and n sigma usually becomes completely intractable for simple methods <ref:2603.03353#pg2>.

Mira: The paper shows that even at these higher momenta, the extracted yields for both pions and kaons still agree very well with what was input from the AMPT model, showing they didn't introduce big biases in the results.

Lev: They confirmed that this method doesn't introduce biases in flow measurements either because their reconstructed v two values for both pions and kaons match the AMPT input within statistical uncertainties across the entire pT range studied.

Kai: So for someone just looking at the data, this means we can trust the identification purity to be above ninety-eight percent even at those highest transverse momenta where things usually get messy.

Mira: The implication is that we gain a reliable way to study how particles behave in the quark-gluon plasma without having to worry as much about those specific momentum overlaps complicating our measurements.

Conclusion: Kai: To wrap up, the paper "Improved Pion-Kaon Identification in Heavy-Ion Collisions with a Two-Dimensional Transformation" shows that by using a two dimensional shift and rotation method, we can tackle the severe overlap between m two and n sigma distributions.

Mira: They achieved this by building a new variable space where pions and kaons are clearly separated, allowing for reliable identification up to pT equals three point zero GeV per c with purity over ninety-eight percent.

Lev: And they confirmed that this separation method is robust, because the extracted v two coefficients for both species match the AMPT input within statistical uncertainties across the whole transverse momentum range studied.

Kai: This capability is crucial for precision flow measurements and differential studies because it gives us access to those high momentum regimes where identification used to be really difficult.

Mira: Overall, this method provides a controlled framework for evaluating particle identification performance by incorporating realistic detector resolution effects into the model in a data-driven way.

Lev: It's important to remember that the paper notes its limitation, which is that it relies on the AMPT model output for validation, and they have to tune those transformation parameters carefully.

Kai: So this work gives us a more powerful tool for studying heavy-ion collisions by extending our reliable identification range significantly.

Shaowei Lan, Bijun Fan, Like Liu

School of Electrical and Mechanical Engineering, Pingdingshan University · Key Laboratory of Quark & Lepton Physics (MOE) and Institute of Particle Physics, Central China Normal University

physics.ins-det, hep-ex, nucl-ex

Submitted: 2026-02-27

Updated: 2026-11-17

Comments: 10 pages, 7 figures

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

Importance score: 86/100

The gist: Accurate identification of charged pions and kaons is essential for precision measurements in relativistic heavy-ion collisions…

Key concepts

Time-of-Flight Mass Square ($m^2$)
This variable relates to the particle's mass and velocity, providing information about its mass. It is one of the key inputs used in the identification process. When combined with energy loss data, it helps distinguish between different particle species, especially at higher momenta where other methods struggle.
Ionization Energy Loss ($n rac{dE}{dx}$)
This variable measures how much energy a charged particle loses as it passes through matter. It is another crucial piece of information used alongside $m^2$. The correlation between this energy loss and the mass-squared allows the proposed two-dimensional transformation to create a more distinct separation between pions and kaons.
Two-Dimensional Shift and Rotation Transformation
This is the core methodology. Instead of using simple one-dimensional fits, this method mathematically transforms the existing $m^2$ and $n rac{dE}{dx}$ data into a new coordinate system. This transformation is optimized to align the distributions of pions and kaons along different axes, making them much easier to separate reliably.
Purity and Transverse Momentum (pT)
Purity refers to how accurately the identification system can tell a particle apart from others; high purity means very few misidentifications. The study shows this method maintains a purity exceeding 98% up to pT = 3 GeV/c, which is essential for precise measurements in the high-momentum regime.

Terminology

Summary

Accurate identification of charged pions and kaons is essential for precision measurements in relativistic heavy-ion collisions, but becomes increasingly challenging at intermediate and high transverse momentum due to the overlap between time-of-flight mass-square (m2) and ionization energy loss (nσ) distributions. p. 0

Where the paper says it

Improved Pion–Kaon Identification in Heavy-Ion Collisions with a Two-Dimensional Transformation Shaowei Lan,1, ∗ Bijun Fan,2, † and Like Liu2, ‡ 1School of Electrical and Mechanical Engineering, Pingdingshan University, 467000 Pingdingshan, China 2Key Laboratory of Quark & Lepton Physics (MOE) and Institute of Particle Physics, Central China Normal University, Wuhan 430079, China (Dated: March 5, 2026) Accurate identification of charged pions and kaons is essential for precision measurements in relativistic heavy-ion collisions, but becomes increasingly challenging at intermediate and high transverse momentum due to the overlap between time-of-flight mass-square (m2 ) and ionization energy loss (nσ) distributions. In this work, we present a two-dimensional shift and rotation method that exploits the correlated information between m2 and nσ to enhance particle identification performance.

The reconstructed pion and kaon transverse momentum distributions show excellent agreement with the AMPT input, maintaining a purity exceeding 98% at high pT and extend the reliable identification range up to pT ≈ 3 GeV/c. p. 0
Where the paper says it

The method is validated using Au+Au collision events generated with the AMPT model, where detector response effects are incorporated through a data-driven smearing procedure tuned to reproduce the particle identification performance of the STAR experiment. The reconstructed pion and kaon transverse momentum distributions show excellent agreement with the AMPT input, maintaining a purity exceeding 98% at high pT and extend the reliable identification range up to pT ≈ 3 GeV/c.

Introduction to the Problem

The challenge in hadron identification lies in the separation of pions and kaons at intermediate to high pT (typically above 2.0 GeV/c) because conventional methods based on one-dimensional fits often suffer from overlap in these momentum regions, leading to reduced purity and increased systematic uncertainty. p. 1

Where the paper says it

A persistent challenge in hadron identification lies in the separation of pions and kaons at intermediate to high pT (typically above 2. Due to their same electric charge and relatively small mass difference [14], the conventional methods, based on one-dimensional fits to dE/dx or m2 distributions [15], often suffer from overlap in these momentum regions, leading to reduced purity and increased systematic uncertainty in physics analyses.

This overlap is particularly problematic because of their same electric charge and relatively small mass difference. p. 1
Where the paper says it

A persistent challenge in hadron identification lies in the separation of pions and kaons at intermediate to high pT (typically above 2. Due to their same electric charge and relatively small mass difference [14], the conventional methods, based on one-dimensional fits to dE/dx or m2 distributions [15], often suffer from overlap in these momentum regions, leading to reduced purity and increased systematic uncertainty in physics analyses.

In the STAR experiment, the conventional procedure involves an initial selection using the nσπ variable, typically requiring nσπ < 3, followed by a one-dimensional m2 distribution fit. p. 2
Where the paper says it

In the conventional particle identification procedure commonly employed in STAR analyses, an initial selection is applied using the nσπ variable, typically requiring |nσπ| <3, to enhance pion purity. Within each pT interval, a one-dimensional m2 distribution is then constructed from the selected tracks, as shown by the black solid circles in Fig.

This one-dimensional framework begins to exhibit pronounced limitations in the intermediate pT region because the kaon peak becomes strongly overlapped with the dominant pion distribution and is no longer clear separable. p. 2
Where the paper says it

While this method performs adequately at low transverse momentum, it begins to exhibit pronounced limitations in the intermediate pT region. In particular, the kaon peak becomes strongly overlapped with the dominant pion distribution and is no longer clear separable.

Methodology of the Transformation

The proposed solution introduces a two-dimensional shift and rotation method that exploits the correlated information between m2 and nσ to enhance particle identification performance. This approach redefines the coordinate system through an optimal transformation, constructing a new variable space where pions and kaons are more clearly separated. p. 1

Where the paper says it

This approach, originally inspired by techniques employed in the STAR collaboration [19–23], utilizing both the nσ (normalized dE/dx) [24] from the TPC and the m2 from the TOF detector. By redefining the coordinate system through an optimal transformation, the method simultaneously exploits the PID information from both detectors to construct a new variable space where pions and kaons are more clearly [20, 23].

The transformation procedure consists of three sequential steps: scaling, shifting, and rotation. p. 2
Where the paper says it

This redefinition of the variable space maximizes the separation between particle species and facilitates a more robust and stable particle identification. The transformation procedure consists of three sequential steps – scaling, shifting, and rotation – and is mathematically defined by Eqs.

Specifically, the scaling step equalizes the effective resolution of the two variables to ensure neither dimension dominates the particle separation in the transformed space. p. 4
Where the paper says it

5 First, a scale factor fscale is introduced to normalize the widths of the distributions along the nσπ and m2 axes. The scaling step equalizes the effective resolution of the two variables, ensuring that neither dimension dominates the particle separation in the transformed space.

Optimization and Fitting Strategy

The transformation parameters are optimized such that the effective Gaussian widths of the particle peaks along x and y directions become comparable, while the pion and kaon distributions are predominantly aligned along the x axis. The transformation is mathematically defined by equations involving a scale factor fscale, linear translation parameters µ(nσπ) and µ(m2(π)), and a rotation angle α. p. 2

Where the paper says it

This procedure defines a new coordinate system, x(nσπ, m2 ) and y(nσπ, m2 ), obtained through a linear shift and rotation followed by a translation of the original variables. The transformation parameters are optimized such that the effective Gaussian widths of the particle peaks along x and y directions become comparable, while the pion and kaon distributions are predominantly aligned along the x axis.

The resulting distributions are described using a two-dimensional Student’s t function in the x-y plane, which includes an additional degree-of-freedom parameter ν for enhanced flexibility. p. 5
Where the paper says it

In this analysis, the two-dimensional distributions of individual particle species are modeled using a two-dimensional Student’s t function, as defined in Eq. Compared to a standard Gaussian function, the Student’s t function includes an additional degree-of-freedom parameter, ν, which provides enhanced flexibility in describing non-Gaussian tails that are commonly observed in experimental particle identification distribution.

This multi-particle fit function contains 13 free parameters, and the fitting procedure is performed iteratively to ensure a stable and physically meaningful solution. p. 5
Where the paper says it

In total, the combined two-dimensional multi-particle fit function describing the pion and kaon distributions contains 13 free parameters, as listed in Eq. Given the relatively large parameter space and potential correlations among parameters, the fitting procedure is performed iteratively.

Performance Evaluation and Results

The performance of the proposed technique is evaluated using simulated Au+Au collisions generated with the AMPT model, which incorporates realistic, pT-dependent PID resolution effects via a data-driven smearing procedure. The results show that the extracted yields for both pions and kaons remain in excellent agreement with the AMPT input over the entire studied transverse momentum range. p. 7

Where the paper says it

By construction, the method preserves event-by-event correlations and is therefore suitable for anisotropic flow analyses. Comparisons with the known AMPT input demonstrate that the extracted pion and kaon yields obtained with the two-dimensional method remain in excellent agreement with the input over the entire studied transverse momentum range.

Specifically, even at high-pT up to pT = 3. p. 6
Where the paper says it

Over the full transverse momentum range considered, the extracted yields show excellent agreement with the model input. In particular, even in the high-pT region up to pT = 3.

0 GeV/c, the method reproduces input yields with a deviation of less than 2% for both species. p. 6
Where the paper says it

In particular, even in the high-pT region up to pT = 3. 0 GeV/c, where particle identification becomes increasingly challenging due to reduced detector resolution, the two-dimensional method reproduces the input yields with a deviation of less than 2% for both pions and kaons.

Furthermore, the extracted purity remains close to unity across the entire pT range, indicating a particle identification purity exceeding 98% even at the highest transverse momenta. p. 6
Where the paper says it

5 show the extraction purity, defined as the ratio of the extracted yield to the AMPT input yield. The purity remains close to unity across the entire pT range, indicating a particle identification purity exceeding 98% even at the highest transverse momenta.

Impact on Flow Measurements

The impact of the improved particle identification on elliptic flow measurements is examined by extracting the elliptic flow coefficient v2 via azimuthal yield fits. The reconstructed v2 values for both pions and kaons are consistent with the AMPT input within statistical uncertainties over the entire pT range. p. 8

Where the paper says it

The impact of the improved particle identification on elliptic flow measurements is further examined by extracting v2 via azimuthal yield fits. The reconstructed v2 values for both pions and kaons are consistent with the AMPT input within statistical uncertainties, confirming that the proposed method does not introduce biases in flow observables.

This confirms that the proposed method does not introduce biases in flow observables. p. 8
Where the paper says it

The impact of the improved particle identification on elliptic flow measurements is further examined by extracting v2 via azimuthal yield fits. The reconstructed v2 values for both pions and kaons are consistent with the AMPT input within statistical uncertainties, confirming that the proposed method does not introduce biases in flow observables.

The extended pT reach provided by this technique is crucial for precision flow measurements and differential studies relying on identified hadrons. p. 8
Where the paper says it

The extended pT reach provided by this technique is particularly relevant for studies of identified-hadron flow and number-of-constituent-quark scaling, which rely critically on precise measurements at intermediate and high transverse momentum. Overall, the two-dimensional shift and rotation method offers a robust and general framework for improving PID performance in future high-precision heavy-ion measurements.

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

  1. Data-driven particle identification method based on a two-dimensional shift and rotation transformation can be applied to enhance pion and kaon separation in heavy-ion collisions. This method exploits correlated information between the time-of-flight mass square (m2) and ionization energy loss (nσ) distributions to construct a new variable space where species are more clearly separated.

  2. The improved AI system can accurately extract particle yields up to pT = 3.0 GeV/c with a purity exceeding 98%, extending the reliable identification range beyond the limitations of conventional one-dimensional fitting methods. This capability is crucial for precision measurements in high momentum regimes where species overlap is severe.

  3. The system can robustly extract elliptic flow coefficients (v2) for charged pions and kaons over the entire transverse momentum range studied, demonstrating that the improved particle separation achieved by the two-dimensional shift and rotation method provides enhanced robustness, particularly at higher transverse momentum where species overlap becomes increasingly severe.

Abstract

Accurate identification of charged pions and kaons is essential for precision measurements in relativistic heavy-ion collisions, but becomes increasingly challenging at intermediate and high transverse momentum because of the increasing overlap of the TPC and TOF particle identification responses. In this work, we quantitatively benchmark a two-dimensional shift-and-rotation PID procedure that combines the nσ π and mass-squared (m squared) information. The benchmark is performed using Au+Au collision events generated with the AMPT model, where the detector response is modeled using p T-dependent PID parameters extracted from STAR data. Since the true particle identities and input yields are known in AMPT but are not used during the reconstruction, the procedure can be evaluated through a direct closure test that is not possible with experimental data alone. The two-dimensional procedure is compared with a reference one-dimensional method consisting of an nσ π< 3 preselection followed by a Gaussian fit to the m squared distribution. While both methods perform well when pion and kaon PID responses are well separated, the two-dimensional procedure provides improved yield closure as the PID distributions increasingly overlap at high p T. The reconstructed-to-input yield ratio remains within 2% of unity up to p T = 3 GeV/ c, whereas the reference one-dimensional procedure shows a larger deviation. Reaction-plane v 2 closure tests show that both the reference one-dimensional and two-dimensional PID procedures reproduce the AMPT input v 2 within the statistical uncertainties over the studied p T range. These results demonstrate that the two-dimensional procedure substantially improves the high- p T yield reconstruction without introducing a detectable bias in the extracted v 2 under the STAR-like detector response considered here.

Sources

Related papers