Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459.2-604207

arXiv:2608.21196 · astro-ph.HE · Submitted 2026-08-21 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: We're looking at the paper: "Constraining neutron-star properties with ensembles of thermonuclear bursts".

Jocelyn: Here's the gist of the paper. Deducing the properties of the host neutron stars from thermonuclear (type-I) X-ray bursts remains a challenge, due to incomplete data, the large (multidimensional) parameter space, and dearth of suitable models and analysis tools The paper describes a new “ensemble” analysis mode for the BEANSP package that utilizes consistent and regular “clocked” bursting to compare model predictions to epoch averages This approach is applied to the accretion-powered millisecond pulsar SRGA J144459.2−604207, yielding reasonably good agreement with model predictions for a mildly sub-solar fuel composition The paper introduces a modified approach that takes advantage of ‘ensembles’ of observed bursts with unambiguous recurrence times to constrain system parameters In “ensemble” mode, BEANSP calculates the accretion rate for each epoch given the estimated bolometric persistent flux and model parameters d, ξp, MNS, RNS The predicted quantities are then converted to observational quantities and compared against the observations to generate an overall likelihood comprising contributions for all epochs Data Assembly and Analysis The researchers assembled a set of 14 daily burst epochs covering the 2024 outburst of SRGA J144459.2−604207 by comparing observations to SETTLE model predictions They focused on systems exhibiting “clocked” bursting, such as GS 1826−24, and utilized data from X-ray satellites including Rossi X-ray Timing Explorer, Insight-HXMT and IXPE For SRGA J144459.2−604207, they derived daily mean recurrence times ∆t and calculated averages of the burst fluence Eb, persistent (bolometric) flux Fp and α value for each epoch Key Findings on System Properties The analysis of SRGA J144459.2−604207 resulted in reasonably good agreement with model predictions for a mildly sub-solar fuel composition, with H-fraction X ≃ 0.54 and CNO metallicity ZCNO ≃ 0.01 The inferred distance depends on assumptions about the system inclination and corresponding anisotropy of the persistent emission, likely in the range 6–11 kpc Constraints on Neutron Star Parameters The posterior distributions for SRGA J144459.2−604207 showed that while parameters like X and ZCNO were well-constrained, the anisotropy parameters ξb and ξp were relatively poorly constrained The constraints on distance, which is often one of the most sought-after quantities for such systems, depend inextricably on assumptions about the degree of anisotropy of the emission Model Fidelity and Future Work The study demonstrated that the ensemble model comparison can replicate input parameters in simulations irrespective of whether α values are included in the likelihood calculation The authors suggest that future work comparing SRGA J144459.2−604207 with a more accurate model such as KEPLER could provide more reliable constraints on the fuel composition However, constraints on distance, anisotropy and neutron star mass and radius are less well constrained in the absence of informative priors The ensemble model comparison is broadly successful in accurately constraining the model parameters relating to burst ignition with or without including the α values in the likelihood calculation The authors recommend that best estimates of observational α values, where available, be included in likelihoods for use with BEANSP runs to provide an important consistency check for the model Reference List Chenevez, J., Galloway, D.

Vera: K., in ’t Zand, J.

Jocelyn: J.

Vera: M., Tomsick, J.

Paper summary: Vera: So we're looking at this new work called "Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459 point 2-six hundred four thousand two hundred seven <ref:2608.21196#pg1,Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA>." The main problem they're tackling is figuring out the actual properties of these host neutron stars using those X-ray bursts. They point out that we have incomplete data, a huge parameter space to navigate, and there just aren't enough good models or tools around right now.

Jocelyn: It sounds like they are trying to build a better way to use the BEANSP package, which is what they used before, by adding this new "ensemble" analysis mode. What’s the core idea behind this ensemble approach?

Vera: Essentially, they are using bursts that have consistent and regular recurrence times. They compare what their model predicts against the average of these clocked bursts across different epochs. This helps them constrain those system parameters we're trying to figure out, like how fast the accretion rate is changing.

Subrahmanyan: So they’re moving from looking at single events to looking at a whole collection of consistent timing information to narrow down the possible physical settings of these neutron stars and their environments. That's a big step toward understanding the physics happening on those surfaces.

Vera: Exactly. They apply this ensemble mode to an accretion-powered millisecond pulsar, SRGA J144459 point 2−six hundred four thousand two hundred seven and they found that it gives them reasonably good agreement when they test a mildly sub-solar fuel composition for the burst material <ref:2608.21196#pg1>.

Jocelyn: What does that mean for the physics of the burst itself? Are we getting any concrete numbers on what’s actually fueling those X-ray explosions?

Subrahmanyan: They found specific values for that fuel composition, stating it had an H-fraction of about zero point five four and a CNO metallicity around zero point zero one. These are parameters that describe the material being burned on the neutron star's surface, which is key to understanding the explosion mechanism itself and how much mass is involved in these events.

Vera: And they also inferred a distance for this system, though that distance depends on some assumptions about how we think the persistent emission from the neutron star is beamed or anisotropic. They estimated it to be somewhere in the range of six to eleven kiloparsecs.

Jocelyn: So what's still left fuzzy? If they got those fuel composition numbers and a distance estimate, what are the uncertainties that are still hanging over this result?

Subrahmanyan: The constraints on parameters like X and CNO metallicity were pretty well-defined, but the anisotropy parameters, specifically xi b and xi p, were relatively poorly constrained. And as they noted in their work "Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459 point 2-six hundred four thousand two hundred seven", the constraints on distance depend on assumptions about the degree of anisotropy of the persistent emission > <ref:2608.21196#pg1,Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA>

Vera: That's true. The paper shows that while their ensemble model comparison can replicate input parameters in simulations whether or not they include those alpha values in the likelihood calculation, it doesn't help as much with system parameters like distance or neutron star mass and radius unless they have some strong priors to start with.

Jocrahmanyan: So, even though they can constrain the ignition properties of the burst well, we still need more information about the neutron star itself—its exact mass and radius—unless we get really good data on those anisotropy parameters. That suggests that future work comparing SRGA J144459 point 2−six hundred four thousand two hundred seven with a more accurate model like KEPLER could provide better constraints on the fuel composition, but not necessarily the neutron star structure itself > <ref:2608.21196#pg1>

Jocelyn: It sounds like the ensemble analysis is a really useful tool for checking consistency in their models, even if it doesn't solve every mystery about what’s happening on those stars yet. What should we take away from this study?

Vera: The main point is that using these clocked bursts in an ensemble mode helps us test model fidelity. They suggest including the best estimates of observational alpha values when running BEANSP to get a better consistency check for the model >

Subrahmanyan: For cosmology, this work is important because it shows how we can use these specific astrophysical phenomena to constrain the properties of dense matter and extreme environments, which are often related to processes happening in the early universe structures >

Jocelyn: So, it’s about getting a more solid handle on what’s going on during those bursts, even if pinning down everything about the neutron star remains difficult without more data. That sets up nicely for what we look at next.

Conclusion: Vera: So we're wrapping up on this paper about constraining neutron star properties using bursts from SRGA J144459 point two-six hundred four thousand two hundred seven by looking at ensembles of data that have regular recurrence times > <ref:2608.21196#pg1>

Jocelyn: It’s interesting because they’re not just looking at one burst; they're comparing model predictions to the average behavior across a whole set of bursts, which helps them nail down those system parameters >

Subrahmanyan: The core idea is using this ensemble mode to test how well their models predict the physical conditions for thermonuclear ignition on these neutron stars >

Vera: Right, and what they found is that for a mildly sub-solar fuel composition, like an H-fraction of about zero point five four and CNO metallicity around zero point zero one, the model predictions match the observations reasonably well >

Jocelyn: And they also managed to infer a distance for this system, though that distance depends on some assumptions about how we think the persistent emission from the neutron star is beamed >

Subrahmanyan: That distance estimate is really interesting because it links these burst observations to the broader picture of how matter accretes onto these compact objects >

Vera: What’s still a bit fuzzy, though, are those constraints on anisotropy parameters like xi b and xi p; they weren't very well constrained in their final posterior distributions >

Jocelyn: And that distance estimate itself is really tied up in those assumptions about the emission's shape, so we still have a lot of wiggle room there >

Subrahmanyan: So even with these results, pinning down the neutron star mass and radius remains tough without more information on those anisotropy parameters or some strong prior knowledge >

Vera: It shows that while this ensemble comparison is good for checking if their ignition model works, it doesn't immediately give us a solid number for the neutron star itself >

Jocelyn: They also suggest that including the observed alpha values when running BEANSP provides a better consistency check, even though it didn't drastically change the constraints on distance or mass >

Subrahmanyan: So to sum up, this work is good at nailing down the fuel composition and checking model fidelity for these burst ignition properties >

Vera: And that sets us up nicely because we still need those stronger priors on things like neutron star radius and mass to really tighten those constraints >

School of Physics & Astronomy, Monash University · Institute for Globally-Distributed Research and Education (IGDORE) · International Centre for Radio Astronomy Research, Curtin University Institute for Globally-Distributed Research and Education (IGDORE) · Key Laboratory of Stars and Interstellar Medium, Xiangtan University · School of Science, Qingdao University of Technology

astro-ph.HE

Submitted: 2026-08-21

Updated: 2026-11-16

Comments: 15 pages, 8 figures, 7 tables; accepted by PASA. Supporting data and code at https://doi.org/10.26180/33065816

Code: https://github.com/adellej/beansdhttps:

License: http://creativecommons.org/licenses/by-nc-sa/4.0/

Importance score: 73/100

The gist: Deducing the properties of the host neutron stars from thermonuclear (type-I) X-ray bursts remains a challenge, due to incomplete data, the large (multidimensional) parameter space…

Key concepts

Ensemble Analysis Mode
This is a new analysis technique that uses multiple observed bursts with known recurrence times to constrain system parameters. It calculates the accretion rate for each burst epoch based on model predictions and compares the resulting observational quantities against all recorded epochs to build an overall likelihood.
Clocked Bursting
This refers to systems where thermonuclear bursts exhibit consistent and regular recurrence times. The paper specifically focused on these 'clocked' systems because their regular timing provides unambiguous constraints for fitting the physical parameters of the neutron star system.
Fuel Composition (H-fraction X and ZCNO)
The analysis determined the chemical makeup of the fuel on the neutron star surface. The results indicated a mildly sub-solar composition, specifically an H-fraction (Hydrogen) of about 0.54 and a CNO metallicity (Z_CNO) around 0.01, which are key inputs for understanding how the burst occurs.
Anisotropy Parameters ($\xi_b$ and $\xi_p$)
These parameters describe the directionality or 'anisotropy' of the persistent X-ray emission from the neutron star. While they were constrained, their values are relatively poorly determined in this study because constraints on distance and other properties depend heavily on assumptions about how anisotropic this emission is.

Terminology

Summary

Deducing the properties of the host neutron stars from thermonuclear (type-I) X-ray bursts remains a challenge, due to incomplete data, the large (multidimensional) parameter space, and dearth of suitable models and analysis tools The paper describes a new “ensemble” analysis mode for the BEANSP package that utilizes consistent and regular “clocked” bursting to compare model predictions to epoch averages This approach is applied to the accretion-powered millisecond pulsar SRGA J144459. p. 0

Where the paper says it

3 , and Li, Zhaosheng5 1School of Physics & Astronomy, Monash University, Clayton, VIC 3800, Australia and 2 Institute for Globally-Distributed Research and Education (IGDORE); 3 International Centre for Radio Astronomy Research, Curtin University, GPO Box U1987, Perth, WA 6845, Australia; 4 Key Laboratory of Stars and Interstellar Medium, Xiangtan University, Xiangtan 411105, Hunan, China; 5School of Science, Qingdao University of Technology, Qingdao 266525, China Abstract Deducing the properties of the host neutron stars from thermonuclear (type-I) X-ray bursts remains a challenge, due to incomplete data, the large (multidimensional) parameter space, and dearth of suitable models and analysis tools. Here we describe further development of the BEANSP package, used previously to analyse several burst samples.

2−604207, yielding reasonably good agreement with model predictions for a mildly sub-solar fuel composition

Methodology of Ensemble Analysis

The paper introduces a modified approach that takes advantage of ‘ensembles’ of observed bursts with unambiguous recurrence times to constrain system parameters In “ensemble” mode, BEANSP calculates the accretion rate for each epoch given the estimated bolometric persistent flux and model parameters d, ξp, MNS, RNS The predicted quantities are then converted to observational quantities and compared against the observations to generate an overall likelihood comprising contributions for all epochs

Data Assembly and Analysis

The researchers assembled a set of 14 daily burst epochs covering the 2024 outburst of SRGA J144459.2−604207 by comparing observations to SETTLE model predictions They focused on systems exhibiting “clocked” bursting, such as GS 1826−24, and utilized data from X-ray satellites including Rossi X-ray Timing Explorer, Insight-HXMT and IXPE For SRGA J144459. p. 1

Where the paper says it

Data and analysis We adopted new and previously-analysed observations of thermonuclear bursts made by X-ray satellites including the Rossi X-ray Timing Explorer, Insight-HXMT and IXPE. As the new “ensemble” approach to matching burst observations requires regular, consistent bursts, we focused on a few systems that show such behaviour; the best known is GS 1824−238 (e.

2−604207, they derived daily mean recurrence times ∆t and calculated averages of the burst fluence Eb, persistent (bolometric) flux Fp and α value for each epoch

Key Findings on System Properties

The analysis of SRGA J144459.2−604207 resulted in reasonably good agreement with model predictions for a mildly sub-solar fuel composition, with H-fraction X ≃ 0. p. 0

Where the paper says it

2−604207, and attempted to constrain the system properties of this object by comparing to SETTLE model predictions. We find reasonably good agreement between the observations and model predictions for a mildly sub-solar fuel composition, with H-fraction X ≃ 0.

54 and CNO metallicity ZCNO ≃ 0. p. 0
Where the paper says it

We find reasonably good agreement between the observations and model predictions for a mildly sub-solar fuel composition, with H-fraction X ≃ 0. 54 and CNO metallicity ZCNO ≃ 0.

01 The inferred distance depends on assumptions about the system inclination and corresponding anisotropy of the persistent emission, likely in the range 6–11 kpc

Constraints on Neutron Star Parameters

The posterior distributions for SRGA J144459.2−604207 showed that while parameters like X and ZCNO were well-constrained, the anisotropy parameters ξb and ξp were relatively poorly constrained The constraints on distance, which is often one of the most sought-after quantities for such systems, depend inextricably on assumptions about the degree of anisotropy of the emission

Model Fidelity and Future Work

The study demonstrated that the ensemble model comparison can replicate input parameters in simulations irrespective of whether α values are included in the likelihood calculation The authors suggest that future work comparing SRGA J144459.2−604207 with a more accurate model such as KEPLER could provide more reliable constraints on the fuel composition However, constraints on distance, anisotropy and neutron star mass and radius are less well constrained in the absence of informative priors The ensemble model comparison is broadly successful in accurately constraining the model parameters relating to burst ignition with or without including the α values in the likelihood calculation The authors recommend that best estimates of observational α values, where available, be included in likelihoods for use with BEANSP runs to provide an important consistency check for the model

Reference List

Chenevez, J., Galloway, D. p. 0

Where the paper says it

xx Research Paper Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459. 2−604207 Galloway, Duncan K.

p. 0
Where the paper says it

xx Research Paper Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459. 2−604207 Galloway, Duncan K.

p. 0
Where the paper says it

xx Research Paper Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459. 2−604207 Galloway, Duncan K.

p. 0
Where the paper says it

xx Research Paper Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459. 2−604207 Galloway, Duncan K.

p. 0
Where the paper says it

xx Research Paper Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459. 2−604207 Galloway, Duncan K.

p. 0
Where the paper says it

xx Research Paper Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459. 2−604207 Galloway, Duncan K.

K., in ’t Zand, J. p. 6
Where the paper says it

, Galloway, D. , in ’t Zand, J.

J. M., Tomsick, J. p. 6
Where the paper says it

, in ’t Zand, J. , Tomsick, J.

A., Barret, D. p. 6
Where the paper says it

, Tomsick, J. , Barret, D.

, Chakrabarty, D. p. 6
Where the paper says it

, Barret, D. , Chakrabarty, D.

, Furst, F. p. 6
Where the paper says it

, Chakrabarty, D. , Furst, F.

, Boggs, S. p. 6
Where the paper says it

, Furst, F. , Boggs, S.

E., Christensen, F. p. 6
Where the paper says it

, Boggs, S. , Christensen, F.

E., Craig, W. p. 6
Where the paper says it

, Christensen, F. , Craig, W.

W., ¨ Hailey, C. p. 6
Where the paper says it

, Craig, W. , ¨ Hailey, C.

J., Harrison, F. p. 6
Where the paper says it

, ¨ Hailey, C. , Harrison, F.

A., Romano, P. p. 6
Where the paper says it

, Harrison, F. , Romano, P.

, Stern, D. p. 6
Where the paper says it

, Romano, P. , Stern, D.

, & Zhang, W. p. 6
Where the paper says it

, Stern, D. , & Zhang, W.

W. 2016 Di Salvo, T. p. 1
Where the paper says it

That and other instruments that observed the source early in the outburst also revealed frequent, regular thermonuclear bursts without strong evidence for photospheric radius-expansion, unusual for accretion-powered millisecond pulsars (e. Di Salvo and Sanna 2022).

& Sanna, A Accretion Powered X-ray Millisecond Pulsars Foreman-Mackey, D. p. 6
Where the paper says it

& Sanna, A. Accretion Powered X-ray Millisecond Pulsars.

, Hogg, D. W., Lang, D., & Goodman, J. p. 3
Where the paper says it

Usually this fraction would start out > 0. 5 and decrease throughout the run; reducing the stretch a parameter, which quantifies the scale of the “stretch move” algorithm for updating the walker positions (Goodman and Weare 2010), has the effect of stopping (or even reversing) this decrease.

2013 Fu, T. p. 2
Where the paper says it

MCMC runs We used the EMCEE code of Foreman-Mackey et al. (2013) for our MCMC runs, as described in Table 2.

, Li, Z., Pan, Y., Ji, L., Chen, Y., Kuiper, L. p. 6
Where the paper says it

Publications of the Astronomical Society of the Pacific, 125(925), pp. , Kuiper, L.

p. 6
Where the paper says it

Publications of the Astronomical Society of the Pacific, 125(925), pp. , Kuiper, L.

, Galloway, D. K., Falanga, M. p. 6
Where the paper says it

, Galloway, D. , Falanga, M.

, Xu, R., Li, X., Ge, M., Song, L. M., Zhang, S. p. 6
Where the paper says it

, Stern, D. , & Zhang, W.

., & Zhang S. p. 6
Where the paper says it

, Stern, D. , & Zhang, W.

-N. 2025 Fujimoto, M. p. 3
Where the paper says it

The relatively low values of both anisotropy parameters would seem to favour the lower inclination range, but the situation may not be this clear-cut. A comparison of our combined posteriors with the predictions of models by Fujimoto (1988) and He and Keek (2016) suggest that the ξb is typically too large compared to the ξp value for consistency with the models (Fig.

Y. 1988 Galloway, D. p. 3
Where the paper says it

The relatively low values of both anisotropy parameters would seem to favour the lower inclination range, but the situation may not be this clear-cut. A comparison of our combined posteriors with the predictions of models by Fujimoto (1988) and He and Keek (2016) suggest that the ξb is typically too large compared to the ξp value for consistency with the models (Fig.

K. & Cumming, A 2006 Galloway, D. p. 0
Where the paper says it

2019; Galloway et al. 2024; see also Galloway and Cumming 2006).

K., Goodwin, A. p. 0
Where the paper says it

2−604207 Galloway, Duncan K. 1,2 , Jones, Serena3 , Waterson, Luke1 , Fu, Tao4 , Goodwin, Adelle J.

p. 0
Where the paper says it

2−604207 Galloway, Duncan K. 1,2 , Jones, Serena3 , Waterson, Luke1 , Fu, Tao4 , Goodwin, Adelle J.

J., Hilder, T. p. 6
Where the paper says it

, Goodwin, A. , Hilder, T.

, Waterson, L. p. 0
Where the paper says it

2−604207 Galloway, Duncan K. 1,2 , Jones, Serena3 , Waterson, Luke1 , Fu, Tao4 , Goodwin, Adelle J.

, & Cupak, M. p. 6
Where the paper says it

, Waterson, L. , & Cupak, M.

2024 Galloway, D. p. 0
Where the paper says it

2019; Galloway et al. 2024; see also Galloway and Cumming 2006).

K., Johnston, Z. p. 1
Where the paper says it

We used an ensemble of bursts from GS 1824−238 assembled previously by Galloway et al. (2017), and compared against KEPLER models by Johnston et al.

, Goodwin, A., & He, C.-C. 2022 Galloway, D. p. 2
Where the paper says it

Third, we include the distance d, and burst and persistent flux emission anisotropy (ξb and ξp, respectively; see Galloway et al. 2022) as separate parameters, while Johnston et al.

K. & Keek, L. 2021 Goodman, J. p. 0
Where the paper says it

The burst ignition conditions are primarily determined by the rate of mass accretion onto the neutron star, and the composition of the accreted fuel (e. Galloway and Keek 2021); the majority of systems have main-sequence mass donors and consequently accrete H-rich fuel.

& Weare, J. p. 3
Where the paper says it

Usually this fraction would start out > 0. 5 and decrease throughout the run; reducing the stretch a parameter, which quantifies the scale of the “stretch move” algorithm for updating the walker positions (Goodman and Weare 2010), has the effect of stopping (or even reversing) this decrease.

2010 Goodwin, A. p. 6
Where the paper says it

2010, CAMS, 5(3), 65–80. Goodwin, A.

J., Galloway, D. K., Heger, A. p. 5
Where the paper says it

consistent with this limit), in a model which fully takes into account the heating effect of steady burning prior to ignition, also providing “thermal inertia” which allows earlier ignition (e. Heger et al.

p. 5
Where the paper says it

consistent with this limit), in a model which fully takes into account the heating effect of steady burning prior to ignition, also providing “thermal inertia” which allows earlier ignition (e. Heger et al.

p. 5
Where the paper says it

consistent with this limit), in a model which fully takes into account the heating effect of steady burning prior to ignition, also providing “thermal inertia” which allows earlier ignition (e. Heger et al.

, Cumming, A. p. 0
Where the paper says it

2019; Galloway et al. 2024; see also Galloway and Cumming 2006).

p. 0
Where the paper says it

2019; Galloway et al. 2024; see also Galloway and Cumming 2006).

, & Johnston, Z. p. 1
Where the paper says it

We used an ensemble of bursts from GS 1824−238 assembled previously by Galloway et al. (2017), and compared against KEPLER models by Johnston et al.

2019 He, C. p. 0
Where the paper says it

The BEANSP code has been used to determine NS system parameters by comparing observations of thermonuclear bursts to numerical models, for several sources (Goodwin et al. 2019; Galloway et al.

-C. & Keek, L. 2016 Heger, A. p. 5
Where the paper says it

consistent with this limit), in a model which fully takes into account the heating effect of steady burning prior to ignition, also providing “thermal inertia” which allows earlier ignition (e. Heger et al.

, Cumming, A., Galloway, D. K., & Woosley, S. p. 6
Where the paper says it

, Galloway, D. , & Woosley, S.

E. 2007 Johnston, Z. p. 1
Where the paper says it

(2017), and compared against KEPLER models by Johnston et al. These data include measurements of the recurrence time ∆t, integrated burst fluence Eb, and bolometric persistent flux Fp, over three epochs observed between 1998 and 2007.

, Heger, A., & Galloway, D. p. 0
Where the paper says it

xx Research Paper Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459. 2−604207 Galloway, Duncan K.

p. 0
Where the paper says it

xx Research Paper Constraining neutron-star properties with ensembles of thermonuclear bursts: application to SRGA J144459. 2−604207 Galloway, Duncan K.

K. 2018 Johnston, Z. p. 2
Where the paper says it

We also used the burst simulations in Table 2 of Johnston et al. (2018) to provide synthetic burst data based on the 2002 outburst of SAX J1808.

, Heger, A., & Galloway, D. K. 2020 Li, Z. p. 1
Where the paper says it

For now, the choice is limited to the interpolated KEPLER grid used by Johnston et al. (2020) also to match the bursts from GS 1824−238.

, Kuiper, L., Pan, Y., Xu, R., Chen, Y., Ge, M., Huang, Y. p. 6
Where the paper says it

, Kuiper, L. , Huang, Y.

, Jia, S., Li X., Song L., Qu J., Zhang S.., Tao L., Feng H., Zhang S.-N.., & Falanga M. p. 6
Where the paper says it

, Galloway, D. , Falanga, M.

2025 Malacaria, C. p. 2
Where the paper says it

(2025) to derive i = (50. 3 ) ◦ , consistent with the value adopted by Malacaria et al.

, Papitto, A. p. 1
Where the paper says it

(2025) derived an upper limit on the Hydrogen mass fraction of X < 0. Detection of polarisation in the persistent emission led to constraints on the inclination (Papitto et al.

, Campana, S. p. 6
Where the paper says it

, Papitto, A. , Campana, S.

, Di Marco, A. p. 6
Where the paper says it

, Campana, S. , Di Marco, A.

p. 6
Where the paper says it

, Campana, S. , Di Marco, A.

, Di Salvo, T. p. 1
Where the paper says it

That and other instruments that observed the source early in the outburst also revealed frequent, regular thermonuclear bursts without strong evidence for photospheric radius-expansion, unusual for accretion-powered millisecond pulsars (e. Di Salvo and Sanna 2022).

, Cristina Baglio, M. p. 6
Where the paper says it

, Di Salvo, T. , Cristina Baglio, M.

, Illiano, G. p. 6
Where the paper says it

, Cristina Baglio, M. , Illiano, G.

, La Placa, R. p. 6
Where the paper says it

, Illiano, G. , La Placa, R.

, Miraval Zanon, A. p. 6
Where the paper says it

, La Placa, R. , Miraval Zanon, A.

, Pilia M., Poutanen J. p. 6
Where the paper says it

, Pilia, M. , Poutanen, J.

p. 6
Where the paper says it

, Pilia, M. , Poutanen, J.

p. 6
Where the paper says it

, Pilia, M. , Poutanen, J.

, Salmi T., Sanna A.., & Mandal M. p. 2
Where the paper says it

2−604207 system has been estimated by a number of methods. Fitting the reflection spectrum led Mandal et al.

2025 Mandal, M. p. 2
Where the paper says it

Fitting the reflection spectrum led Mandal et al. (2025) to derive i = (50.

, Naik, S., & Chhotaray B. p. 6
Where the paper says it

A&A, 699, A288. , & Chhotaray, B.

2025 Molkov, S. p. 2
Where the paper says it

3 ) ◦ , consistent with the value adopted by Malacaria et al. (2025) based on the same approach (see also Molkov et al.

V., Lutovinov A. p. 6
Where the paper says it

MNRAS, 542(1), 443–455. , Lutovinov, A.

A., Tsygankov S. p. 6
Where the paper says it

, Lutovinov, A. , Tsygankov, S.

S., Suleimanov V. p. 6
Where the paper says it

, Tsygankov, S. , Suleimanov, V.

F., Poutanen J., Lapshov I. p. 6
Where the paper says it

, Poutanen, J. , Lapshov, I.

Y., Mereminskiy I. p. 6
Where the paper says it

, Lapshov, I. , Mereminskiy, I.

A., Semena A. p. 6
Where the paper says it

, Mereminskiy, I. , Semena, A.

N., Arefiev V. p. 6
Where the paper says it

, Semena, A. , Arefiev, V.

A., & Tkachenko A. p. 6
Where the paper says it

, Arefiev, V. , & Tkachenko, A.

Y. 2024 Ng, M. p. 0
Where the paper says it

We performed a number of experiments on simulated data, demonstrating that we can recover input parameters related to the burst ignition with reasonable accuracy, but less so for the system distance, emission anisotropy and neutron star mass and radius. Finally, we assembled a set of 14 daily burst epochs covering the 2024 outburst of the accretion-powered millisecond pulsar SRGA J144459.

, Sanna, A. p. 1
Where the paper says it

That and other instruments that observed the source early in the outburst also revealed frequent, regular thermonuclear bursts without strong evidence for photospheric radius-expansion, unusual for accretion-powered millisecond pulsars (e. Di Salvo and Sanna 2022).

, Strohmayer T. p. 7
Where the paper says it

, Sanna, A. , Strohmayer, T.

p. 7
Where the paper says it

, Sanna, A. , Strohmayer, T.

E., Arzoumanian Z. p. 7
Where the paper says it

, Strohmayer, T. , Arzoumanian, Z.

, Gendreau K. p. 7
Where the paper says it

, Arzoumanian, Z. , Gendreau, K.

C., Ray P. S., Coley J. B., Chakrabarty D. p. 6
Where the paper says it

, Barret, D. , Chakrabarty, D.

, Guillot S. p. 7
Where the paper says it

, Chakrabarty, D. , Guillot, S.

, Bogdanov S. p. 7
Where the paper says it

, Guillot, S. , Bogdanov, S.

p. 7
Where the paper says it

, Guillot, S. , Bogdanov, S.

, Altamirano D. p. 7
Where the paper says it

, Bogdanov, S. , Altamirano, D.

, Chenevez J. p. 1
Where the paper says it

These systems are often termed “clocked bursters” in the literature, although this term is a misnomer as such sources show regular (“clocked”) bursting behaviour only episodically (see e. Chenevez et al.

, Hare J., Wolff M. T., Guver T., Jaisawal G. K., Wadiasingh Z., & Ferrara E. C. 2024 Papitto, A. p. 1
Where the paper says it

(2025) derived an upper limit on the Hydrogen mass fraction of X < 0. Detection of polarisation in the persistent emission led to constraints on the inclination (Papitto et al.

, Di Marco, A., Poutanen J., Salmi T., Illiano G. p. 6
Where the paper says it

, Cristina Baglio, M. , Illiano, G.

, La Monaca F. p. 7
Where the paper says it

, Illiano, G. , La Monaca, F.

, Ambrosino F. p. 7
Where the paper says it

, La Monaca, F. , Ambrosino, F.

, Bobrikova A. p. 7
Where the paper says it

, Ambrosino, F. , Bobrikova, A.

, Baglio M. p. 6
Where the paper says it

, Di Salvo, T. , Cristina Baglio, M.

C., Ballocco C. p. 7
Where the paper says it

, Baglio, M. , Ballocco, C.

, Burderi L. p. 7
Where the paper says it

, Ballocco, C. , Burderi, L.

, Campana S. p. 6
Where the paper says it

, Papitto, A. , Campana, S.

., Coti Zelati F. p. 7
Where the paper says it

, Campana, S. , Coti Zelati, F.

, Di Salvo T. p. 1
Where the paper says it

That and other instruments that observed the source early in the outburst also revealed frequent, regular thermonuclear bursts without strong evidence for photospheric radius-expansion, unusual for accretion-powered millisecond pulsars (e. Di Salvo and Sanna 2022).

, La Placa R. p. 6
Where the paper says it

, Illiano, G. , La Placa, R.

, Loktev V. p. 7
Where the paper says it

, La Placa, R. , Loktev, V.

, Long S., Malacaria C. p. 2
Where the paper says it

(2025) to derive i = (50. 3 ) ◦ , consistent with the value adopted by Malacaria et al.

, Miraval Zanon A. p. 6
Where the paper says it

, La Placa, R. , Miraval Zanon, A.

, Ng M., Pilia M., Sanna A., Stella L. p. 7
Where the paper says it

, Sanna, A. , Stella, L.

, Strohmayer T., & Zane S. p. 7
Where the paper says it

, Strohmayer, T. , & Zane, S.

2025 Steiner, A. p. 7
Where the paper says it

A&A, 694, A37. Steiner, A.

W., Heinke, C. p. 7
Where the paper says it

Steiner, A. , Heinke, C.

O., Bogdanov S., Li C. K., Ho W. C. G., Bahramian A.," & Han S. 2018 Thompson, T. p. 2
Where the paper says it

Such an approach might be necessary if there were systematic errors in the estimation of the bolometric flux (e. Thompson et al.

W. J., Galloway, D. K., Rothschild R. p. 6
Where the paper says it

, Chakrabarty, D. , & Rothschild, R.

p. 6
Where the paper says it

, Chakrabarty, D. , & Rothschild, R.

p. 6
Where the paper says it

, Chakrabarty, D. , & Rothschild, R.

E., & Homer L. p. 7
Where the paper says it

, Rothschild, R. , & Homer, L.

p. 7
Where the paper says it

, Rothschild, R. , & Homer, L.

2008 Thompson, T. p. 7
Where the paper says it

, & Homer, L. 2008, Deviations from the Flux-Recurrence Time Relationship in GS 1826-238: Potential Transient Spectral Changes.

W. J., Rothschild R. E., Tomsick J. p. 6
Where the paper says it

, in ’t Zand, J. , Tomsick, J.

A., & Marshall H. p. 7
Where the paper says it

, Tomsick, J. , & Marshall, H.

L. 2005 Thompson, T. p. 7
Where the paper says it

, & Marshall, H. 2005, Chandra and RXTE Spectra of the Burster GS 1826-238.

W. J., Galloway, D. K., Rothschild R. E., & Homer L. 2008

Appendix Details

In Appendix A, the authors describe experiments to illustrate the effect of including observed α values in the burst model-observation comparison The inclusion of α in the likelihood when comparing models to observations confirms that this parameter, when measured reliably, provides an important consistency check for model fidelity In Appendix B, they describe simulation studies intended to test the reliability of BEANSP by generating synthetic burst data based on SETTLE mimicking SRGA J144459.2−604207 The simulations consistently replicated the burst ignition properties, including the H-fraction X, CNO mass fraction ZCNO, and the base flux Qb irrespective of the data treatment The ensemble model comparison is broadly successful in accurately constraining the model parameters relating to burst ignition with or without including the α values in the likelihood calculation but less so for system parameters including distance, anisotropy and neutron star mass and radius particularly in the absence of an informative prior on these parameters

Table Summaries

Table 1 lists averaged burst properties of SRGA J144459.2−604207 during (approximately) daily epochs of the 2024 outburst, derived from burst measurements made by IXPE and Insight-HXMT Table 3 lists inferred system parameters for SRGA J144459. p. 5

Where the paper says it

We applied the new method to 14 daily epochs of data from the profligate burster SRGA J144459. 2−604207, assembled from a sample of 94 unique events observed by Insight-HXMT and IXPE during the 2024 outburst.

2−604207, showing the posterior distributions for various runs Table B1 shows simulated data generated for testing BEANSP in ensemble mode Table B2 lists inferred properties for the simulated data for SRGA J144459. p. 2
Where the paper says it

In lieu of more certainty about the inclination and distance we retain the flat priors for ξb and ξp for most runs, but experiment with one more restricted prior in §4. We also generated a set of simulated data (using the sim data method of BEANSP) based loosely on the properties of SRGA J144459.

2−604207 as listed in Table B1, for comparison with the model input values Table B3 shows system parameters derived from the comparison with KEPLER simulations for SAX J1808. p. 8
Where the paper says it

Thus, we conclude that the ensemble model comparison is broadly successful in accurately constraining the model parameters relating to burst ignition, with or without including the α values in the likelihood calculation; but less so for the system parameters including distance, anisotropy and neutron star mass and radius, particularly in the absence of an informative prior on these parameters. 4−3658 burst train mode We performed two BEANSP runs on the simulated SAX J1808.

4−3658

Figure Summaries

Figure 1 compares observed and predicted burst properties for SRGA J144459.2−604207 from the last 105 samples of run base sel Figure 2 shows posterior distributions for SRGA J144459. p. 4

Where the paper says it

Posterior distributions for SRGA J144459. 2−604207 from the last 5 × 105 samples of run base sel mr.

2−604207 from the last 5 × 10 5 samples of run base sel mr Figure 3 shows combined posterior distributions for the anisotropy parameters ξb and ξp SRGA J144459. p. 5
Where the paper says it

Combined posterior distributions for the anisotropy parameters ξb and ξp SRGA J144459. 2−604207 from the last 5 × 105 samples of run base sel mr.

2−604207 from the last 5 × 10 5 samples of run base sel mr Figure 4 shows posterior distributions (blue shaded contours) for selected parameters for GS 1826−24, derived from the comparison of the KEPLER grid interpolation model to the 3-epoch reference data Figure B1 compares posterior distributions for runs sim1 alpha mr and sim1 prior Figure B2 compares predicted and input (simulated) burst properties for run sim1 alpha mr Figure B3 shows neutron-star parameter posterior distributions for 105 samples of the johnston18 simulation run to match KEPLER predictions for SAX J1808. p. 5
Where the paper says it

2 ξp/ξb Figure 4. Posterior distributions (blue shaded contours) for selected parameters for GS 1826−24, derived from the comparison of the KEPLER grid interpolation model to the 3-epoch reference data.

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. Ensemble analysis mode refinement: The system can now utilize a new 'ensemble' analysis mode that compares model predictions to epoch averages instead of individual events, which is much more efficient than the previous train mode. This allows for a single model evaluation per epoch.

  2. Constraint on fuel composition: The improved system can constrain the accreted composition by finding reasonably good agreement between observations and model predictions for a mildly sub-solar fuel composition, specifically inferring an H-fraction X of 0.54 and CNO metallicity ZCNO of 0.01 for SRGA J144459.2−604207.

  3. Distance estimation refinement: The system can constrain the inferred distance to be in the range of 6–11 kpc, although it notes that this inferred distance depends on assumptions about the system inclination and corresponding anisotropy of the persistent emission.

  4. Model parameter recovery: The system can recover input parameters related to burst ignition with reasonable accuracy when using simulated data, though less so for system distance, emission anisotropy and neutron star mass and radius.

  5. Incorporation of alpha values: The system can include the alpha-value in the likelihood calculation when comparing models to observations, which confirms that this parameter provides an important consistency check for model fidelity when measured reliably.

  6. Prior application refinement: The system can incorporate a prior intended to reflect one of the inclination constraints derived from polarization measurements, which results in a lower distance but unrealistically massive and compact neutron star, allowing researchers to test the impact of these assumptions.

Abstract

Deducing the properties of the host neutron stars from thermonuclear (type-I) X-ray bursts remains a challenge, due to incomplete data, multidimensional parameter space, and dearth of suitable models and analysis tools. Here we describe further development of the BEANSP package, with a new "ensemble" analysis mode utilising the consistent and regular "clocked" bursting exhibited by some sources. This mode requires only one model evaluation per epoch, and so is much more efficient than the previous "train" mode. We apply the code to the best-known source exhibiting "clocked" bursting, GS 1826-24, utilising a grid of KEPLER models pre-calculated for this purpose, and find good agreement with a previous study. We performed experiments on simulated data, and recovered input parameters related to the burst ignition with reasonable accuracy, but less so for the system distance, emission anisotropy and neutron star mass and radius. Finally, we assembled a set of 14 daily burst epochs covering the 2024 outburst of the accretion-powered millisecond pulsar SRGA J144459.2-604207, and attempted to constrain the system properties of this object by comparing to SETTLE model predictions. We find reasonably good agreement between the observations and model predictions for a mildly sub-solar fuel composition, with H-fraction X about0.54 and CNO metallicity Z CNO about0.01. However, we note that the inferred H-fraction is in excess of the limit of 0.4 established separately, and the adopted model may not provide sufficiently accurate predictions for this burst ignition regime. The inferred distance depends on assumptions about the system inclination and corresponding anisotropy of the persistent emission, and is likely in the range 6-11 kpc. Future applications with more physically realistic models are a promising avenue for this and other sources with H-rich bursts.

Related papers