BSN Application in Astrophysics

• BSN application is a specialized software platform for analyzing photometric light curves of contact binary systems, specifically W UMa-type eclipsing binaries.
• It integrates data preprocessing, Roche geometry modeling, and both Levenberg–Marquardt and MCMC optimization for robust parameter estimation.
• Empirical evaluations reveal high computational efficiency and precision, with applications across TESS and ground-based multiband observations.

A BSN application (Binary Systems of South and North application) in contemporary astrophysics refers to a specialized software platform for photometric light-curve modeling and physical parameter estimation of contact binary star systems, with a particular focus on the W UMa-type eclipsing binaries. The most current and widely referenced implementation is the “BSN application” version 1.0 described and deployed in several recent studies for rapid, robust analysis of multiband time-series data from both ground-based and space-based observatories (Paki et al., 15 Aug 2025, Poro et al., 20 Oct 2025).

1. Software Architecture and Workflow

The BSN application is a modular, Windows-based program written in .NET (C#), optimized for the analysis of contact binary light curves. The main dataflow consists of:

• Data acquisition: Supports ASCII, CSV, and FITS-formatted light-curve data, including SAP_FLUX from TESS, with multiple photometric filters (Johnson UBVRI, Cousins Rc, Ic, TESS).
• Preprocessing: Conversion of time to orbital phase using provided ephemerides or minima lists; normalization; optional detrending.
• Geometric modeling: Computation of Roche equipotential surfaces according to Mochnacki (1984), with user-switchable contact configuration and calculation of the fill-out factor,

$$f = \frac{\Omega_\text{in} - \Omega}{\Omega_\text{in} - \Omega_\text{out}}$$

where $\Omega_\text{in}$ and $\Omega_\text{out}$ are Roche potentials at the inner and outer Lagrange points (Poro et al., 20 Oct 2025).

• Surface discretization and flux synthesis: Calculation of monochromatic intensity for each surface element, incorporating limb-darkening, gravity-darkening (default exponent $\beta=0.32$ for convective envelopes), and bolometric reflection effects. Synthetic light curves are generated via integration over the visible surface at all phases and bands.
• Parameter optimization: Offers both Levenberg–Marquardt differential-correction (fast local minimization) and ensemble Markov-Chain Monte Carlo (MCMC) for robust posterior estimation. The likelihood function is based on

$$\chi^2(\mathbf{p}) = \sum_{i=1}^N \Bigl[\frac{f_{\rm obs,i} - f_{\rm mod,i}(\mathbf{p})}{\sigma_i}\Bigr]^2$$

with flexible parameterization (Paki et al., 15 Aug 2025).

• Output: Full posterior parameter chains, summary tables (.csv/.txt), phase-sorted synthetic light curves, and diagnostic corner or residual plots.

2. Physical and Mathematical Models

BSN’s modeling framework is rooted in Roche binary geometry and established prescriptions for stellar surface physics:

• Roche geometry forms the basis for star shapes, with explicit calculation of the equipotential for overcontact binaries.
• Limb-darkening is supported via the linear and logarithmic laws; user-provided coefficients or van Hamme (1993) tables. Surface intensities per band are computed as

$$I(\mu, \lambda) = I_0(\lambda) \left[1 - x_1(1-\mu) - x_2\,\mu\ln\mu \right] g^\beta$$

where $\mu$ is the cosine of the angle to the local surface normal and $g$ is surface gravity.

• Gravity-darkening and albedo: Default $\beta=0.32$ (Lucy 1967) for convective stars; albedo $A=0.5$ (Ruciński 1969).
• Starspot modeling: Circular spots are parameterized by angular radius, co-latitude, longitude, and temperature ratio. Inside the spot, the local temperature is modified, which directly impacts both intensity and limb-darkening at that location (Poro et al., 20 Oct 2025, Paki et al., 15 Aug 2025).

Derived absolute parameters are obtained using empirical period–semimajor axis ($P$–$a$) relations,

$$a = (0.372_{-0.114}^{+0.113}) + (5.914_{-0.298}^{+0.272}) P\,,$$

where $a$ is in $R_\odot$ and $P$ in days (Poro et al., 20 Oct 2025). Masses, radii, and luminosities follow from standard formulae:

$$M_1 = \frac{4\pi^2 a^3}{G P^2 (1+q)}, \quad M_2 = q M_1$$

$$R_{1,2} = a\,r_{\rm mean\,1,2},\qquad L = 4\pi R^2 \sigma T^4$$

3. Optimization, Inference, and Performance

The central inference mechanism is an affine-invariant ensemble MCMC algorithm (single stretch move, Goodman & Weare 2010), enabling robust recovery of posterior distributions for geometric and physical parameters including inclination $i$, mass ratio $q$, fill-out factor $f$, and polar temperatures $T_1$ and $T_2$ (Paki et al., 15 Aug 2025). The software achieves high computational efficiency:

• Synthetic light curve generation is accelerated by >40$\times$ relative to PHOEBE 2.4.9, via compiled .NET code, multi-threading, and caching of geometric and atmosphere computations.
• The optimization can routinely handle multiband datasets and hundreds to thousands of phase points per system.

Convergence and uncertainty reporting are based on standard MCMC diagnostics (autocorrelation, acceptance fraction, marginal quantiles).

4. Multiband, Starspot, and Advanced Features

Key capabilities of the BSN application as used in (Poro et al., 20 Oct 2025) and (Paki et al., 15 Aug 2025):

• Native multiband fitting: Simultaneous adjustment of model across arbitrary bands, accounting for band-dependent limb-darkening and zero points, but with shared geometric parameters.
• Starspot implementation: Starspot parameters may be fixed or allowed to vary within the MCMC; the model includes the effect on both local intensity and limb-darkening. Spot models are critical to replicate O’Connell-effect asymmetries commonly seen in W UMa-type systems (Poro et al., 20 Oct 2025).
• User interface: Graphical display of model and data, 3D visualization of the binary configuration at any phase, and pipeline for exporting results for further astrophysical analysis.

5. Empirical Evaluation and Application

The BSN application has been validated on a wide range of real systems—from total eclipse binaries using TESS data (Paki et al., 15 Aug 2025) to multiband solutions of ground-based observations (Poro et al., 20 Oct 2025):

System Type	Dataset	Model Features Used	Parameter Recovery	Key Results
TESS contact	TESS SAP_FLUX	MCMC, starspots	$q$ uncertainty $<$0.01	Absolute param. match empirical laws
Multiband W UMa	B, V, $R_c$, $I_c$	Starspots, multiband	$<$10\% error on $M$, $R$, $L$	90% of systems: $\Delta T < 9.4\%$
Shallow fillout binaries	Ground-based	Multiband, MCMC	Robust spot modeling	Fillout factor: $f=0.08\!-\!0.22$

BSN’s template and parameter conventions follow established codes (Wilson–Devinney, PHOEBE), ensuring interoperability and reproducibility.

6. Scientific and Astrophysical Applications

BSN’s primary domain is empirical and evolutionary study of contact binaries:

• Accurate classification into A-/W-subtypes based on $M_{1,2}$ and $T_{1,2}$.
• Quantitative assessment of thermal contact via analysis of $|\Delta T/T|$ for two stars—finding $<9.4\%$ temperature contrast in 90% of cases (Poro et al., 20 Oct 2025).
• Long-term evolutionary status assessed using period changes ($O-C$ diagrams), mass transfer rates, and positions in mass-radius and mass-luminosity planes.
• Estimation of initial component masses and total mass loss from absolute parameters.
• Investigation of dynamical stability, evidence for tertiary components, and comparison with empirical and theoretical angular momentum relations.

7. Future Directions and Limitations

Limitations of the current (v1.0) BSN implementation include:

• Platform restriction to Windows; no native support for macOS/Linux as of the described studies.
• Atmosphere model grid limited to $T_{\rm eff}=3500$–$8500$ K and solar metallicity.
• Spot models are fixed-shape, no spot evolution or migration.
• Third-light treatment available, but not integrated into full posterior sampling.

Planned advancements include cross-platform compatibility, expanded atmosphere and filter support, adaptive/multispot modeling, joint photometric+spectroscopic analysis, and community distribution of v2.0 with scripting interfaces (Paki et al., 15 Aug 2025).

---

References:

• "The BSN Application-I: Photometric Light Curve Solutions of Contact Binary Systems" (Paki et al., 15 Aug 2025)
• "BSN-IV: The First Multiband Light Curve Study of Five W UMa-type Contact Binary Systems" (Poro et al., 20 Oct 2025)