Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Identifying the physical periods in the radio emission from the $γ$-ray emitting binary LS I +61 303

Frédéric Jaron, Valentí Bosch-Ramon

TL;DR

This work tackles the question of which radio-periods in the $ ext{LS I +61 303}$ binary reflect intrinsic physical processes versus interference effects. By applying a generalized Lomb-Scargle timing analysis to carefully selected radio intervals and by generating synthetic datasets composed of sums of sinusoids at candidate periods, the authors demonstrate that the periods $P_1 = 26.5$ d, $P_2 = 26.9$ d, and $P_{ m long} = 4.6$ yr can all be intrinsic, with $P_3 \approx 26.3$ d arising as a beat-related side-lobe rather than a separate physical rhythm. The full OVRO and GBI radio datasets, along with interval-specific analyses and synthetic-data tests, show that long-term modulation can survive after filtering out short-term periods, supporting a scenario in which a long-term process modulates orbit-by-orbit emission. The results provide a model-agnostic constraint on the system's dynamics and underscore the need for continued, high-cadence multiwavelength monitoring to refine the physical interpretation of these periods.

Abstract

The $γ$-ray emitting binary LS I +61 303 exhibits periodic emission across the electromagnetic spectrum, from radio up to the very-high-energy regime. The most prominent features are the three periods $P_1 = 26.5$ d, $P_2 = 26.9$ d, and $P_{\rm long} = 4.6$ years. Occasionally, a fourth period of 26.7 d is also detected. Mathematically, these four periods are interrelated via the interference pattern of a beating. Competing scenarios that seek to determine which of these periods are physical and which are secondary are under debate. The detection of a fifth period, $P_3 = 26.3$ d, was recently claimed. Our aim is to determine which of these periods are intrinsic (likely related to physical processes) and which of these are secondary (resulting from interference). We avoided any assumption about the physical scenario and restricted our analysis to the phenomenology of the radio emission variability. We selected intervals from archival radio data and applied the generalized Lomb-Scargle periodogram. We fit the observational data to generate synthetic data that only contain specific signals. We analyzed these synthetic data to assess the impact of these signals and their interference on the light curves and the periodogram. The two-peaked profile, consisting of $P_1$ and $P_2$, was detected in the periodogram of the actual data for intervals that are significantly shorter than $P_{\rm long}$, provided that these intervals contain a minimum of the long-term modulation. The characteristics of the observational data and their periodogram could only be reproduced with synthetic data if these explicitly included all three periods $P_1$, $P_2$, and $P_{\rm long}$, the residuals being limited by noise. We have found that all three periods, i.e., $P_1$, $P_2$, and $P_{\rm long}$, could correspond to physically real processes occurring in LS I +61 303.

Identifying the physical periods in the radio emission from the $γ$-ray emitting binary LS I +61 303

TL;DR

This work tackles the question of which radio-periods in the binary reflect intrinsic physical processes versus interference effects. By applying a generalized Lomb-Scargle timing analysis to carefully selected radio intervals and by generating synthetic datasets composed of sums of sinusoids at candidate periods, the authors demonstrate that the periods d, d, and yr can all be intrinsic, with d arising as a beat-related side-lobe rather than a separate physical rhythm. The full OVRO and GBI radio datasets, along with interval-specific analyses and synthetic-data tests, show that long-term modulation can survive after filtering out short-term periods, supporting a scenario in which a long-term process modulates orbit-by-orbit emission. The results provide a model-agnostic constraint on the system's dynamics and underscore the need for continued, high-cadence multiwavelength monitoring to refine the physical interpretation of these periods.

Abstract

The -ray emitting binary LS I +61 303 exhibits periodic emission across the electromagnetic spectrum, from radio up to the very-high-energy regime. The most prominent features are the three periods d, d, and years. Occasionally, a fourth period of 26.7 d is also detected. Mathematically, these four periods are interrelated via the interference pattern of a beating. Competing scenarios that seek to determine which of these periods are physical and which are secondary are under debate. The detection of a fifth period, d, was recently claimed. Our aim is to determine which of these periods are intrinsic (likely related to physical processes) and which of these are secondary (resulting from interference). We avoided any assumption about the physical scenario and restricted our analysis to the phenomenology of the radio emission variability. We selected intervals from archival radio data and applied the generalized Lomb-Scargle periodogram. We fit the observational data to generate synthetic data that only contain specific signals. We analyzed these synthetic data to assess the impact of these signals and their interference on the light curves and the periodogram. The two-peaked profile, consisting of and , was detected in the periodogram of the actual data for intervals that are significantly shorter than , provided that these intervals contain a minimum of the long-term modulation. The characteristics of the observational data and their periodogram could only be reproduced with synthetic data if these explicitly included all three periods , , and , the residuals being limited by noise. We have found that all three periods, i.e., , , and , could correspond to physically real processes occurring in LS I +61 303.

Paper Structure

This paper contains 15 sections, 2 equations, 10 figures, 1 table.

Table of Contents

  1. Introduction
  2. The stellar system LS I +61 303
  3. The binary system
  4. Physical scenarios for periodic emission
  5. The data
  6. Methods
  7. Lomb-Scargle periodogram
  8. Period determination
  9. Results
  10. Timing-analysis of restricted intervals
  11. Interlude: Reasoning behind the Taylor1982 detection of only $P_1$
  12. Filtering out $P_1$ and $P_2$
  13. Synthetic data
  14. Conclusions
  15. Parameters from synthetic data fit

Figures (10)

  • Figure 1: Radio light curve resulting from long-term monitoring of LS I +61 303 by the GBI at 8 GHz. The lower time axis is expressed in terms of MJD and the upper one in cycles of the long-term modulation, using $P_{\rm long} = 1667$ d and $T_0 = \mathrm{MJD}~43366.275$. These data have been averaged in time bins of width seven days. Two intervals are highlighted (green and yellow), which are analyzed individually, as explained in the text.
  • Figure 2: Radio light curve of LS I +61 303 observed by the OVRO at 15 GHz. The two color-shaded intervals are analyzed separately.
  • Figure 3: GLS periodograms of the GBI 8 GHz data. The horizontal lines indicate FAP levels of 10 % (dotted), 1 % (dashed), and 0.1 % (solid). (a) Periodogram of the full light curve shown in Fig. \ref{['fig:GBI8GHz-lc']}. A peak at $P_{\rm long}$ is present along with a feature at $P_1$. (b) Zoom onto the region around the orbital period. A distinct two-peaked profile is present. The vertical dashed lines marks the position of the possible third period, $P_3$. The result of fitting the periodogram with Gaussian functions is shown as the red solid curve. We report the uncertainties of the resulting periods using the $1\sigma$ standard deviations of these Gaussians. (c) Periodogram of interval I (green area in Fig. \ref{['fig:GBI8GHz-lc']}). The peak at $P_{\rm long}$ is absent. (d) The zoom onto the orbital region reveals the presence of only one peak, at $P_{\rm average}$. (e) Periodogram of interval II (yellow area in Fig. \ref{['fig:GBI8GHz-lc']}). (f) Zoom onto the orbital region. There is a two-peaked profile, with the periods being largely compatible with the literature values of $P_1$ and $P_2$.
  • Figure 4: GLS periodograms of the OVRO 15 GHz data. (a) Periodogram of the entire light curve shown in Fig. \ref{['fig:OVRO15GHz-lc']}Jaron2024. (b) Zoom onto the region around the orbital period. A distinct two-peaked profile is present. The vertical dashed line marks the position of $P_3$, as reported by Zhang2024. A significant feature is indeed detected at this position. (c) Periodogram of interval I (green-shaded area in Fig. \ref{['fig:OVRO15GHz-lc']}). The only significant feature is at approximately the orbital period. (d) Zoom. There is only one peak. (e) Periodogram of interval II (yellow-shaded area in Fig. \ref{['fig:OVRO15GHz-lc']}). The only significant feature is at approximately the orbital period. (f) Zooming in on the orbital range reveals a two-peaked profile.
  • Figure 5: Radio observations of LS I +61 303 taken from Taylor1982. (a) Light curve at 5 and 10.5 GHz. An interval for separate analysis is shaded in green. (b) Periodogram of the full light curve. The most powerful peak is consistent with the orbital period. The fit to the periodogram, however, requires the superposition of two Gaussians. (c) Periodogram of super-orbital interval $\Theta = 0.0 - 0.5$ (green-shaded area in panel a). The one-peaked profile is shifted toward the middle of the positions of $P_1$ and $P_2$ and is fitted by a single Gaussian.
  • ...and 5 more figures