TL;DR
This paper presents the highest-cadence, highest signal-to-noise millimeter light curves of Sagittarius A (Sgr A) to date, collected during the 2017 Event Horizon Telescope (EHT) campaign. These new data reveal detailed variability patterns on minute-to-hour timescales, crucial for understanding the physics near the supermassive black hole and for calibrating EHT observations.
What Problem Does This Paper Address?
Understanding the behavior of Sgr A* at millimeter wavelengths is key to probing the physics of its accretion flow and the immediate environment of the supermassive black hole. Previous millimeter light curves suffered from low cadence and signal-to-noise, limiting detailed variability studies and making it difficult to compare with other wavelengths (like X-ray and infrared) or to fully utilize the data for EHT calibration. This paper addresses this gap by providing unprecedentedly detailed millimeter light curves, enabling investigations into source variability on short timescales and improving the calibration of VLBI observations.
Methods
The authors leverage data from the Atacama Large Millimeter/submillimeter Array (ALMA) and the Submillimeter Array (SMA) collected simultaneously with the EHT in 2017. They employ specialized data reduction pipelines (A1 and A2 for ALMA, SM for SMA) to extract the compact source emission. Key techniques include:
- Intra-field calibration (A1): A novel method that models both the static extended emission (minispiral) and the variable compact source (Sgr A*) within the same field of view to remove instrumental and atmospheric effects.
- SEFD-based calibration (A2): A more standard approach using system equivalent flux density measurements.
- Gaussian Process (GP) modeling: Used to characterize the variability, specifically employing a Damped Random Walk (DRW) model and a more general Matérn covariance model to extract timescales and power spectral density (PSD) slopes.
- Structure Function (SF) analysis: Applied to identify characteristic variability timescales.
- Lomb-Scargle periodogram analysis: Used to search for periodic signals.
- Cross-correlation analysis: To investigate time lags between different frequency bands.
Datasets
The primary datasets are:
- ALMA observations: Collected on 2017 April 6, 7, and 11, with 4-10 hour tracks.
- SMA observations: Collected on 2017 April 5, 6, 7, 10, and 11, with 3.3-7.7 hour tracks.
- Archival millimeter light curve data: From various sources spanning 2005-2019, used for comparison.
Key Insights and Ideas
- Minute-scale variability is real: The high cadence data confirm that Sgr A* exhibits significant variability on timescales as short as one minute, challenging previous assumptions and requiring sophisticated modeling.
- The "flaring" day is different: The light curves from April 11, 2017, which followed an X-ray flare, show dramatically enhanced variability compared to other days, suggesting a change in the source's physical state.
- Red noise dominates, but with nuances: The variability is well-described by a red-noise process, but the analysis suggests a steepening of the power spectral density (PSD) slope at shorter timescales, potentially indicating multiple underlying processes.
- Calibration is a major challenge: The paper highlights that systematic calibration errors, rather than thermal noise, are the dominant source of uncertainty and differences between data reduction pipelines.
- Optically thin at the horizon: The lack of time lags between millimeter bands and a flat spectral energy distribution suggest that the emission region probed by these observations is optically thin, consistent with imaging the event horizon scale.
Key Results
- High-cadence light curves: The paper presents the most detailed millimeter light curves of Sgr A* to date, with cadences as short as 4 seconds.
- Variability characterization: The data are consistent with a red-noise process with a PSD slope between -2 and -3, steepening to approximately -2.6 on timescales shorter than 0.3 hours.
- April 11 anomaly: Variability on April 11, 2017, was significantly enhanced (factor of ~2-10 increase in variance across timescales) compared to other days, particularly after an X-ray flare.
- Spectral index: The average spectral index at 220 GHz is measured to be $0.0 \pm 0.1$, with rapid fluctuations and a lower value (-0.25±0.10) immediately after the X-ray flare.
- No time lags: No detectable time lags were found between the 213 and 229 GHz bands, indicating low optical depth.
- GP model comparison: The Matérn covariance model provides a better fit to the data than the DRW model, particularly in capturing the steepness of the PSD at short timescales.
Limitations and Future Work
- Calibration uncertainties: Systematic calibration errors remain a significant challenge, leading to differences between data reduction pipelines and influencing variability measurements. The authors recommend the A1 pipeline for future analyses due to its robustness.
- Timescale constraints: While timescales of ~0.14-0.30h and ~1.0-1.5h are identified, their significance is limited by the observation duration and sampling. Longer observations are needed for more robust characterization.
- Stationarity assumption: The assumption of a single, stationary stochastic process may not hold over long timescales, especially when combining data from different epochs or source states.
- Future work: The authors suggest that future EHT observations with similar or improved quality will further advance the understanding of Sgr A*'s statistical properties. The analysis of polarization data from this campaign will be presented elsewhere. The paper also emphasizes the need for improved calibration techniques and more sophisticated modeling to account for potential breaks in the PSD.
Summary updated