Steven Silverberg
LightNews LightNews
silverbergastro.bsky.social
Steven Silverberg
@silverbergastro.bsky.social
310 followers 1K following 82 posts
Postdoctoral Fellow, Smithsonian Astrophysical Observatory. Focus on low-mass stars and the material around them. PI of Disk Detective.
Posts Replies Media Videos
silverbergastro.bsky.social
We were able to reprocess and analyze archival Chandra & XMM-Newton observations of HL Tau from 2000, 2004, 2017, and 2018, and look at how it changes (or doesn't) over time. The temperatures stay consistent around 4 keV. The bright 2020 observations are not as bright as a 2004 flare. 9/14
Figure 9 from Silverberg et al. (2025). Comparison of model fits to HL Tau observations from 2000 to 2020. All models use one collisionally-excited plasma component with variable temperature and a fixed abundance for Fe (and elements with Fe-like first ionization potential) set at the best-fit value from a joint fit to the faint XMM-Newton observations from 2020 and a variable absorption component. Top row: flux (in units of pico-ergs per second per square centimeter) is shown as both absorbed (as observed; blue for XMM-Newton, green for Chandra) and unabsorbed (corrected for absorption; brown). Second row: the plasma temperature for the single plasma component in keV (red) in comparison to 4 keV (dashed gray line). Third row: the emission measure (in units of 10^52 per cubic centimeter) for the single plasma component (red). Bottom row: the hydrogen column density NH (in units of 10^22 square centimeters) of the absorbing material (black). Observing windows are labeled above each column. The observations during the flare identified by G. Giardino et al. (2006) in 2004 are highlighted in gray. The data show that HL Tau is variable, that the by-eye period in 2020 does not necessarily extend to earlier epochs (but might be there?), that the variation in 2020 is due to changing emission rather than changing absorption, and that this pattern does not necessarily hold at earlier epochs.
March 5, 2025 at 4:59 PM
silverbergastro.bsky.social
So, what all does this tell us? Our best fit models and the consistent presence of 6.7 keV emission indicate that the spectrum of HL Tau is hot (consistently ~4 keV), on the hot end and bright end of YSOs compared to the COUP observations, but generally consistent with being a Class I YSO. 8/14
Figure 8 from Silverberg et al. (2025). Temperature vs. unabsorbed X-ray luminosity for single-temperature fits to HL Tau (green crosses), in comparison to temperature and unabsorbed luminosities from two-temperature fits (blue and orange circles) to the COUP pre-main-sequence stars (T. Preibisch et al. 2005). The plot shows temperature (measured in keV) in log space on the y axis, as a function of log(X-ray luminosity) on the x axis. The plot depicts blue, orange, and green scatter plots. Blue points represent the temperatures and luminosities associated with cool plasma from COUP, while orange points represent temperatures and luminosities associated with hot plasma. The blue scatter points are flat in temperature as a function of log(luminosity), while the orange points increase linearly as a function of log(luminosity). HL Tau is plotted as green "x" shapes, which overlap with the high-luminosity end of the COUP data. The right y axis shows temperature in millions of Kelvin.
March 5, 2025 at 4:59 PM
silverbergastro.bsky.social
We also got a high-resolution spectrum using the High-Energy Transmission Grating on Chandra! These data show clear emission lines that help constrain the temperature and density of the plasma. We also see indications of cooler plasma hidden by the absorption, and a hint at iron fluorescence! 6/14
Figure 7. Combined grating data from all 11 Chandra/ACIS/HETG observations of HL Tau, with spectral lines of interest labeled. The best-fit model fit jointly to the individual zeroth-order and combined grating spectra is shown in orange. Left inset: a zoom-in on the region around the Fe xxv feature at 1.85 Å (6.7 keV) shows a clear detection of the Fe xxv line and a marginal detection of the 6.4 keV neutral Fe fluorescence line in ∼300 ks. Right inset: a zoom-in on the 6–7 Å region shows a strong detection of the f line of the Si xiii He-like triplet and a fainter but still present detection of the r line. Inset units are same as in the larger figure. The plot itself is of counts per kilosecond per angstrom as a function of wavelength in angstroms (on the lower x axis) and energy in keV (on the upper x axis). The observed data are blue, while the best fit model is in orange. A left inset shows a strong peak in Fe XXV emission at 1.85 angstroms, and marginal emission at 1.94 angstroms. The right inset shows strong Si XIV emission, and clear detections of the Si XIII r and f lines. The f line is stronger than the r line.
March 5, 2025 at 4:59 PM
silverbergastro.bsky.social
We did similar with the zeroth-order data from our Chandra gratings observations in 2018. The data were taken over 18 days, so we don't have sensitivity to a 21-day signal. There is some rise and fall in the data, but not nearly as clear as in the 2020 data. 5/14
Figure 6 from Silverberg et al. (2025). Left: the zeroth-order light curve from the 2018 Chandra observations. Right: the Lomb–Scargle periodogram for these data (solid) and its window function (dotted). The strong peak in the window function near 2.5 days is consistent with the orbital period of Chandra. The left plot shows count rate (in counts per second) as a function of time (in MJD) for Chandra---the count rates are plotted as blue scatter plots. There are brief gaps between the first nine observations, then a longer duration gap before the final two. The right plot shows power as a function of period in days. There are no clear peaks in this data.
March 5, 2025 at 4:59 PM
silverbergastro.bsky.social
We looked at the variation over time in 2020, and saw something that looked like it might be a period. A Lomb-Scargle periodogram of the data shows a peak signal at 21 days (albeit with other peaks at lower levels at other periods). So, variability on a 21-day timescale! 4/14
The high-energy (>2 keV) XMM/PN light curve of HL Tau from 2020, in chronological order (left) and phased to a ∼21-day period (right). Colors distinguish data from each observation and are the same for each observation in both panels. Count rates are derived from binning by 1 ks. Two plots, side by side. The first one plots a series of points as a function of time--the y axis is counts per second, while the x axis is time in modified Julian days. Data are color-coded chronologically as black, blue, orange, red, purple, green. On the right, the data are phase folded to a period of ~21 days. Y axis is counts per second, while the x axis is phase (in days). The data are at a low point in green and orange observations at phase -10, rise to their highest in the red data around phase -4.5, decrease slightly to the black data at phase 0 days, and then diminish to nearly the level of the green and orange points in the purple and blue points at phase +4.5 days. Each data set is a large clump of scatter plot in the appropriate color, with gaps of a few days to a week between observations. Figure 4 from Silverberg et al. (2025): Lomb–Scargle periodograms of the high-energy light curves for HL Tau from the XMM-Newton MOS1 (solid blue), MOS2 (solid orange), and PN (solid green) detectors and all three jointly (black). The Lomb–Scargle periodograms of the window functions (i.e., flat data with identical timing and uncertainties to the observed data) for the three detectors are depicted as fainter dotted lines with the same colors as those derived from data. The strongest peak of the data is at a minimum of the periodogram of the window function, indicating that this is not due solely to the observing window. The strongest peak of the window function, at ∼2 days, corresponds to XMM-Newton's orbital period. A plot showing Power (from 0 to 1) on the y axis, as a function of period (in days) on the x axis. Blue solid lines represent the periodogram of data from the MOS1 detector on XMM-Newton, while orange represent data from the MOS2 detector and green represent data from the PN detector. A black solid line is a joint fit to data from all three detectors. Thin dotted lines in the colors for each detector show the periodogram of the window function for each detector. The observed data show a peak at ~21 days, while the window function shows a local minimum near 21 days. This suggests that there is a signal of variability on a 21-day timescale.
March 5, 2025 at 4:59 PM
silverbergastro.bsky.social
In our paper, we began with monitoring data of HL Tau from the XMM-Newton satellite in 2020, and high-resolution spectroscopy from Chandra in 2018. The monitoring data show that the spectrum is consistently hot, and heavily absorbed below 2 keV. 3/14
A plot from Silverberg et al. (2025) shows XMM-Newton EPIC-PN spectra from each of the six ∼40 ks observations (represented by different colors as listed in the legend) in the the 2020 monitoring campaign. Spectra are binned by 15 counts per bin. The y axis is counts per second per kilo-electron volt (keV) on a logarithmic scale, while the x axis is energy (in keV) on a linear scale. Observations are plotted at scatter plots showing all data points for each spectrum. Spectra are color coded chronologically as black, blue, orange, green, purple, brown. All spectra show a clear drop-off of emission below ~2 keV, indicating very high absorption. The other spectra show clear decreasing emission as a function of energy at energies above 2 keV, with the exception of a broad emission line that appears in all spectra at 6.7 keV. While all spectra show the same decreasing pattern, some of them are brighter than others--the continuum for the black and green spectra are above the other four. Each point has associated uncertainties in plotted error bars.
March 5, 2025 at 4:59 PM
silverbergastro.bsky.social
You may recall HL Tau from such hits as "a notable infrared source" and "subject of absolutely beautiful ALMA disk images." It's also an excellent subject for observation in X-rays, which can get at the stellar surface in ways other wavelengths can't as easily (because of absorption). 2/14
A 2014 ALMA image of the Class I Young Stellar Object HL Tau. Image is an red-orange-yellow disk on a black field. The disk has alternating light and dark (black) rings, indicating gaps in the disk. The light gets brighter (more orange-yellow) as you look towards the center. It is canted at an angle. Image credit: NRAO/ESO/NAOJ
March 5, 2025 at 4:59 PM
silverbergastro.bsky.social
Counterpoint: that's not even the best Cardale tweet. Evidence:
An old screenshot of a tweet by Ohio State University quarterback Cardale Jones from October 5, 2012, which reads, "Why should we have to go to class if we came here to play FOOTBALL, we ain't come to play SCHOOL, classes are POINTLESS"
November 1, 2024 at 1:35 AM