NenuFAR detects giant hydrogen and carbon atoms in front of the remnant of Supernova Cassiope A
Recombination lines (including the iconic Hα in the optical domain) allow us to study the diffuse phases of the interstellar medium (ISM). In the Bohr formalism, the electrons of an excited atom are in discrete orbits; the diameter of an excited atom is then given by d = n2 x 1.05 x 10-101m (n being the main quantum level of excitation).In the diffuse interstellar medium, the recombination of an electron with an ion produces atoms sometimes excited in extremely high quantum levels (n > 100). These are called Rydberg atoms. These atoms have consequently diameters of the order of a few microns, the typical size of a bacteria or a fly ash on earth. When these giant atoms de-excite from such levels, they radiate their energy in the low frequency radio domain (< 1 GHz). These are the radio recombination lines (RRL). We have used NenuFAR, the large low frequency radio telescope (10 MHz to 85 MHz) of the Nançay station, precursor of SKA to observe these lines. We have detected them very easily in front of the synchrotron continuum emission source of the Cas A supernova remnant (SNR), confirming and complementing recent LOFAR and earlier UTR-2results. The analysis of these new data, obtained in the framework of the Early Science Key Project ES10, (i) underlines the interest of RLLs (still little observed) for the study of the ISM, (ii) confirms the potential of NenuFAR in this field and (iii) shows the possible perspectives in the context of SKA.
Fig. 1 Figure adapted from Salas et al., (2018). The Cas A SNR was observed by NenuFAR in September 2019 in the 10-85 MHz frequency band. Data reduction and analysis was performed in spring 2021 by Lucie Cros. The results presented here correspond to a 2h sample of observations.
NenuFAR
NenuFAR (New extension in Nançay Upgrading LOFAR) is a radio-telescope located on the Nançay site in Sologne. At the time of the observations it consisted in 56 mini-arrays of 1064 dual polarization-sensitive antennas, gradually updated now to 80 mini-arrays and soon to 96 mini-arrays (1824 antennas) in 12 to 18 months. Its early science phase started mid 2019 and will last until the end of 2022. A dozen of Key Programmes are being observed, among which the ES10 is the only one dedicated to the observation of the Galactic ISM. Our program uses the remarkable capacities of NenuFAR in standalone beam-forming mode: a spectral resolution up to 95 Hz/channel over the whole bandwidth of the telescope (10-85 MHz), that is 786432 simultaneous spectral channels. In addition to these capacities, NenuFAR offers a standalone imager mode, not used for the results presented here. Complementary to this, first attempts to couple NenuFAR with LOFAR have been performed, and open the perspective for ES10 to carry out very high angular resolution observations of RRLs at low frequencies. High angular imaging at high spectral resolution will also be possible in the limit of the data flux that the instrument can handle (78 Go/sec). These achievements are the basis of the preparation of the commissioning of the Square Kilometer Array. This future low-frequency telescope will operate from the southern hemisphere and above 50 MHz, which warrants long-term complementarities with NenuFAR.
Observations
Within the framework of the Early Science ES10 project, the objective of these first observations was to determine precisely the capabilities of NenuFAR for the study of the Interstellar Medium via the observation of recombination lines (RRL) in front of an emblematic source. We chose to point NenuFAR towards Cas A, whose continuum level is particularly high, which facilitates the detection of the lines in absorption. The results show that the instrument is up to its specifications: with 2h of observation, NenuFAR and its time-frequency receiver ‘UnDySPuTeD’ (developed at LPC2E) reach a signal-to-noise level of the order of SNR~3-10 on individual lines and of the order of 7 times higher (SNR~20-70) when the stacking method is used on samples of 50 lines.
Fig. 2 Left: individual Carbon Cα recombination lines at quantum levels n=690 and n=691, detected in absorption by NenuFAR. Right: after stacking, the increase of the signal-to-noise ratio allows to resolve different velocity structures in the line. Fitting Voigt profiles on each of them thus allows us to identify ISM clouds at different radial velocities. The spectral resolution is 190 Hz (~2 km/s).
Data Analysis
The data analysis required the implementation of a pipeline optimized for the detection of weak (absorption depth ~10-3) and narrow (a few km/s) absorption lines. A calibration of the frequency response with good accuracy was necessary to correctly estimate the continuum level over the whole observed band and thus to determine the depths and line-widths of the different absorption lines. For this purpose, it was necessary to (i) remove the very large number of interferences (RFI) which pollute the low radio frequencies, (ii) correct the signal fluctuations with frequency at small scales (in each of the 192 sub-bands) and (iii) calibrate the absolute flux as a function of frequency at large scales from the power spectrum of the theoretical emission expected for Case A, thus allowing an estimate of the noise level reached in agreement with the predictions of the NenuFAR time estimator (https://github.com/AlanLoh/nenupy).
Results : Absorption line profiles
Figure 2 on the left shows two individual absorption lines at n=690 and n=691. It can be seen that the continuum level has been removed; it is therefore a negative signal that we are trying to measure. On the right of this figure it is the stacked signal of several quantum levels around n=599 that is illustrated. We have also overlaid absorption models to account for the contribution of the three gas clouds located along the line of sight. These well known clouds in the Perseus arm are at velocities of -47 km/s, -38 km/s and 0 km/s. What we seek to determine then, is the width of these absorption lines order to constrain the properties of the absorbing Interstellar Medium (electron density and temperature). For instance, for the first component at -47 km/s it is necessary to determine precisely the broadening in km/s (or equivalently in Hz) of the green profile of Figure 2. This broadening depends of course on the contribution of the other components.
The line width, a valuable indicator
Fig. 3 Comparison of results obtained with NenuFAR and with LOFAR (Salas et al., 2017, Oonk et al., 2017). The measurement points represent the total widths of the Voigt profiles in kHz as a function of quantum number (average n of the stacked sample). The solid color curves represent individual model predictions for different local properties of the interstellar medium. The dashed curve is the sum of the three contributions.
Why determine line widths? The broadening of the recombination lines depends on three different physical mechanisms: (i) a Doppler broadening related to the thermal agitation and turbulence of the medium, (ii) a radiative broadening, related to the interaction of the Carbon atoms with an external radiation field and (iii) a pressure broadening (also called collisional broadening), related to the interactions with the electrons of the medium. If the Doppler broadening is well represented by a Gaussian, the other mechanisms affect the wings of the line and require a Lorentzian profile. These two mathematical functions combined correspond to what is called a Voigt profile.
For small quantum numbers (around 500), ie for smaller atoms, the Doppler broadening dominates and we expect the other contributions to be negligible. But when the atoms are larger (beyond n = 500), then the Lorentzian line wings contribute to the bulk of the broadening. If we measure the line profiles for different quantum numbers, it is then possible to observe this transition and to determine the properties of the interstellar medium (thermal and turbulent velocity at low n; temperature of the external radiation; density and temperature of electrons at high n). This is represented in Figure 3, which also shows the agreement between the NenuFAR data and those obtained with LOFAR over the whole low frequency range covered.
References
Cros L. et al., in prep.
Gordon M. A. et Sorochenko, R. L., 1992, SSRv, 59, 412G
Konovalenko, A. A. et Sodin, L. G., 1981, Nature, 294, 5837
Oonk R. et al. 2017, MNRAS, 465, 1066O
Salas P. et al., 2018, MNRAS, 475, 2496S
Salas P. et al., 2017, MNRAS, 467; 2274S
Salgado, F. et al., 2017, ApJ, 837; 142
NenuFAR (https://nenufar.obs-nancay.fr/, https://nenufar2019.sciencesconf.org/, https://nenufaruser2021.sciencesconf.org/)
LOFAR (https://www.astron.nl/telescopes/lofar/)
UTR-2 (https://en.wikipedia.org/wiki/Ukrainian_T-shaped_Radio_telescope,_second_modification)