EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 678 (October 2023) A&A, 678 (2023) A42 Full HTML
Open Access
Issue
A&A
Volume 678, October 2023
Article Number A42
Number of page(s) 12
Section Extragalactic astronomy
DOI https://doi.org/10.1051/0004-6361/202347117
Published online 29 September 2023

© The Authors 2023

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe to Open model. Subscribe to A&A to support open access publication.

1. Introduction

Even in the harsh central regions of galaxies hosting an active galactic nucleus (AGN), observations of the cold gas (in the form of atomic neutral hydrogen and cold molecular gas) allow us to trace a variety of phenomena (Morganti & Oosterloo 2018; Veilleux et al. 2020; Combes 2022). Not only can cold gas be present in circum-nuclear disc (CND) structures with sizes of a few hundred parsec, it also traces infalling gas associated with the feeding of the supermassive black hole (SMBH; Tremblay et al. 2016; Maccagni et al. 2018; Rose et al. 2019; Combes 2022) and gas that is outflowing under the effect of winds and radiation (Brusa et al. 2018; Veilleux et al. 2020, and references therein) or under the effect of the mechanical impact from radio jets (Oosterloo et al. 2017; Murthy et al. 2022; Girdhar et al. 2022; Audibert et al. 2023). One of the advantages of tracing this cold gas is the possibility of observing it at the very high spatial resolution allowed by radio telescopes. The gas can thus be traced in the circum-nuclear region, very close to the SMBH and in the region that is co-spatial with the inner parts of the jets.

The central regions of galaxy clusters, and in particular, cool-core galaxy clusters, are examples of the complicated symbiosis between AGN and the gaseous intra-cluster medium that includes the cold gas. This makes them ideal targets for the study of all the phenomena listed above. The vast majority of these clusters host a radio galaxy in their centre. The radio plasma emitted by these galaxies in the form of jets and lobes is thought to play a key role in preventing the catastrophic cooling of the hot cluster medium. Nevertheless, many of these clusters do contain cold gas. Cold molecular gas mostly traced by CO lines has been observed in many cases (see e.g. Tremblay et al. 2016; Russell et al. 2019; Rose et al. 2023). ALMA observations have shown that the molecular gas in cool-core clusters is often distributed in dynamically unstable filamentary structures associated with X-ray bubbles inflated by radio lobes and resulting from the cooling induced by this interaction (see Lim et al. 2008; Russell et al. 2019). So far, only a minority of cases appear to have CO in circum-nuclear discs, such as Hydra A and 3C 84 (Rose et al. 2019; Nagai et al. 2019 respectively), with the latter being the target of our study.

Atomic neutral hydrogen (H I) is also found to be present in radio galaxies hosted at the centre of these clusters via the detection of the 21 cm line in absorption (see Morganti & Oosterloo 2018 for an overview). The studies of this particular topic have not been extensive so far, however. A search for H I absorption in a small sample of bright central galaxies (BCGs) in clusters by Hogan (2014) found a few detections in which the absorption was either at the systemic velocity or mildly redshifted. A number of studies focusing on individual radio sources in clusters were conducted, for example PKS 2322–123 in Abell 2597 (O’Dea et al. 1994; Taylor et al. 1999), 3C 84 in the Perseus cluster (Sijbring et al. 1989; Jaffe 1990), 4C 26.42 in Abell 1795 (van Bemmel et al. 2012), 4C 35.06 in Abell 407 (Shulevski et al. 2015), Hydra A (Taylor 1996), Cygnus A (Conway & Blanco 1995; Struve & Conway 2010), and the Norma cluster (Saraf et al. 2023). Thus, radio galaxies in cool-core clusters are interesting objects in which several sometimes competing processes can be studied using the combination of CO in emission and H I absorption (see e.g. Tremblay et al. 2016).

Here we present new observations that trace the H I in absorption in 3C 84 (Perseus A). This is one of the most famous radio galaxies and is located in the centre of the iconic gas-rich cool-core Perseus cluster. The observations cover parsec to kiloparsec scales. We connect the results with the properties of the molecular gas and of the radio continuum. As described in Sect. 2, molecular gas in this source has been detected in the form of filaments and a CND. Interaction between the jet and surrounding gas clouds has been suggested based on the VLBI continuum structure of the radio jet (Nagai et al. 2017; Kino et al. 2018, 2021). With the H I absorption presented here, we investigate the interplay between these structures.

This paper is structured in the following way. In Sect. 2 we briefly summarise some of the relevant results from the extensive literature of this well-known source to set the context for our study. In Sect. 3 we describe new observations we obtained with the Karl G. Jansky Very Large Array (JVLA) and with the Very Long Baseline Array (VLBA). In Sect. 4 we present the results on the radio continuum and the H I on arcsec and milliarcsec scales. In Sect. 5 we combine these results and connect them with those obtained on the molecular gas from ALMA observations (Nagai et al. 2019; Oosterloo et al. 2023) in order to derive a picture of the distribution of H I. We conclude by discussing the implications of our results.

2. 3C 84 and its host galaxy

The radio galaxy 3C 84 is associated with the BCG NGC 1275 in the Perseus cluster1. It has an intermediate radio power (P1.4 GHz ∼ 1025 W Hz−1; Pedlar et al. 1990; Fabian et al. 2011), and the radio continuum emission covers a large set of scales, each showing interesting properties. The morphology of the radio continuum emission is shown (in contours) in Fig. 1, ranging from kiloparsec to parsec scales and highlighting the features relevant for the present study. The image on the left shows the large-scale structure (about 80 arcsec corresponding to about 29 kpc), which appears relatively symmetric. This symmetry is lost at arcsecond resolution (second panel from the left, from Pedlar et al. 1990), where the southern jet is clearly brighter. The bright core also shows a complex structure, as is visible in the two right panels of Fig. 1. A combination of diffuse emission (detected by Silver et al. 1998 and called a mini-halo) and a bright jet to the south are present. These structures on the sub-arcsecond scale are particularly relevant for our study.

thumbnail Fig. 1.

Images of 3C 84 system from the kiloparsec to parsec scales relevant for this study. Left: continuum map at 90 cm (Pedlar et al. 1990; angular resolution of ∼5″ ≃ 1.8 kpc) in grey contours superimposed on the map of molecular gas (Lim et al. 2008; resolution 3″ ≡ 1.1 kpc, black contours), with the X-ray image from Fabian et al. (2011) in the background. Second panel from the left: image at ∼1″ ≡ 360 pc resolution at 1380 MHz from Pedlar et al. (1990) showing the continuum structure against which the H I is seen in absorption. Last two panels on the right: were obtained using the VLBA at 330 MHz and 1.4 GHz by Silver et al. (1998) and show, in addition to the core-jet structure (right), a diffuse, extended structure (which they call mini-halo) on larger scales (∼200 pc ≡ 0 . ″ 5 $ 0{{\overset{\prime\prime}{.}}}5 $ radius).

The inner region of the 3C 84 radio jet has been studied in detail at very high spatial resolution (Giovannini et al. 2018). Monitoring of the parsec-scale region has revealed the continuous emergence of new components in the jet (Nagai et al. 2010; Suzuki et al. 2012; Savolainen et al. 2023). A sharp bending of the jet has also been observed (Nagai et al. 2017; Kino et al. 2018, 2021). Nagai et al. (2017) reported an abrupt change in the position of the hotspot of the innermost jet, as well as an increase in polarised emission on parsec scales that they interpreted as showing that the jet interacts with the surrounding inhomogeneous and clumpy medium. However, the change in the position angle of the jet was also explained by Suzuki et al. (2012) as caused by the difference in the path followed by newly ejected components from the path of the previously ejected components. If we were able to identify whether gas is present with disturbed kinematics that trace this interaction, we would be able to distinguish between these two scenarios and shed light on the inner structure of this complex radio galaxy. This is one of the goals of our study.

As expected in a cool-core cluster, its central region is gas rich. 3C 84 is indeed known to be embedded in ionised and molecular gas as well as H I. Ionised gas is seen in the form of a complicated web of large-scale Hα filaments out to 50 kpc (Conselice et al. 2001; Fabian et al. 2008) and is traced in the central regions by the detection of free-free absorption (Walker et al. 2000) and through [FeII] and Brγ lines from the CND (Scharwächter et al. 2013; Riffel et al. 2020). Molecular gas is also seen in large-scale filaments out to large radii (Salomé et al. 2006), as well as on much smaller arcsecond and sub-arcsecond scales in the central regions (Lim et al. 2008; Scharwächter et al. 2013; Nagai et al. 2019). Lim et al. (2008) used the Submillimeter Array (SMA) and detected filaments in CO(2−1) on kiloparsec scales, with kinematics consistent with radial infalling motion and without a clear signature of rotation. Interestingly, at the higher spatial resolution ( 0 . ″ 1 × 0 . ″ 05 ≡ 36 × 18 $ 0{{\overset{\prime\prime}{.}}}1 \times 0{{\overset{\prime\prime}{.}}}05 \equiv 36 \times 18 $ pc) obtainable with ALMA, Nagai et al. (2019) detected CO(2−1), HCN(3−2) and HCO+(3−2) emission distributed in a CND-like structure with a diameter of 100 pc ( 0 . ″ 3 $ 0{{\overset{\prime\prime}{.}}}3 $), with a mass of cold H2 of 4 × 108 M⊙. This CND is also detected in warm H2 (Wilman et al. 2005; Scharwächter et al. 2013) and could provide a reservoir of gas to feed the SMBH. This CND consists of a fast Keplerian rotating inner part with a diameter of about 50 pc and position angle 68° that is surrounded by a warped disc-like structure of gas (about 100 pc in diameter), the kinematics of which is less regular because large filaments of gas accrete onto the CND (see Oosterloo et al. 2023 for more details).

Unsurprisingly, H I has also been detected in the form of absorption against 3C 84. The presence of this H I has been known since the low-resolution (kiloparsec scale) observations by Crane et al. (1982), Sijbring et al. (1989), and Jaffe (1990). They found two absorbing systems: a narrow, highly redshifted system at ∼8100 km s−1 (the so-called high-velocity system), and a broad, so-called low-velocity, system at the systemic velocity. This is illustrated in the full H I profile in Fig. 2 from the JVLA observations presented in this paper.

thumbnail Fig. 2.

Full H I profile obtained with the JVLA showing the various absorption systems. We focus on the low-velocity system, centred on the systemic velocity of NGC 1275. The high-velocity absorption is briefly discussed in the appendix. Top: spectrum using the full intensity range. Bottom: same as above, but with a different vertical scaling. The newly discovered component is indicated with an arrow.

The high-velocity absorption is a very high opacity system that has been studied in great detail at high spatial resolution with the VLBA by Momjian et al. (2002). This component is thought to be associated with a foreground cluster galaxy that is falling towards NGC 1275 but is still ∼100 kpc distant from the AGN host galaxy (Gillmon et al. 2004; Sanders & Fabian 2007).

The broader low-velocity system, on the other hand, has not been studied as extensively because the strong continuum emission of 3C 84 requires high-quality bandpass calibration. Given the complexity of the radio continuum of 3C 84, observations spanning a range of spatial resolutions are required to give us a better picture of the location and conditions of the H I in the heart of Perseus A. This is the focus of our study. We used 15″ scales (i.e. ∼5 kpc) with the observations performed during the commissioning of the Multi-Frequency Front End of the Westerbork Synthesis Radio Telescope (WSRT) in April 2001 (Morganti et al. 2020), at arcsecond scales (i.e. a few hundred parsecc) with the JVLA and at milliarcsecond scales (i.e. a few dozen parsec) using the VLBA.

3. Observations

3.1. JVLA observations

These observations were carried out with the JVLA A array in August 2019 (Project ID: VLA/2018-06-101) over three epochs of four hours each. We spent a total of nine hours on the target with interleaved scans on 3C 147 every 40 min that we used for flux, phase, and bandpass calibration. Since at the resolution of our observation the target is very bright with a peak continuum flux density of 14.3 Jy beam−1, we spent a total of three hours on the calibrator, which has a total flux of about 22 Jy, in order to reach a good S/N for the bandpass calibration. The observational setup consisted of nine spectral windows, roughly covering the frequency range of 1−2 GHz. Eight of these were dedicated to continuum imaging with a bandwidth of 128 MHz subdivided into 64 frequency channels for each window. The remaining spectral window was used to study the H I and had a bandwidth of 32 MHz, centred at 1393.7 MHz (i.e. close to the absorption frequency of the low-velocity system of 3C 84), divided into 2560 frequency channels, giving a velocity resolution of 2.7 km s−1. The eight continuum spectral windows were found to be strongly affected by radio frequency interference (RFI), and much of the data could not be used. Therefore, all our results are based on the spectral-line window alone, upon which the impact of RFI was minimal.

The data reduction was made using the classic Astronomical Image Processing Software (AIPS; Greisen 2003). We used 3C 147 to estimate the antenna-dependent gain solutions. Since 3C 147 is slightly resolved with the A array, we imaged the calibrator and used the model for bandpass calibration. Since 3C 147 is only 1.5 times brighter than the target, we followed the strategy used by Oosterloo & Morganti (2005) for the bandpass calibration of a very bright source to reduce the impact of the propagation of bandpass errors into the data cube. This entailed Hanning-smoothing the bandpass solutions by eight channels and then applying these solutions to the unsmoothed 3C 84 data. Next, we carried out self-calibration using only the line-free channels and performed continuum subtraction and de-redshifting of the uv data to the systemic velocity of 3C 84 using z = 0.01756. A detailed description of these steps can be found in Murthy et al. (2021).

We made the continuum images and data cubes at two spatial resolutions: (a) one cube using natural weighting to recover the diffuse continuum emission and obtain the highest sensitivity. We made this cube at the raw resolution of 2.7 km s−1. Then, we used (b) a cube using robust = −1 to obtain a better resolution. This latter cube was made at a velocity resolution of 21.6 km s−1. The continuum, the spectral beam sizes, and the rms noise levels we achieved are listed in Table 1. The naturally weighted continuum image is shown in Fig. 3, and the integrated spectrum obtained from the naturally weighted cube is shown in Fig. 4. Although the sensitivity of the higher-resolution cube is lower, the H I absorption is slightly resolved at this spatial scale, and we use this to compare the kinematics of H I with that of molecular gas in Sect. 5.2.

thumbnail Fig. 3.

Natural weighted JVLA continuum image. The image has an rms noise of ∼0.4 mJy beam−1 and a beam size of 1.64″ × 1.52″ ≡ 585 pc × 543 pc with a position angle of 68 . ° 9 $ 68{{\overset{\circ}{.}}}9 $, shown in the bottom right corner. The contours start at 3σ and increase with factors of 2 $ \sqrt{2} $; the 3σ negative contours are shown in grey. The image shows a range of structures starting from the bright unresolved core, which has a peak flux density of 14.3 Jy beam−1, the southern jet, diffuse lobes, and also diffuse radio emission connecting all these structures.

thumbnail Fig. 4.

Comparison between the H I spectra of the low-velocity absorption system obtained with the JVLA and the WSRT. The WSRT spectrum has a velocity resolution of ∼34 km s−1, and the JVLA spectrum, originally at a velocity resolution of 2.7 km s−1, has been binned to roughly the WSRT channel resolution, followed by a Hanning smoothing. It thus has the same velocity resolution as the WSRT spectrum. Despite the difference in spatial resolution between the two observations (∼5 kpc for the WSRT and ∼350 pc for the JVLA), the strong similarity between the two profiles indicates that the H I producing the absorption and/or the background continuum is located on sub-kiloparsec scales from the centre (see text for details). Velocities are relative to systemic.

Table 1.

Observations and imaging results.

where Tspin is the H I spin temperature, gives a H I column density of 2.4 ± 0.1 × 1020 cm−2 for Tspin = 100 K. However, we show in Sect. 5.1 that the absorption does not arise against the bright inner core, but against a much fainter extended continuum component, implying much higher column densities.

4.2. Milliarcsecond scale

The VLBA continuum image is shown in Fig. 5. Following the designations of Silver et al. (1998), we detect the core (C), the bright southern jet (S1), a diffuse extension from the jet (S2), and the diffuse lobe (S3) about ∼80 mas (285 pc) farther south from the core (see Fig. 1, right). The peak emission is 3.5 Jy beam−1, coincident with S1, while the core at the resolution of our observations is ∼1 Jy. We measure an integrated flux of 9.1 ± 0.9 Jy. Silver et al. (1998) reported a total flux of 22.1 Jy at 1.3 GHz, ∼2.4 times brighter than our measurement. The flux of 3C 84 has varied strongly in recent years (see Paraschos et al. 2023, although their data cover higher frequencies than ours). This could explain the difference between the flux we measured and that of previous VLBA observations at 1.4 GHz (Silver et al. 1998; Momjian et al. 2002).

thumbnail Fig. 5.

Naturally weighted VLBA continuum map. The map has an rms noise of 2.1 mJy beam−1 and a beam of 9.4 mas × 6.7 mas (3.4 pc × 2.4 pc), with a position angle of − 28 . ° 5 $ -28{{\overset{\circ}{.}}}5 $. The contours start at 3σ and increase with a factor of 2 $ \sqrt{2} $. The 3σ negative contours are shown in grey. We detect the core (C), the southern bright jet (S1), its diffuse extension (S2), and the diffuse lobe farther south of the core (S3). We measure a peak flux density of 3.5 Jy beam−1 at S1 and an integrated flux of 9.14 Jy (which includes C+S1+S2+S3).

3.2. VLBA observations

The VLBA observations were carried out in March 2019 (Project ID: VLBA/19A-043). No fringes were found to the Pie Town station, and all the baselines to the Mauna Kea station produced strong ripples in the final data cube and thus were flagged. We observed both left and right circular polarisation, but we did not produce cross-polarisation products as we were only interested in the total intensity. We used a 32 MHz bandwidth, subdivided into 1024 channels, giving a velocity resolution of 6.7 km s−1, centred at the redshifted frequency of the broad H I component. The total observation time was 12 h. Eight hours were spent on 3C 84, and the remaining time was spent on interleaved scans on calibrator sources. We used 3C 454.3 and DA 193 as fringe finders and B0307+380 (J0310+3814) as the phase-reference calibrator.

The data reduction was made following the standard VLBI data reduction procedure using classic AIPS and DIFMAP (Shepherd 1997). In AIPS we first flagged the data with low weights, corrected for Earth-orientation parameters, applied the ionospheric corrections, corrected for sampler threshold errors from the correlator, and estimated the sampler and gain corrections using the standard AIPS tasks. Although we observed various calibrators, precision astrometry was not a main goal of the project, and therefore, we solved for residual fringe rate, delay, and phase solutions directly using 3C 84 data (i.e. fringe-fitting was made on the target source).

We used 3C 454.3 for the bandpass calibration. Since 3C 84 is brighter than 3C 454.3 and the latter is also extended at VLBI scales, we again used the same technique as with the JVLA data mentioned in Sect. 3.1 to estimate the bandpass solutions. The calibrated visibilities were then exported to DIFMAP for a few rounds of self-calibration and imaging. This is a powerful and quick way for iterative imaging and self-calibration cycles with additional flagging of data that are poorly (and irreversibly) calibrated or had various issues (instrumental or due to RFI). For the initial steps, we started with uniform weighting, followed by a weighting scheme that balanced resolution and sensitivity (uvw = 2, 0 and uvw = 2, −1 in DIFMAP, respectively). For the first several rounds, only the phases were calibrated, initially with a 30-min solution interval, and gradually going down to 3 min. The remaining calibration errors after applying a priori gains (from measured Tsys and gains) and the loss of Pie Town meant that it was challenging to recover the most extended emission. Before amplitude and phase self-calibration on shorter timescales, finally down to 1 min, we recovered some of the extended structure by fitting large circular Gaussian components to the residual data, and we continued cleaning with uvw = 0, −1 and finally uvw = 0, −2 (natural weighting). While this process helped us identify sections of bad data to flag and resulted in a good calibration, the image is still limited by the dynamic range. Some spurious cleaning errors were noted as well. Therefore, we imported the calibrated data to AIPS, which was used to make the final continuum image and the cube. The subsequent steps in the making of the cube are the same as for the JVLA data in Sect. 3.1. The final parameters of the images and H I cube are listed in Table 1.

4. Results

4.1. Arcsecond scale

The JVLA continuum image is shown in Fig. 3. We detect a range of diffuse structures in addition to the core and jets. Consistent with Pedlar et al. (1990), we detect the bright central component with a peak brightness of 14.3 ± 0.7 Jy beam−1 and a bright southern jet and lobe extending to ∼16″ (5.6 kpc) from the core. We do not detect any counter-jet to the north. However, we detect a faint continuum component with a total flux density of ∼25 mJy about 4″ north-east from the core. Further to the north, we detect a lobe at ∼11″ with a flux density of ∼480 mJy. We also detect diffuse radio emission that connects all these components.

The H I spectrum against the bright radio core as obtained from our JVLA data over the entire observing band is shown in Fig. 2. Because the observing band is so wide, we detect the broad H I absorption at the systemic velocity of NGC 1275 as well as the well-known high-velocity absorber redshifted by ∼2850 km s−1 with respect to the systemic velocity of NGC 1275. Additionally, we detect a hitherto unknown absorption feature redshifted by ∼2660 km s−1 from the systemic velocity of NGC 1275. These two high-velocity absorbers are briefly discussed in the appendix. All three absorption features are only detected against the central core, and we do not detect absorption against any other continuum components.

Our focus is on the low-velocity absorber, which we show in detail in Fig. 4. This absorption feature is particularly broad, with a FWHM ∼500 km s−1, and the profile is relatively symmetric about the systemic velocity. We also show an H I spectrum that we obtained during test observations with the WSRT in April 2001 (Morganti et al. 2020). The spatial resolution of these observations is 15″. The two profiles are very similar, which confirms that all H I absorption comes from the central arcsecond region. If the absorption is assumed to arise against the core of 3C 84, we obtain a peak optical depth of τ = −ln(1 + ΔS/cfS​c) = 0.004 for a peak absorption of 40 mJy beam−1, where ΔS is the depth of the absorption line, S​c is the continuum flux density, and cf is the covering factor, which is assumed to be equal 1. This is lower than the typical optical depths of 0.01−0.02 found in radio galaxies (Morganti & Oosterloo 2018). From integrating the VLA spectrum, we find ∫τ dv = 1.34 ± 0.07 km s−1. The standard formula for estimating the H I column density

N H I = 1.82 × 10 18 T spin ∫ τ d v cm − 2 , $$ \begin{aligned} N_{{\text{ H}}{\small {\uppercase {\text{ i}}}}} = 1.82 \times 10^{18} \ T_{\rm spin} \int \tau \ \mathrm{d}v \;\; \mathrm{cm}^{-2}, \end{aligned} $$

However, the main result from the VLBA observations is that no trace of the H I absorption feature at the systemic velocity seen with the JVLA is detected. When the VLBA cube is smoothed to a velocity resolution of 50 km s−1, the noise level is 0.4 mJy beam−1. Much of the absorption in the JVLA spectrum is at the level of 30−40 mJy, therefore we should have detected this absorption feature at high significance if it occurred against the bright inner continuum. Narrow features at the 3σ level are tentatively seen, but because they are typically very narrow, none was considered a convincing detection. In any case, even if they were real detections, they would be too faint and narrow to account for even a small fraction of the JVLA H I absorption. Using the noise level, we can derive the upper limit in optical depth and column density of the H I on pc scales. For an rms noise level of 1 mJy beam−1 and a peak flux density of 3.5 Jy beam−1 at S1 and 0.8 Jy beam−1 at the core position, we derive a 3σ upper limit of τ < 0.0006 against S1 and τ < 0.0026 against the core.

5. Discussion

5.1. Origin of the H I absorption

The main observational result of this paper is that we clearly detect the low-velocity H I absorption against the central ≲1 arcsec region (≲350 pc) of 3C 84 in the JVLA data, while there is no trace of this absorption in the higher-resolution VLBA observations. Together with the information from the WSRT observations, this allows us to determine the location of the absorbing gas and learn more about the properties of the circum-nuclear structures of 3C 84.

As mentioned above, the low-velocity absorber was also observed with the WSRT at a spatial resolution of ∼15″. This scale is an order of magnitude larger than was achieved with the JVLA. A comparison of the WSRT and the JVLA spectra is shown in Fig. 4, and it is clear that the two spectra agree very well, both in terms of the depth of absorption an in the shape of the absorption profile. This good agreement implies that the gas giving rise to the low-velocity absorption is distributed on sub-arcsecond scales (i.e. < 350 pc, the size of the JVLA beam), and that the background continuum structure against which the absorption is seen is this large at most.

Because the absorption comes from a small region, an obvious candidate for the continuum emission against which the absorption is seen would be the bright VLBA continuum structure, which has a size of about 50 mas (i.e. a few dozen parsecs), and which includes the core and the bright inner jet. However, our VLBA data show no absorption against these components, even though the sensitivity of our observations is such that we would easily have detected this absorption if it originated there. This means that the background continuum must be located away from the core (i.e. at radii larger than a few dozen parsec). It also implies that the absorption seen in the VLA data is not due to a foreground cloud. An unrelated cloud like this would be expected to produce absorption against the 14 Jy core of 3C 84 in a similar way as the high-velocity system.

Another indication about the nature of the H I structure causing the absorption is that the velocity range of the H I profile is very similar to that of the molecular CND as observed (in emission) in warm H2 by Wilman et al. (2005) and Scharwächter et al. (2013) and in CO(2−1) by Nagai et al. (2019) and Oosterloo et al. (2023). The left panel in Fig. 6 shows a comparison between the integrated spectrum of the CO(2−1) emission of the inner ∼50 pc radius as obtained by Oosterloo et al., together with the H I absorption profile (extracted from the central beam of the natural weighting cube). The two spectra clearly agree very well in terms of total width and of asymmetry. This suggests that the H I absorption comes from atomic gas that is located in a structure similar to the molecular CND, which has a diameter of about 100 pc (∼0.3 arcsec) and extends in the direction roughly perpendicular to the jet.

thumbnail Fig. 6.

Comparison between the JVLA H I absorption data and the kinematics of the molecular CND as observed with ALMA by Oosterloo et al. (2023). Left: the JVLA H I absorption profile (extracted from the central beam of the natural weighting cube and shown in red; inverted to facilitate comparison) and the profile of the integrated CO(2−1) emission from the circum-nuclear molecular disc shown in blue (Oosterloo et al. 2023). The H I profile has been smoothed to the same velocity resolution as the CO(2−1) profile, and the profiles have been normalised to the peak. Right: position-velocity diagram of the circum-nuclear CO disc taken along the kinematic major axis (from Oosterloo et al. 2023). The full range of velocities clearly only occurs for radii smaller than 50 pc. In both panels the velocities are with respect to the systemic velocity of NGC 1275.

To determine this, we extracted spectra from the JVLA data cube made with robust = −1, which has a resolution of 1.10 × 1.02 arcsec, at two diametrically opposite locations, offset 0 . ″ 5 $ 0{{\overset{\prime\prime}{.}}}5 $ north-east and 0 . ″ 5 $ 0{{\overset{\prime\prime}{.}}}5 $ south-west from the peak continuum of 3C 84 (Fig. 7). Although the peak continuum is only very slightly resolved, we find that the peak of the H I absorption shifts bluewards from the north-eastern part of the centre to the south-western part. This is exactly the behaviour expected from the kinematics of the CND, whose kinematical major axis has a position angle of 68°, and the NE part has receding velocities, as shown in the right panel of Fig. 6. This is another strong indication that the H I gas seen in absorption arises from the molecular circum-nuclear gas disc and is probably located relatively close to the core because the highest velocities of the gas are observed there.

thumbnail Fig. 7.

H I spectra extracted from the JVLA robust = −1 data cube 0 . ″ 5 $ 0{{\overset{\prime\prime}{.}}}5 $ north-east and 0 . ″ 5 $ 0{{\overset{\prime\prime}{.}}}5 $ south-west of the peak continuum, showing the same rotation signature as the CND. For clarity, the spectra have been smoothed to a resolution of 21.6 km s−1. Velocities are relative to systemic.

If the CND causes the low-velocity H I absorption system, the radio continuum causing the absorption probably also spans similar spatial scales. Intriguingly, Silver et al. (1998) used multi-wavelength VLBI observations and reported diffuse radio emission spanning ∼100 pc that spatially overlapped with the circum-nuclear disc. Nagai & Kawakatu (2021) reanalysed the 330 MHz VLBA data of Silver et al. (1998) and highlighted the spatial coincidence between this extended continuum emission and the molecular CND. We illustrate this in Fig. 8, where this extended continuum emission is shown on top of the velocity field of the molecular CND. Nagai et al. (2019) argued that this strongly suggests that the relativistic electrons emitting the radio continuum and the molecular gas disc are co-spatial, and that supernova explosions in the CND are at the origin of these electrons, and not the AGN. Silver et al. (1998) reported that the radio luminosity of the CND agreed well with the infrared luminosity as seen in normal galaxies (Condon 1992). This indicates on-going star formation, which is further supported by the high density of the molecular gas. Based on the calculations and assumptions presented in Silver et al. (1998), a star formation rate (SFR) of 3 M⊙ yr−1 is expected (see the discussion in Nagai & Kawakatu 2021). This SFR is also capable of inducing turbulence in the CND, where Nagai et al. (2019) and Oosterloo et al. (2023) observed a relatively high velocity dispersion of ∼25 km s−1. All this strongly suggests that the H I absorption occurs against this radio continuum, which arises from the CND itself.

thumbnail Fig. 8.

Overlay of the diffuse radio continuum emission at 330 MHz from the VLBA data of Silver et al. (1998) and the velocity field of the central CO(2−1) disc from the reanalysed ALMA data presented in Oosterloo et al. (2023).

Finally, another line of investigation also supports the idea that the background radio continuum emission very likely does not arise from the radio core and jets: the low-velocity absorber has been observed at least four times over the last four decades (Crane et al. 1982; Sijbring 1993; Morganti et al. 2020, and the present work). In this time span, the radio AGN has shown significant variability of over 10 Jy at milliarcsecond scales (e.g. Taylor 1996; Silver et al. 1998; Momjian et al. 2002). However, remarkably, the depth of the absorbed flux reported by all these studies has remained the same. Because we do not expect such dramatic changes (and on such short timescales) in the optical depth and column density of the screen producing the H I absorption, this implies that the absorption does not occur against the variable radio component because if it did, it should have changed in flux in step with the continuum. Sijbring (1993) hypothesised that the absorption may arise only against the milliarcsecond-scale radio continuum, which is not variable. However, our VLBA observations clearly show that this is not the case either.

5.2. Properties of the H I gas and the CND

Although we determined the location of the gas producing the H I absorption, we cannot derive the exact distribution and extent of the H I on subkiloparsec scales. Furthermore, with the available data, it is difficult to comment on whether a vertical stratification of multiphase gas is present in the CND, as predicted by numerical simulations (Wada et al. 2016). However, we can attempt to derive some of the properties of the CND from our observations. Figure 6 presents a position-velocity diagram of the molecular CND, taken from the CO(2−1) data published by Oosterloo et al. (2023) along the kinematic major axis of the CND (position angle 68°). This shows the semi-Keplerian behaviour of the CND (see Nagai et al. 2019; Oosterloo et al. 2023 for details), and in particular, that the full velocity range of the CO(2−1) only occurs in the inner part of the CND. This suggests that at least part of the H I absorption comes from this inner region. Assuming the gas is in Keplerian motion, we can convert the highest detected velocities into the radius at which gas must be present. The largest deviations in the H I profile from the systemic velocity are about 300 km s−1, which gives a maximum deprojected velocity of 425 km s−1 when the inclination of the CND is assumed to be about 45° (Scharwächter et al. 2013). Bettoni et al. (2003) estimated the mass of the SMBH to be MBH = 108.61 M⊙. Based on this mass, this maximum velocity means that the smallest radius at which gas is present is about 20 pc. Nagai et al. (2019) estimated the sphere of influence of the SMBH in 3C 84 to be 60 pc. Thus, the H I appears to reach this inner region despite the influence of the active SMBH.

That the broad low-velocity H I absorption is not seen against the core and inner jet of 3C 84 but against much fainter diffuse continuum emission has strong implications for estimating the column density of the H I gas. Silver et al. (1998) not only detected this diffuse emission at 330 MHz, but were also able to estimate its flux at 1.4 GHz (and thus its spectral index). To do this, they tapered their 1.4-GHz VLBA data extremely, which allowed them to detect some of the extended continuum emission at this frequency as well, and they derived an integrated flux of 80 mJy at 1.4 GHz for the diffuse emission. If the H I absorption detected with the JVLA is seen against this diffuse CND emission, it would mean that the gas has a high optical depth of τ ∼ 0.80 and thus a high column density of 4.3 ± 0.2 × 1022 cm−2, assuming Tspin = 100 K and cf = 0.5. However, the conditions of H I gas located so close to the AGN might well be such that the spin temperature is at least 1000 K, leading to column densities above 1023 cm−2. These column densities are quite similar to those observed for the warm and cold molecular gas (Scharwächter et al. 2013; Oosterloo et al. 2023).

Nagai & Kawakatu (2021) used the approach presented in Kawakatu et al. (2020) to derive the thickness of the CND. They concluded that it is expected to be thin, with a height of about 10 pc at a radius of 100 pc. We do not see H I absorption against the central core and the beginning of the jet. This gives some weak constraints on the thickness of the H I distribution. If the CND has a constant thickness, there has to be a central hole in the CND, otherwise, we would always see gas in absorption. We estimated the radius of this hole to be about 20 pc above. Inside this radius, the gas is likely ionised throughout. When we take the inclination of the CND to be about 45°, this means that the CND cannot be thicker than the size of the central hole, otherwise, absorption would still be observed, implying a thickness of at most 20 pc. This may suggest the H I is distributed in a thicker disc than the molecular gas, similar to what was found in the case of Circinus (Izumi et al. 2018) and is predicted by numerical simulations (Wada et al. 2016). However, when we assume the CND to be flaring (i.e. a thickness proportional to the radius), our data cannot set constraints on the thickness of the distribution of the H I.

5.3. Similarity with Mrk 231 and overall relevance

Our data show that the broad H I absorption detected in 3C 84 originates from the CND of this AGN. It is particularly interesting to note that this finding bears a striking similarity to the case of Mrk 231 as presented by Carilli et al. (1998). Mrk 231 is one of the very few objects for which the multi-phase structure of the CND, including the continuum emission from its star formation, has been studied in detail.

In Mrk 231, broad H I absorption is also detected on arcsecond scales, but not on milliarcsecond scales. Based on this, Carilli et al. (1998) suggested that the absorption comes from H I in the circum-nuclear disc which has a size of about 60 pc, instead of being caused by clouds in front of the radio jet. Taylor et al. (1999) concluded that diffuse synchrotron emission is present in Mrk 231, which is most likely related to star formation activity in the CND. Carilli et al. (1998) were able to confirm their scenario by directly detecting the H I absorption against the continuum emission from the star-forming CND by using the short baselines of the VLBA (including the phased-up VLA). Unfortunately, our 3C 84 VLBA data do not have enough sensitivity on the shorter baselines to produce a deep enough image to trace such a disc at 1.4 GHz in 3C 84. The strength of the continuum, limiting the dynamic range reached, creates an additional complication, and dedicated observations are needed. Another similarity is that the optical depth of the absorption against the faint CND in Mrk 231 is also very high (τ = 0.17 ± 0.02), implying column densities of about 1023 cm−2.

3C 84 and Mrk 231 are very different types of radio sources and also live in very different environments. Their CNDs appear to have very similar structures, however. This might suggest that these CNDs are relatively common, but can only be identified when a rich set of observations is available.

5.4. Impact of the radio jet

A number of radio continuum studies have suggested interaction between the radio plasma and gas clouds in the central region in order to explain the observed change in position angle of the radio jet (e.g. Nagai et al. 2017; Kino et al. 2018 and references therein). This is particularly the case for the C3 component in the inner jet, which is located only a few milliarcseconds (i.e. a few parsec) from the SMBH. At the resolution of our observations, this region is included in our C component. At this location, a hot spot was detected that changed position at the various epochs of the available observations. This could indicate that the jet meanders in an inhomogeneous ambient medium, which results in the movement of the termination shock, or that it is due to time-variable ejection of jet flows (Nagai et al. 2017). In the scenario described above as obtained from our H I observations, the lack of H I absorption on parsec scales indicates that no H I is seen at the location of the jet and, most importantly, no kinematically disturbed cold gas is seen to result from gas interacting with the radio jet.

Because the cold gas seems to be absent in the central region (< 20 pc), it is interesting to study whether these interactions might instead be present in the ionised gas. A large amount of ionised gas in the central regions has been traced by the detection of free-free absorption (Walker et al. 2000) and of emission lines. Interestingly, when emission lines are used ([FeII] and Brγ), ionised gas is also seen to be distributed in a CND, which includes gas characterised by a relatively high velocity dispersion (Scharwächter et al. 2013) and possibly also outflowing gas resulting from the shocks (Riffel et al. 2020). However, even in the latter case, the data do not show any coincidence between the possibly disturbed kinematics of the gas and the radio jet. We did not detect gas with disturbed kinematics at the core location of H I either, even though the upper limit to the optical depth of the gas against the core C region is quite low, and lower than the typical values seen for the shallow and blueshifted H I absorption tracing jet-cloud interactions as seen in some radio galaxies (e.g. Morganti et al. 2005; Mahony et al. 2013).

Interestingly, another possible explanation could be that the observed change in position angle of the jet is instead due to wobbling of the CND caused by slight changes in the direction of the ongoing gas deposition from the surrounding filaments of molecular gas onto the disc. This would result in a small reorientation of the CND and hence of the jet (see the discussion in Oosterloo et al. 2023).

If the jet and the gas in the centre of 3C 84 do not interact directly, the type of feedback produced by 3C 84 is affected as well. It suggests that in this source, the jet has already opened a channel in the surrounding gaseous medium and has already expanded beyond the gas-rich central part of the galaxy. Because of this, we do not observe gas outflows driven by the jet, and the effect of the radio jet seems to have shifted to the “maintenance mode”, creating bubbles in the circum-galactic medium (CGM) and preventing much of the cooling of the CGM on galaxy scales and larger (as also suggested by Lim et al. 2008). Because of the size of the radio jet, which in 3C 84 extends to several kiloparsec, the scenario is consistent with what is seen in other radio galaxies (e.g. PKS 0023–63 and 3C 305; Morganti et al. 2021, 2023), as well as in low-luminosity radio sources (e.g. Rosario et al. 2019; Feruglio et al. 2020; Cresci et al. 2023). Thus, the impact of the radio jets on producing outflows appears to be limited to the first phase of their evolution, while at a later stage, their main impact is in the maintenance mode, affecting galaxy scales and beyond (see also Oosterloo et al. 2023).

6. Conclusions

We have presented a multi-resolution study of the H I absorption in 3C 84. The combination of these data with information on the circum-nuclear structure seen in molecular gas at a similar resolution, as well as detailed images of the radio continuum, have allowed us to expand our understanding of this iconic radio galaxy. We concentrated on the H I absorption centred on the systemic velocity of NGC 1275. Our data confirm that this absorption originates from the inner 350 pc (the scale of the JVLA A-array beam) central region. On the other hand, the higher-resolution (a few pc) VLBA data showed that the absorbing gas is not located against the core or inner jet.

The striking agreement we find between the velocities and shape of the H I profile and those of the molecular gas of the CND (Nagai et al. 2019; Oosterloo et al. 2023) strongly suggests that the H I absorption is associated with the CND. This is further confirmed by using the JVLA data at the highest possible resolution to show that apart from the velocity range, the H I absorption shows the same spatial kinematic signature as the CND.

These results tell us about the composition of the CND, and they now can form the basis for more detailed studies to investigate whether stratification is present in the different phases of the gas, as predicted by simulations (Wada et al. 2016).

The star formation in this circum-nuclear structure appears to be the key not only to explain the radio continuum emission, which provides the background for detecting H I absorption, but also to explain the feeding of the AGN by inducing turbulence to the gas (Nagai et al. 2019).

The similarity of the circum-nuclear H I (and cold molecular) properties between 3C 84 and Mrk 231, combined with the fact that these two objects are quite different as radio sources and in terms of their environment, suggests that CNDs with these properties may be relatively common. Observations with a wide dynamic range and high sensitivity are needed to explore this further.

We confirm the presence of H I in 3C 84 down to ∼20 pc from the SMBH through the comparison of the velocities observed with respect to those of the molecular gas observed in emission. This occurs despite possible effects of the AGN on the physical conditions of the gas. However, the non-detection of absorption against the core and inner jet suggests that the gas in the very inner region is mostly ionised. Thus, high spatial resolution H I absorption observations do not necessarily trace the H I gas in the very inner regions, and it is essential to have a good knowledge of the radio continuum that may provide the background for the absorption: sensitive observations with different spatial resolution are crucial for this. Furthermore, complementing the H I data with information on the molecular gas has been the key for the interpretation.

We did not detect kinematically disturbed H I gas despite the low optical depth reached by our observations. This result supports the picture in which the impact of jets may change as they evolve and expand farther into the ISM. Although new components are regularly ejected from the nucleus of 3C 84 that feed the jet, the fact that the overall jet is already evolved and many dozen kiloparsec in size suggests that it has already opened a channel in the ISM and that its main impact is in a maintenance mode and not in producing gas outflows. This is consistent with the proposed scenario in which the jets, while travelling through the hot cluster gas, create cavities in the hot cluster gas, dump energy into the cluster gas, and quench to some extent the cooling flow of the hot ICM of the Perseus cluster (Salomé et al. 2006; Lim et al. 2008; Fabian et al. 2011). On the other hand, around the edges of these rising plasma bubbles, the gas is compressed, which induces cooling of this gas, leading to the formation of the filaments of warm and cold gas (see e.g. Lim et al. 2008). These filaments can fuel the CND, which in turn is likely to cause the fuelling of the AGN (Nagai et al. 2019; Oosterloo et al. 2023). The results from our H I observations are consistent with this scenario.

Our study has highlighted the importance of combining the view of the line and the radio continuum on different scales (up to high spatial resolution) and of doing this for different phases of the gas to interpret H I absorption studies of AGN.

1

For the cosmology adopted in this paper, we assumed a flat Universe with the following parameters: H0 = 70 km s−1 Mpc−1, ΩΛ = 0.7, and ΩM = 0.3. For the redshift of 3C 84 of z = 0.01756 (systemic velocity Vsys = 5264 km s−1; Huchra et al. 1999), this gives a luminosity distance of 76.2 Mpc and an angular scale distance of 73.6 Mpc so that 1 mas = 0.357 pc.

Acknowledgments

C.C. acknowledges support via the ASTRON/JIVE summer student program. H.N. is supported by JSPS KAKENHI grants No. JP18K03709 and No. JP21H01137. S.M. thanks Junghwan Oh and Gabor Orosz for useful discussions on the VLBA data reduction. The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. The Westerbork Synthesis Radio Telescope was operated by ASTRON (Netherlands Institute for Radio Astronomy) with support from the Netherlands Foundation for Scientific Research (NWO). This paper makes use of the ALMA data of ADS/JAO. ALMA#2017.0.01257.S. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada), MOST and ASIAA (Taiwan), and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ.

References

  1. Audibert, A., Ramos Almeida, C., García-Burillo, S., et al. 2023, A&A, 671, L12 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  2. Bettoni, D., Falomo, R., Fasano, G., et al. 2003, A&A, 399, 869 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  3. Brusa, M., Cresci, G., Daddi, E., et al. 2018, A&A, 612, A29 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  4. Carilli, C. L., Wrobel, J. M., & Ulvestad, J. S. 1998, AJ, 115, 928 [NASA ADS] [CrossRef] [Google Scholar]
  5. Combes, F. 2022, Eur. Phys. J. Web Conf., 265, 00047 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  6. Condon, J. J. 1992, ARA&A, 30, 575 [Google Scholar]
  7. Conselice, C. J., Gallagher, J. S., & Wyse, R. F. G. 2001, AJ, 122, 2281 [NASA ADS] [CrossRef] [Google Scholar]
  8. Conway, J. E., & Blanco, P. R. 1995, AJ, 449, L131 [NASA ADS] [Google Scholar]
  9. Crane, P., van der Hulst, J., & Haschick, A. 1982, IAU Symp., 97, 307 [NASA ADS] [Google Scholar]
  10. Cresci, G., Tozzi, G., Perna, M., et al. 2023, A&A, 672, A128 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  11. De Young, D. S., Roberts, M. S., & Saslaw, W. C. 1973, ApJ, 185, 809 [NASA ADS] [CrossRef] [Google Scholar]
  12. Fabian, A. C., Johnstone, R. M., Sanders, J. S., et al. 2008, Nature, 454, 968 [NASA ADS] [CrossRef] [Google Scholar]
  13. Fabian, A. C., Sanders, J. S., Allen, S. W., et al. 2011, MNRAS, 418, 2154 [Google Scholar]
  14. Feruglio, C., Fabbiano, G., Bischetti, M., et al. 2020, ApJ, 890, 29 [Google Scholar]
  15. Gillmon, K., Sanders, J. S., & Fabian, A. C. 2004, MNRAS, 348, 159 [NASA ADS] [CrossRef] [Google Scholar]
  16. Giovannini, G., Savolainen, T., Orienti, M., et al. 2018, Nat. Astron., 2, 472 [Google Scholar]
  17. Girdhar, A., Harrison, C. M., Mainieri, V., et al. 2022, MNRAS, 512, 1608 [NASA ADS] [CrossRef] [Google Scholar]
  18. Greisen, E. W. 2003, Astrophys. Space Sci. Lib., 285, 109 [NASA ADS] [Google Scholar]
  19. Hogan, M. T. 2014, The radio properties of brightest cluster galaxies. PhD Thesis, University of Durham. http://etheses.dur.ac.uk/11008/ [Google Scholar]
  20. Huchra, J. P., Vogeley, M. S., & Geller, M. J. 1999, ApJS, 121, 287 [Google Scholar]
  21. Izumi, T., Wada, K., Fukushige, R., et al. 2018, ApJ, 867, 48 [NASA ADS] [CrossRef] [Google Scholar]
  22. Jaffe, W. 1990, A&A, 240, 254 [NASA ADS] [Google Scholar]
  23. Lim, J., Ao, Y., Dinh-V-Trung, et al. 2008, ApJ, 672, 252 [NASA ADS] [CrossRef] [Google Scholar]
  24. Kawakatu, N., Wada, K., & Ichikawa, K. 2020, ApJ, 889, 84 [NASA ADS] [CrossRef] [Google Scholar]
  25. Kino, M., Wajima, K., Kawakatu, N., et al. 2018, ApJ, 864, 118 [NASA ADS] [CrossRef] [Google Scholar]
  26. Kino, M., Niinuma, K., Kawakatu, N., et al. 2021, ApJ, 920, L24 [CrossRef] [Google Scholar]
  27. Maccagni, F. M., Morganti, R., Oosterloo, T. A., et al. 2018, A&A, 614, A42 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  28. Mahony, E. K., Morganti, R., Emonts, B. H. C., et al. 2013, MNRAS, 435, L58 [NASA ADS] [CrossRef] [Google Scholar]
  29. Minkowski, R. 1955, Carnegie Yrb., 54, 25 [Google Scholar]
  30. Momjian, E., Romney, J. D., & Troland, T. H. 2002, ApJ, 566, 195 [NASA ADS] [CrossRef] [Google Scholar]
  31. Morganti, R., & Oosterloo, T. 2018, A&ARv, 26, 4 [Google Scholar]
  32. Morganti, R., Tadhunter, C. N., & Oosterloo, T. A. 2005, A&A, 444, L9 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  33. Morganti, R., Oosterloo, T., Schulz, R., et al. 2020, Proc. Int. Astron. Union, 342, 85 [NASA ADS] [Google Scholar]
  34. Morganti, R., Oosterloo, T., Tadhunter, C., et al. 2021, A&A, 656, A55 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  35. Morganti, R., Murthy, S., Guillard, P., et al. 2023, Galaxies, 11, 24 [NASA ADS] [CrossRef] [Google Scholar]
  36. Murthy, S., Morganti, R., Oosterloo, T., et al. 2021, A&A, 654, A94 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  37. Murthy, S., Morganti, R., Wagner, A. Y., et al. 2022, Nat. Astron., 6, 488 [CrossRef] [Google Scholar]
  38. Nagai, H., & Kawakatu, N. 2021, ApJ, 914, L11 [NASA ADS] [CrossRef] [Google Scholar]
  39. Nagai, H., Suzuki, K., Asada, K., et al. 2010, PASJ, 62, L11 [NASA ADS] [CrossRef] [Google Scholar]
  40. Nagai, H., Fujita, Y., Nakamura, M., et al. 2017, ApJ, 849, 52 [NASA ADS] [CrossRef] [Google Scholar]
  41. Nagai, H., Onishi, K., Kawakatu, N., et al. 2019, ApJ, 883, 193 [NASA ADS] [CrossRef] [Google Scholar]
  42. O’Dea, C. P., Baum, S. A., & Gallimore, J. F. 1994, ApJ, 436, 669 [CrossRef] [Google Scholar]
  43. Oosterloo, T. A., & Morganti, R. 2005, A&A, 429, 469 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  44. Oosterloo, T., Raymond Oonk, J. B., Morganti, R., et al. 2017, A&A, 608, A38 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  45. Oosterloo, T., Morganti, R., & Murthy, S. 2023, Nat. Astron., in press [Google Scholar]
  46. Paraschos, G. F., Mpisketzis, V., Kim, J.-Y., et al. 2023, A&A, 669, A32 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  47. Pedlar, A., Ghataure, H. S., Davies, R. D., et al. 1990, MNRAS, 246, 477 [NASA ADS] [Google Scholar]
  48. Riffel, R. A., Storchi-Bergmann, T., Zakamska, N. L., et al. 2020, MNRAS, 496, 4857 [NASA ADS] [CrossRef] [Google Scholar]
  49. Rosario, D. J., Togi, A., Burtscher, L., et al. 2019, ApJ, 875, L8 [NASA ADS] [CrossRef] [Google Scholar]
  50. Rose, T., Edge, A. C., Combes, F., et al. 2019, MNRAS, 485, 229 [Google Scholar]
  51. Rose, T., McNamara, B. R., Combes, F., et al. 2023, MNRAS, 518, 878 [Google Scholar]
  52. Russell, H. R., McNamara, B. R., Fabian, A. C., et al. 2019, MNRAS, 490, 3025 [Google Scholar]
  53. Salomé, P., Combes, F., Edge, A. C., et al. 2006, A&A, 454, 437 [CrossRef] [EDP Sciences] [Google Scholar]
  54. Sanders, J. S., & Fabian, A. C. 2007, MNRAS, 381, 1381 [Google Scholar]
  55. Saraf, M., Wong, O. I., Cortese, L., et al. 2023, MNRAS, 519, 4128 [NASA ADS] [CrossRef] [Google Scholar]
  56. Savolainen, T., Giovannini, G., Kovalev, Y. Y., et al. 2023, A&A, 676, A114 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  57. Scharwächter, J., McGregor, P. J., Dopita, M. A., et al. 2013, MNRAS, 429, 2315 [CrossRef] [Google Scholar]
  58. Shepherd, M. C. 1997, ASP Conf. Ser., 125, 77 [Google Scholar]
  59. Shulevski, A., Morganti, R., Barthel, P. D., et al. 2015, A&A, 579, A27 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  60. Sijbring, L. G. 1993, Ph.D. Thesis, Groningen University, The Netherlands [Google Scholar]
  61. Sijbring, D., de Bruyn, A. G., Jaffe, W. J., et al. 1989, European Southern Observatory Conference and Workshop Proceedings, 32, 107 [NASA ADS] [Google Scholar]
  62. Silver, C. S., Taylor, G. B., & Vermeulen, R. C. 1998, ApJ, 502, 229 [NASA ADS] [CrossRef] [Google Scholar]
  63. Struve, C., & Conway, J. E. 2010, A&A, 513, A10 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  64. Suzuki, K., Nagai, H., Kino, M., et al. 2012, ApJ, 746, 140 [Google Scholar]
  65. Taylor, G. B. 1996, ApJ, 470, 394 [NASA ADS] [CrossRef] [Google Scholar]
  66. Taylor, G. B., O’Dea, C. P., Peck, A. B., & Koekemoer, A. M. 1999, ApJ, 512, L27 [NASA ADS] [CrossRef] [Google Scholar]
  67. Tremblay, G. R., Oonk, J. B. R., Combes, F., et al. 2016, Nature, 534, 218 [NASA ADS] [CrossRef] [Google Scholar]
  68. van Bemmel, I. M., Morganti, R., Oosterloo, T., et al. 2012, A&A, 548, A93 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  69. Veilleux, S., Maiolino, R., Bolatto, A. D., et al. 2020, A&ARv, 28, 2 [NASA ADS] [CrossRef] [Google Scholar]
  70. Wada, K., Schartmann, M., & Meijerink, R. 2016, ApJ, 828, L19 [Google Scholar]
  71. Walker, R. C., Dhawan, V., Romney, J. D., et al. 2000, ApJ, 530, 233 [NASA ADS] [CrossRef] [Google Scholar]
  72. Wilman, R. J., Edge, A. C., & Johnstone, R. M. 2005, MNRAS, 359, 755 [NASA ADS] [CrossRef] [Google Scholar]
  73. Yu, A. P.-Y., Lim, J., Ohyama, Y., et al. 2015, ApJ, 814, 101 [NASA ADS] [CrossRef] [Google Scholar]

Appendix A: Two high-velocity absorbers

In addition to the broad H I absorption detected at the systemic velocity of NGC 1275, two other absorption systems were detected in the JVLA data. One of these is a deep, narrow absorption line at Vhel = 8115 km s−1 (Fig. A.1) that was discovered by De Young et al. (1973) and was studied at high spatial and velocity resolution with the VLBA by Momjian et al. (2002). This absorption is associated with a foreground cluster galaxy, commonly referred to as the high-velocity system (HVS), that is falling towards NGC 1275, but is still ∼100 kpc distant from the AGN host galaxy. The HVS was first identified by Minkowski (1955) and causes most of the dust lines seen in optical images of NGC 1275. It is also detected in optical emission lines (e.g. Yu et al. 2015) and in absorption in X-rays (e.g. Sanders & Fabian 2007).

thumbnail Fig. A.1.

Spectrum of the high-velocity absorption system redshifted 2850 km s−1 with respect to the systemic velocity of NGC 1275. It was earlier studied by Momjian et al. (2002).

For this absorption, we measure an integrated optical depth of ∫τdv = 1.63 ± 0.07 km s−1, which gives a column density of 2.9 ± 0.1 × 1020 cm−2 assuming Tspin = 100 K. This is consistent with the results of Momjian et al. (2002).

In addition to this previously known HVS absorption system, we discovered a much fainter absorption component redshifted by 2660 km s−1 with respect to the systemic velocity of NGC 1275 (Fig. A.2), but slightly blueshifted with respect to the HVS. This component is also likely associated with the infalling galaxy because ionised gas with velocities blueshifted up to 800 km s−1 with respect to those of the main body of the HVS has been found in optical data (Yu et al. 2015). This gas is likely stripped from the HVS by the cluster medium, and this absorption is likely part of this gas.

thumbnail Fig. A.2.

Spectrum of the newly discovered absorption system redshifted by 2660 km s−1 with respect to the systemic velocity of NGC 1275. The two Gaussian components fitted to the profile are indicated in black. Their parameters are listed in Table A.1.

This newly discovered absorption system appears to consist of two components. We fitted Gaussians to these two components, the parameters of which are given in Table A.1. The column densities of the two components are 5.9 ± 1.6 × 1017 cm−2 and 1.3 ± 0.2 × 1018 cm−2 (assuming Tspin = 100 K).

Table A.1.

Parameters of Gaussian fits to the newly discovered absorption system.

All Tables

Table 1.

Observations and imaging results.

In the text
Table A.1.

Parameters of Gaussian fits to the newly discovered absorption system.

In the text

All Figures

thumbnail Fig. 1.

Images of 3C 84 system from the kiloparsec to parsec scales relevant for this study. Left: continuum map at 90 cm (Pedlar et al. 1990; angular resolution of ∼5″ ≃ 1.8 kpc) in grey contours superimposed on the map of molecular gas (Lim et al. 2008; resolution 3″ ≡ 1.1 kpc, black contours), with the X-ray image from Fabian et al. (2011) in the background. Second panel from the left: image at ∼1″ ≡ 360 pc resolution at 1380 MHz from Pedlar et al. (1990) showing the continuum structure against which the H I is seen in absorption. Last two panels on the right: were obtained using the VLBA at 330 MHz and 1.4 GHz by Silver et al. (1998) and show, in addition to the core-jet structure (right), a diffuse, extended structure (which they call mini-halo) on larger scales (∼200 pc ≡ 0 . ″ 5 $ 0{{\overset{\prime\prime}{.}}}5 $ radius).
In the text
thumbnail Fig. 2.

Full H I profile obtained with the JVLA showing the various absorption systems. We focus on the low-velocity system, centred on the systemic velocity of NGC 1275. The high-velocity absorption is briefly discussed in the appendix. Top: spectrum using the full intensity range. Bottom: same as above, but with a different vertical scaling. The newly discovered component is indicated with an arrow.
In the text
thumbnail Fig. 3.

Natural weighted JVLA continuum image. The image has an rms noise of ∼0.4 mJy beam−1 and a beam size of 1.64″ × 1.52″ ≡ 585 pc × 543 pc with a position angle of 68 . ° 9 $ 68{{\overset{\circ}{.}}}9 $, shown in the bottom right corner. The contours start at 3σ and increase with factors of 2 $ \sqrt{2} $; the 3σ negative contours are shown in grey. The image shows a range of structures starting from the bright unresolved core, which has a peak flux density of 14.3 Jy beam−1, the southern jet, diffuse lobes, and also diffuse radio emission connecting all these structures.
In the text
thumbnail Fig. 4.

Comparison between the H I spectra of the low-velocity absorption system obtained with the JVLA and the WSRT. The WSRT spectrum has a velocity resolution of ∼34 km s−1, and the JVLA spectrum, originally at a velocity resolution of 2.7 km s−1, has been binned to roughly the WSRT channel resolution, followed by a Hanning smoothing. It thus has the same velocity resolution as the WSRT spectrum. Despite the difference in spatial resolution between the two observations (∼5 kpc for the WSRT and ∼350 pc for the JVLA), the strong similarity between the two profiles indicates that the H I producing the absorption and/or the background continuum is located on sub-kiloparsec scales from the centre (see text for details). Velocities are relative to systemic.
In the text
thumbnail Fig. 5.

Naturally weighted VLBA continuum map. The map has an rms noise of 2.1 mJy beam−1 and a beam of 9.4 mas × 6.7 mas (3.4 pc × 2.4 pc), with a position angle of − 28 . ° 5 $ -28{{\overset{\circ}{.}}}5 $. The contours start at 3σ and increase with a factor of 2 $ \sqrt{2} $. The 3σ negative contours are shown in grey. We detect the core (C), the southern bright jet (S1), its diffuse extension (S2), and the diffuse lobe farther south of the core (S3). We measure a peak flux density of 3.5 Jy beam−1 at S1 and an integrated flux of 9.14 Jy (which includes C+S1+S2+S3).
In the text
thumbnail Fig. 6.

Comparison between the JVLA H I absorption data and the kinematics of the molecular CND as observed with ALMA by Oosterloo et al. (2023). Left: the JVLA H I absorption profile (extracted from the central beam of the natural weighting cube and shown in red; inverted to facilitate comparison) and the profile of the integrated CO(2−1) emission from the circum-nuclear molecular disc shown in blue (Oosterloo et al. 2023). The H I profile has been smoothed to the same velocity resolution as the CO(2−1) profile, and the profiles have been normalised to the peak. Right: position-velocity diagram of the circum-nuclear CO disc taken along the kinematic major axis (from Oosterloo et al. 2023). The full range of velocities clearly only occurs for radii smaller than 50 pc. In both panels the velocities are with respect to the systemic velocity of NGC 1275.
In the text
thumbnail Fig. 7.

H I spectra extracted from the JVLA robust = −1 data cube 0 . ″ 5 $ 0{{\overset{\prime\prime}{.}}}5 $ north-east and 0 . ″ 5 $ 0{{\overset{\prime\prime}{.}}}5 $ south-west of the peak continuum, showing the same rotation signature as the CND. For clarity, the spectra have been smoothed to a resolution of 21.6 km s−1. Velocities are relative to systemic.
In the text
thumbnail Fig. 8.

Overlay of the diffuse radio continuum emission at 330 MHz from the VLBA data of Silver et al. (1998) and the velocity field of the central CO(2−1) disc from the reanalysed ALMA data presented in Oosterloo et al. (2023).
In the text
thumbnail Fig. A.1.

Spectrum of the high-velocity absorption system redshifted 2850 km s−1 with respect to the systemic velocity of NGC 1275. It was earlier studied by Momjian et al. (2002).
In the text
thumbnail Fig. A.2.

Spectrum of the newly discovered absorption system redshifted by 2660 km s−1 with respect to the systemic velocity of NGC 1275. The two Gaussian components fitted to the profile are indicated in black. Their parameters are listed in Table A.1.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.