A tidal tale: detection of several stellar streams in the environment of NGC 1052

The possible existence of two dark-matter-free galaxies (NGC1052-DF2 and NGC1052-DF4) in the ﬁeld of the early-type galaxy NGC1052 presents a challenge to theories of dwarf galaxy formation according to the current cosmological paradigm. We carried out a search for signatures of past interactions connected to the putative hosts of NGC1052-DF2 and NGC1052-DF4 using a very deep L -band image obtained with the 0.7m Jeanne Rich telescope that reach a surface brightness limit of 28.5magarcsec − 2 in the r band. We found several low-surface brightness features, possibly consistent with an ongoing merger history in this group. We ﬁnd a tidal interaction between NGC1052 and NGC1047, conﬁrming a physical association. Furthermore, we ﬁnd a stellar loop around NGC1052 in the direction of NGC1042 and a stellar stream pointing in the direction of NGC1052-DF2, but they are not directly connected. We ﬁnd no evidence for a recent tidal interaction for NGC1052-DF2 and NGC1052-DF4. No LSB features have been uncovered around the spiral galaxy NGC1042, which leaves the association (physical or projected) between NGC1052 and NGC1042 ambiguous, although they have similar radial velocities. Their association will only be established when accurate distances to both objects have been measured.

The apparent lack of dark matter of NGC 1052-DF2 and NGC 1052-DF4 makes these objects interesting and has raised discussions about their origin.A primordial dwarf galaxy should always contain a dynamically significant portion of dark matter -see, e.g. the dual dwarf theorem in Kroupa (2012) and its implications -making the possible absence of dark matter in these galaxies a direct confrontation with the current cosmological paradigm.Several scenarios have been suggested, for example, that they were tidally stripped from their dark matter by a close-passage to NGC 1052 (Ogiya 2018), or that they are the direct results of a tidal interaction, i.e.Tidal Dwarf Galaxies (TDG, Duc 2012;Duc et al. 2014).Mergers producing such TDGs are expected to leave observable signatures in the form of several streams and shells around NGC 1052 (and/or the other massive galaxies in the field) that can potentially be detected by deep imaging for several Gyr after the interaction.
NGC 1052 is an active radio galaxy consisting of two radio sources, a compact one, and an extended one in the form of a two-sided radio jet (Tingay et al. 2003;Kadler et al. 2004).Pierce et al. (2005) studied the recent history of NGC 1052 and argued that the young central starburst had to start from gas enriched in α elements.Such gas is usually found in massive spirals.This young central starburst has an age estimate of ∼1 Gyr coinciding with the estimated age of the accretion of the H I clouds found around NGC 1052 (van Gorkom et al. 1986).Its distance is estimated with surface brightness fluctuations (SBF) and measured to be D = 19 Mpc (Tully et al. 2013).Another giant galaxy in the same field is NGC 1042 -a spiral galaxy separated by 15 from NGC 1052 to its south-east.There have been various distance estimations placing NGC 1042 between 4 and 20 Mpc (Tully et al. 1992;Theureau et al. 2007;Tully et al. 2008;Luo et al. 2016).A reconstruction of the velocity field of the galaxies in the local universe puts NGC 1042 at a distance of 13 Mpc (Theureau et al. 2007), which would mean that NGC 1052 and NGC 1042 are companions only in projection.However, they have similar velocities with v = 1376 km s −1 for NGC 1042 (Adelman-McCarthy & et al. 2009) and v = 1510 km s −1 for NGC 1052 (Denicoló et al. 2005).Furthermore, in the extended field around NGC 1052, there appears to be a plethora of giant galaxies sharing the same velocity (Garcia 1993, see also Section 5 and 6 of Trujillo et al. 2019).
In this work, we raise the question whether NGC 1052-DF2 and NGC 1052-DF4 are physically associated with the giant galaxies, as well as whether tidal interactions can be observed between the galaxies.We have acquired deep images to recover the low-surface brightness features of the field around NGC 1052 and probe the signatures of tidal features.

Observations and data processing
As part of the Halos and Environments of Nearby Galaxies (HERON) survey (Rich et al. 2017(Rich et al. , 2018)), we have acquired deep images with the 0.7 meter Jeanne Rich telescope (Brosch et al. 2015;Rich et al. 2017) located at Lockwood Valley (CA, USA) between October and November 2018.The camera has a pixel scale of 1.1 arcsec and a field of view of ∼ 1 square degree, allowing us to cover NGC 1052, NGC 1042, NGC 1052-DF2, and NGC 1052-DF4 in a single image field.In total, we secured 245 × 300 second exposures, giving a total exposure of 20.4 hours.The observations were conducted at an airmass between 1.4 and 1.8.All observations were taken under photometric conditions when the moon was bellow the horizon.During the observations the sky brightness was between 21.7 to 22.0 mag arcsec −2 (V-band).The reduction process was carried out by correcting bias and flat-fielding.For the creation of the flat, we mask the science images using Noisechisel (Akhlaghi & Ichikawa 2015) and combine them .Due to the presence of gradients, we use Noisechisel to obtain the sky of each image, which is subtracted from the already corrected images from bias and flat-fielding.Astrometry was obtained through the scamp package (Bertin 2006).Finally we combine all exposures with a 3σ resistant mean.
The zero point was derived using APASS (Henden & Munari 2014).To transform from our instrumental luminance L magnitudes to the commonly used SDSS r band, we applied a linear transformation using 25 standard stars from the APASS catalog.We determine the surface brightness limits by measuring the standard deviation of blank pixels (pixels with no sources) in 10×10 arcsec boxes on the luminance L image.The 3σ standard deviation -transformed to the r-band using the linear photomet-ric transformation -is µ lim,r = 28.5 mag arcsec −2 .This limit is comparable to the depth reached by other telescopes of modest aperture size (Martínez-Delgado et al. 2010;Karachentsev et al. 2015;Javanmardi et al. 2016;Henkel et al. 2017).

Low-surface brightness features
Several Low-surface brightness (LSB) features stand out in the field around NGC 1052 by merely having a glance at Figure 1.We have modelled NGC 1052 using IRAF's ellipse algorithm (Tody 1986) after removing the faint objects and masking bright stars together with the companion galaxies.The modelsubtracted image is presented in Figure 2. Most prominently, there is a narrow stream coming from NGC 1052 to the vicinity of NGC 1052-DF2 which we refer to as Stream 1.A loop is visible to the south-west (loop SW), close to NGC 1042 but apparently disconnected from it.The tidal tail as seen in H I around NGC 1052 extends towards NGC 1047 (see Figure 2).NGC 1047 shows a clear tidal distortion/extension to the northwest, which is on the opposing side in respect to NGC 1052.NGC 1047 also shows tidal features towards NGC 1052, which could be a bridge.The outer isophotes of NGC 1047 appear to be boxy.We also note an arc feature to the north-west of NGC 1052.While we find multiple tidal features around NGC 1052, no convincing features are visible around NGC 1042.
Stream 1 is a striking feature.It appears as a straight line coming from NGC 1052 in the direction of NGC 1052-DF2.It stops 100 arcsec before the center of NGC 1052-DF2.The residual image of NGC 1052 shows that Stream 1 is extending towards the center of the bright elliptical galaxy.
To study the LSB features around NGC 1052-DF2 and NGC 1052-DF4 we have modelled and subtracted the two galaxies from the images (Figure 3).After the subtraction, no remaining features are left.While NGC 1052-DF2 has a change of the Position Angle with respect to the radius, NGC 1052-DF4 shows no sign of it.

Discussion
NGC 1052 resides in an environment where multiple interactions occurred in the past and are still on-going.The LSB features between NGC 1052 and NGC 1047 and the lopsided LSB feature on the counterside of NGC 1047 undoubtedly show that NGC 1052 and NGC 1047 have an ongoing interaction and are therefore at the same distance.The boxy appearance of the outer isophotes of NGC 1047 further indicates tidal disturbance.NGC 1052 also contains an irregularly shaped distribution of H I (van Gorkom et al. 1986), with what appears to be a tidal arm extending towards the south-west.van Gorkom et al. (1986) argued that the H I has been acquired 1 Gyr ago.In addition, Tingay et al. (2003) argued that NGC 1052 is a restarted radio source.The extended radio source is likely much older than the compact radio source at the nucleus, and these sources could have been created during different events.
The relative position of Stream 1 and NGC 1052-DF2 may suggest that the galaxy is a TDG formed at a tip of a tidal arm (Bournaud & Duc 2006).Indeed old TDGs have the structural properties (central surface brightness, large effective radius, lack of dark matter content) of NGC 1052-DF2 and NGC 1052-DF4 (Duc et al. 2014).While the hypothesis of a young TDG is falsified by the stellar age of NGC 1052-DF2 (9 Gyr, Fensch et al. 2018), one could argue that it is an old TDG, formed during an interaction 9 Gyr ago.This could solve the dark matterdeficiency issue and would be consistent with the location of NGC 1052-DF2 on the metallicity-stellar mass relation (see discussion in Fensch et al. 2018).However, tidal features, such as Stream 1 should disappear in a few Gyrs, especially if the galaxy suffered posterior mergers.Namely, some of the most extended tidal arms are around 100 kpc long and the free fall time from such a distance to NGC 1052 is about 1-2 Gyr, i.e. substantially less than the estimated age of NGC 1052-DF2.The linear morphology of Stream 1 is not expected if the stream had made several orbits around the host galaxy.Hence we can conclude that the alignment between this stream and NGC 1052-DF2 is pure chance and not linked with any TDG scenario.In addition, Stream 1 will not be the stripped tail of NGC 1052-DF2, because in such a case we would expect the tidal tail and the progenitor being connected (see, e.g. Figure 10 of Müller et al. 2018b).The morphology of the observed tidal features does not agree with the morphology of the tidal features in the simulation by Ogiya (2018).Alternatively and perhaps more likely, Stream 1 could originate from another stripped dwarf galaxy satellite.We note an unresolved source on the axis of the stream which is a candidate for the stripped core of its progenitor (02:41:25.3,−08:22:11.6).All of these suggest that this is likely a chance alignment.This is consistent with the subtracted model of NGC 1052-DF2 (Figure 3) revealing no clear signs of tidal features.
Turning to NGC 1042, also the absence of LSB features contains valuable information.While its systemic velocity (v = 1376 km s −1 ; Adelman-McCarthy & et al. 2009) is consistent with the velocity of the NGC 1052 group (v = 1425 ± 111 km s −1 ; van Dokkum et al. 2018a), i.e. a distance of ≈ 20 Mpc, the redshift-independent distance estimators favor a closer value (13 Mpc Theureau et al. 2007).As with NGC 1047, a visible tidal interaction would, without a doubt, put NGC 1042 at the NGC 1052 group.There is a loop which is in the direction of NGC 1042 but its morphology is rather consistent with a stream associated to another disrupted satellite orbiting around NGC 1052.No clear tidal features are directly connected to NGC 1042.This can either mean that NGC 1042 is within the group or has just arrived in the group, and no visible interac- tion has yet taken place2 , or that NGC 1042 is indeed only associated to NGC 1052 in projection and is, in reality, closer to us.Trujillo et al. (2019) argue that the properties of NGC 1052-DF2 and NGC 1052-DF4 would match with other known dwarf galaxies if they were closer to us, for instance at the distance of 13 Mpc.On the other hand, NGC 1042's spiral structure itself shows sharp changes in pitch angle and loosely wound outer spiral arms which are quite flocculent, which is reminiscent of spirals in the early time steps of merger simulations, not having quite managed to form elongated tidal tails yet.
The remaining giant galaxy to be discussed in the observed field is the spiral galaxy NGC 1035.While it also has a systemic velocity consistent with the NGC 1052 group (v = 1249 km s −1 ; Meyer et al. 2004), its latest distance estimate -based on Tully-Fisher measurements -is 14 ± 2.9 Mpc (Sorce et al. 2014).This places it in the foreground and is consistent with an absence of tidal features associated with this galaxy.

Conclusions
We aim to clarify the origin of the alleged low dark matter content of NGC 1052-DF2 and NGC 1052-DF4 by deep imaging.Deep imaging can, in principle, uncover tidal features whose morphology would reveal the origin of NGC 1052-DF2 and NGC 1052-DF4 either as tidal dwarf galaxies or a tidally darkmatter stripped primordial dwarf.We indeed found signs of many galaxy interactions in the group.Notably, there is an ongoing tidal interaction between NGC 1052 and NGC 1047.Furthermore, around NGC 1052 we find a loop in the direction of NGC 1042 and a stream in the direction of NGC 1052-DF2, but none of them are directly connected to these two galaxies.The features around NGC 1052 are probably much younger (1-3 Gyr) and are inconsistent with the estimated age of NGC 1052-DF2 (9 Gyr).We see no tidal features around NGC 1052-DF2 and NGC 1052-DF4, making a recent tidal stripping scenario unlikely.The tidal dwarf galaxy scenario at high redshift cannot be probed by our data as the tidal features might have been erased and are not detectable today.
We further found no convincing tidal features around the spiral galaxy NGC 1042.This indicates that either it is at a larger separation within the NGC 1052 group, e.g.just falling in, or that its association is only in projection and in reality, NGC 1042 is closer to us.Their final association (physical or projected) will be only firmly established once we have accurate distance estimations to both objects.

Fig. 1 .
Fig. 1.The NGC 1052 environment: stacked luminance (L) band obtained with the Jeanne Rich telescope.True color images from the SDSS are overlayed.The field of view is 1 deg 2 .The indicated 4 bar corresponds to 15 kpc at 13 Mpc or 22 kpc at 19 Mpc.North is to the top, east to the left.

Fig. 2 .
Fig. 2. The residual image after NGC 1052 was subtracted and convolved with a gaussian kernel, revealing the low surface brightness features in its environment.The contours show the H I distribution around NGC 1052 adopted from van Gorkom et al. (1986).