Evidence for low power radio jet-ISM interaction at 10 parsec in the dwarf AGN host NGC 4395

Black hole driven outflows in galaxies hosting active galactic nuclei (AGN) may interact with their interstellar medium (ISM) affecting star formation. Such feedback processes, reminiscent of those seen in massive galaxies, have been reported recently in some dwarf galaxies. However, such studies have usually been on kiloparsec and larger scales and our knowledge on the smallest spatial scales to which these feedback processes can operate is unclear. Here we demonstrate radio jet$-$ISM interaction on the scale of an asymmetric triple radio structure of $\sim$ 10 parsec size in NGC 4395. This triple radio structure is seen in the 15 GHz continuum image and the two asymmetric jet-like structures are situated on either side of the radio core that coincides with the optical {\it Gaia} position. The high resolution radio image and the extended [OIII]$\lambda$5007 emission, indicative of an outflow, are spatially coincident and are consistent with the interpretation of a low power radio jet interacting with the ISM. Modelling of the spectral lines using {\tt MAPPINGS}, and estimation of temperature using optical integral field spectroscopic data suggest shock ionization of the gas. The continuum emission at 237 GHz, though weak, was found to spatially coincide with the AGN. However, the CO(2$-$1) line emission was found to be displaced by around 20 parsec northward of the AGN core. The spatial coincidence of molecular H$_2$$\lambda$2.4085 along the jet direction, the morphology of ionised [OIII]$\lambda$5007 and displacement of the CO(2$-$1) emission argues for conditions less favourable for star formation in the central $\sim$ 10 parsec region.


INTRODUCTION
Active galactic nuclei (AGN), powered by the accretion of matter onto supermassive black holes at the centers of galaxies (Rees 1984), affect their host galaxies through the feedback processes.They can have an impact on the interstellar medium (ISM) of their hosts via energetic outflows (Venturi et al. 2021).These outflows, driven by radiation pressure, jets or winds from AGN, can occur from accretion disk to galaxy scales (Mukherjee et al. 2018;Morganti et al. 2021), and can inhibit (negative feedback; Maiolino et al. 2012) or enhance star formation (SF) (positive feedback; Nesvadba et al. 2020, Maiolino et al. 2017).Both positive and nega-tive feedback processes are also seen in a single system (Shin et al. 2019;García-Bernete et al. 2021).Outflows affecting SF through the interaction of radio jets with the ISM in the host galaxies are known for large massive galaxies (for reviews, see Fabian 2012;Saikia 2022).Recent observational evidence of dwarf galaxies hosting an AGN (Schutte & Reines 2022) challenges theoretical models that generally invoke supernovae feedback in dwarf galaxies (Koudmani et al. 2022;Nandi et al. 2023).Also, recently, outflows have been observed in an AGN hosted by dwarf galaxies (Manzano-King et al. 2019;Bohn et al. 2021).Available observations of AGN have usually identified the impact of jets on the ISM on kpc or larger scales.To have a clear understanding of the effect of jets on the ISM, one needs to study their impact from sub-pc to kpc and larger scales.There is hardly any observational evidence of jet−ISM interaction and its impact on the host galaxies of AGN on parsec scales.
NGC 4395 is a bulgeless dwarf galaxy at a distance of 4.3±0.3Mpc (Thim et al. 2004) and hosting a radioquiet AGN (Filippenko & Sargent 1989).Its nucleus has the optical spectrum characteristic of a Seyfert 1.8 type AGN (Véron-Cetty & Véron 2006), hosts a black hole (Filippenko et al. 1993) in the mass range of 10 4 −10 5 M ⊙ (Peterson et al. 2005;Woo et al. 2019), is variable in X-ray wavelengths (Vaughan et al. 2005).NGC 4395 appeared unresolved in the Very Large Array (VLA) A-configuration image at 1.4 GHz with a flux density of 1.68 mJy (Ho & Ulvestad 2001).In this paper, we present the first observational evidence of a low power radio jet in NGC 4395 interacting with the ISM of its host galaxy, driving shocks and ionised outflow on parsec scale.Adopting a cosmology of H 0 = 70 km s −1 Mpc −1 , Ω M = 0.7, Ω vac = 0.3 and a distance of 4.3±0.3Mpc (Thim et al. 2004), 1 ′′ in NGC 4395 corresponds to 21 parsec.

DATA REDUCTION AND ANALYSIS
To characterise the jet−ISM interaction on the scale of parsecs, we utilised data from both ground and spacebased telescopes from low energy radio waves to high energy X-rays.The details of the data sets used are given in Table 1.

X-ray
We used four epochs of observations (OBSID: 402,882,5301,5302) carried out by the Chandra X-ray observatory with the advanced CCD imaging spectrometer (ACIS, 0.5−7 keV) for exposures ranging from ∼ 1 ks to ∼ 31 ks.We reduced the data using the Chandra Interactive Analysis of Observations (CIAO,version 4.14) software and calibration files (CALDB version 4.9.8).We first downloaded the data and reprocessed them by running the task chandra repro to generate the cleaned and calibrated event files.Next, we combined all the event files, computed the exposure maps and generated an exposure-corrected image in the default 0.5−7 keV energy range for a total exposure of ∼ 79 ks.We adopted the task merge obs for this purpose.We also rebinned the data by one-quarter of the native 0.492 ′′ per pixel giving an effective resolution of 0.123 ′′ per pixel.The image with a size of 1 ′′ ×1 ′′ is shown in Fig. 1.The red cross in the image is the optical Gaia position.

Optical Imaging
Observations of NGC 4395 carried out by HST WFC3-UVIS2 using a range of filters are available in the HST archives 1 (Proposal ID: 12212, PI: D. Michael Crenshaw).Of these, we used the data in two filters, one F502N, which is centred at 5009.87 Å and the other F547M, which is centred around the nearby continuum at 5756.9 Å.Our aim here is to get the [OIII]λ5007 image from the observation done in the F502N filter, which in addition to the [OIII]λ5007 line emission also contains the continuum emission.One of the ideal ways to remove the continuum emission from the observed F502N narrow-band image is to use observations in medium band filters both blue-ward and red-ward of the F502N filter.In such a scenario, observations in the medium band filters that flank the narrow band filter can be used    Table 1.Details of the data used in this work.These are all archival observations.

Telescope
Filter/Wavelength/Energy FoV Resolution/Synthesised Beam size Exposure time Chandra 0.5−7.0keV  to get the continuum slope.The slope thus obtained can give us an estimate of the continuum that can be subtracted from the F502N data to get the line image.However, in the present case, we have only one medium band filter observation which is adjacent and redward of the F502N filter.Given this, we adopted the following to get the line image.Under the assumption that the continuum emission around [OIII]λ5007 line has a zero slope (which is in agreement with spectra in this case), we selected six source-free regions having sizes of 10 ′′ × 10 ′′ in the F547M and F502N observations, determined the scale factors and applied the mean scale factor (c) and subtracted one filter observation from the other as given below Here, f(N) and f(M) are the brightness in counts/sec in the narrow F502N filter (that contains the [OIII]λ5007 emission line) and medium band filter F547M (that contains the continuum emission).The parameter c is the scaling factor, which is the mean of the ratio of the fluxes in the narrow-band filter, F502N and the medium-band filter, F547M determined from six source free regions.Then we converted the observed [OIII]λ5007 (continuum subtracted) images to flux scale using the KEYWORD PHOTFLAM given in the image headers.The observed medium-band F547M image, the narrow-band F502N image and the continuum-subtracted image are shown in Fig. 2.These images have an FWHM angular resolution of ∼ 0.1 ′′ .

Optical/infrared integral field spectroscopy
We used archival near infrared and optical integral field spectroscopic (IFS) observations obtained with the Gemini and SDSS telescopes.

Gemini
For the optical, we used the archival data from the Gemini Multi-Object Spectrograph (GMOS) under the program ID GN-2015A-DD-6 (PI.Mason Rachel).GMOS with a field of view (FoV) of 5.0 ′′ ×3.5 ′′ covers the spectral range from 4500−7300 Å.In the infrared, we used the archival data from the adaptive optics assisted K-band observations acquired with the near infrared integral field spectrograph (NIFS) under the program ID GN-2010A-Q-38 (PI.Anil Seth).The K-band centered at 2.2 µm covers an FoV of 3.4 ′′ ×3.4 ′′ .The data cubes in GMOS observations have a spatial sampling of 0.05 ′′ /pixel.The spectral resolution is 90 km s −1 and the angular resolution is 0.5 ′′ (Brum et al. 2019).Similarly, the data cubes in the NIFS observations have a spatial sampling of 0.05 ′′ /pixel.The spectral resolution is 45 km s −1 and the angular resolution is 0.2 ′′ (Brum et al. 2019).We reduced the GMOS and NIFS data following standard procedures in IRAF (see Brum et al. 2019 for details).The GMOS data is seeing limited, while NIFS data is AO-assisted.
For fitting the emission lines, we followed a nonparametric approach which involves the removal of the underlying continuum and fitting multiple Gaussian components to the emission lines.In the case of GMOS data, we identified line-free regions on either side of our region of interest, namely the [OIII]λ5007 region (λλ = 4990−5040 Å).We fitted a first-order polynomial to the line-free regions and then subtracted the function from the observations.After continuum subtraction, we fitted the [OIII]λ5007 emission line with two Gaussian components to extract the flux and other properties of the line using the non-linear least square minimization algorithm within Curvefit module of Scipy (Virtanen et al. 2020).An example of the fit is shown in Fig. 3.For the broad Hβ line, we fitted three Gaussian components, one for the narrow AGN component, the second for the broad AGN component and the third for the broad outflowing component.While fitting the broad and narrow AGN components we fixed the peak velocity to be the same, however, the line widths were treated as free parameters and were allowed to vary.For the broad outflowing component, no restriction was imposed, either for the peak of the component or the width of the component.From our fits to the Hβ line, we obtained a σ of 385 kms −1 for the broad Hβ component, which is similar to the value of σ = 334 kms −1 obtained by Brum et al. (2019).The procedure adopted in this work to fit the emission lines is thus appropriate.We fitted the lines [NII]λλ6548,6584 and Hα lines together.For the Hα line we used three Gaussian components similar to that used for the Hβ line.In addition, we used two Gaussians for the two [NII]λλ6548,6584 lines.Here, we tied the widths of the [NII]λλ6548,6584 lines to the width of the narrow component of Hα.For the narrow lines such as [NII]λ5755, [SII]λλ6716,6732 and the H2λ2.4085line from NIFS, we followed the same methodology used for the [OIII]λ5007 line.
From the Gaussian fits to the [OIII]λ5007 line emission in the observed spectra (not corrected for instrumental resolution), we estimated non-parametric values (Zakamska & Greene 2014) such as the velocity (v50, the velocity where the cumulative flux of the line becomes half of the total flux), velocity dispersion (W 90 = v95 − v5, where v95 and v5 are the velocities at which the flux becomes 95% and 5% of total flux) and the asymmetry R = (v95−v50)−(v50−v5) v95−v5 of the line.Also, from fits to the [SII] doublet and using the ratio of the [SII]λ6716 to [SII]λ6731 lines, we estimated the electron density.This line ratio is sensitive to electron densities of the order of ∼ 10 2 -10 4 cm −3 .For estimating the electron density, we assumed an electron temperature of 10,000 K.We calculated the internal extinction E(B−V) from Hα and Hβ line ratio using the following formula (Miller & Mathews 1972;Veilleux et al. 1995) The maps for the velocity and velocity dispersion of the [OIII]λ5007 line emitting gas and for the asymmetry parameter of the line are given in Fig. 4, whereas the E(B-V) map and the electron density map are shown in Fig. 5.The molecular H 2 λ2.4085 image is shown in Fig. 6.These figures cover approximately central 1 ′′ ×1 ′′ region of NGC 4395.This is because the radio emission at 15 GHz has an extension of ∼ 0.5 ′′ and our aim is to investigate the interaction of the radio jet with the ISM.

SDSS/MaNGA
From the Sloan Digital Sky Survey (SDSS) we used data of NGC 4395 observed as part of the Mapping Nearby Galaxies at Apache Point Observatory (MaNGA) survey.The pixel scale of MaNGA product is 0.5 ′′ ×0.5 ′′ .It covers the wavelength range of 3600 Å to 10000 Å with a spectral resolution (λ/∆λ) of ≈ 2000.We used the spectrum of the central brightest pixel which covers the central 0.5 ′′ ×0.5 ′′ region of NGC 4395, and largely encompasses the ∼ 0.5 ′′ extent (total) of the radio jet, the region of interest in this paper to investigate the radio jet−ISM interaction.By using the advantage of this wavelength region, we detected shock sensitive lines [OIII]λ4363, HeIIλ4886 (which are beyond the limit of GMOS), [OIII]λλ4959,5007, Hβ, [NII]λ5755 and [NII]λλ6548,6584 lines (as shown in Fig 7) and estimated the parameters.These parameters are thus a representation of the physical conditions in the central 0.5 ′′ ×0.5 ′′ region.We fitted the emission lines in the same way as explained in Section 2.3.1 and estimated the emission line fluxes.

ALMA
We used the archival data, observed with the highresolution Atacama Large Millimeter/submillimeter Array (ALMA) with 12-m antennas (Data-ID: 2017.1.00572.S, PI: Davis, Timothy).The observations were carried out on March 22, 2018, and January 23, 2019, with ALMA band 6 in the frequency range of 227.47−246.43GHz.The on-source integration time was 2037 s and 423 s, respectively.During the observations, a total of 46 antennas were used, with a minimum baseline of 15.1 m and a maximum baseline of 783.5 m.For the observations on both days, the quasar J1221+2813 was observed as a phase calibrator, and J1229+0203 was observed as a flux density and bandpass calibrator.
We reduced the data using the Common Astronomy Software Application (CASA v5.4.1) with the standard data reduction pipeline of the ALMA observatory.We show in Fig. 8   be 274±21 µJy beam −1 and 287±41 µJy respectively.These results from an independent analysis are also in agreement with those of Yang et al. (2022).We used the task TCLEAN to generate the spectral data cubes.The CO line maps are shown in Fig. 8 (lower panels).From both the observations, we found the peak of the CO(2−1) emission to be displaced by around 0.9 ′′ (∼ 20 parsec) from the nucleus (as determined by Gaia) of NGC 4395.

VLA
The source was observed with the VLA Aconfiguration at 15GHz (PI: Payaswini Saikia, Legacy ID: AS1409).We reduced the data using standard procedures that include flagging bad data using CASA (see Saikia et al. 2018 for details).The beam size obtained is 0.129 ′′ ×0.124 ′′ with a PA of −18 deg.The final image at 15 GHz has an rms noise of 11 µJy/beam.The contours of the 15-GHz image shown in the left panel of Fig. 9 are at 0.03, 0.05 and 0.10 mJy/beam.The western jet component, has a peak flux density of 44.6±1.0 µJy/beam.Though it is fainter related to the core and eastern components, it is detected at about the 4σ level.
The source was also observed at 4.8 GHz (C band) in the VLA B-configuration (PI:J.S. Ulvestad, Legacy ID: AU079).We reduced this data using standard procedures in AIPS by using 3C286 as the flux density calibrator and 1227+365 as a phase calibrator.We achieved an rms noise of 48 µJy/beam.The beam size in the final reduced image is 1.75 ′′ ×1.19 ′′ along a position angle of 89 deg.The final images at 15 GHz and 4.8 GHz are shown in Fig. 9.

Radio morphology
The VLA 15 GHz image (Fig. 9, left panel) showed the source to be a triple with E being the eastern component of the triple (Saikia et al. 2018).The weaker central component of the triple is coincident with the optical Gaia position, while component E is displaced from it by about 220 mas, corresponding to a projected distance of 4.6 parsec.The western component (W) is separated from the central component by 4.2 parsec.However, the overall projected extension of the source is ∼ 10 parsec.The source is also highly asymmetric in brightness, the ratio of peak brightness of components E to W is 3.8.Component E has a spectral index, α, of −0.64 ± 0.05 (S∝ ν α ) and a brightness temperature, T B of (2.3±0.4)×10 6K, showing it to be a non-thermal source.For the central feature α = −0.12± 0.08, and non-detection of a sub-parsec scale compact component sets T B < 5.9×10 5 K (Yang et al. 2022).Radio cores being resolved out in low-mass AGN when observed with milliarcsec resolution has been reported earlier (Nyland et al. 2017).Variability or episodic nuclear jet activity could also contribute to the non-detection of a core.
The triple structure is strongly reminiscent of bipolar jet ejection in radio-loud AGN, and we suggest that the outer components (W and E) are formed by weak radio jets from the intermediate-mass black hole, and refer to the central component as the radio core.We refer to W and E, the end-points of the radio emission as jets in this paper to explore jet−ISM interaction.Low power radio jets (P < 10 42 erg s −1 ) can have a significant effect on the ISM of the host galaxy, interacting with clouds of gas and heating the gas, entraining ambient gas, losing collimation and sometimes forming arc-like fronts (Rampadarath et al. 2018).

Nature of radio emission in the central 10 parsec region
The radio emission observed in the central parsecscale region of a dwarf galaxy could be from a variety of physical processes, such as low-power jets, AGNdriven wind, SF, coronal activity and free-free emission from thermal gas (Panessa et al. 2019).Radio structure, spectral index, polarization characteristics of the radio emission if detectable in future, spatial correlation of radio structure and different gaseous components, spectral line diagnostics of the different components of the ISM could provide valuable clues in identifying the dominant processes.In NGC 4395, the radio morphology of a triple radio source clearly shows signs of interaction on the eastern side which appears to bend sharply, as seen in the high-resolution image (see Yang et al. 2022), indicating jet−ISM interaction.This was reinforced by a detailed study of the line-emitting gas.The [OIII]λ5007 emission appears closely associated with the radio source, ionized by shocks as the jets flow outwards.Line-ratio diagnostics, estimation of gas temperatures from line ratios, and comparison of line ratios with theoretical predictions using MAPPINGS models, all showed shocks to be the dominant process responsible for the ionization.The eastern component which is more promi-nent at radio wavelengths showed stronger signs of interaction with the ISM than the western component.
The spectral index (inclusive of all the three components) derived over the 1.4−15 GHz range (see Fig. 10) gives a value of α = −0.51±0.11(S ν ∝ ν α ) and α = −0.61±0.01excluding 8.4 GHz point.We note here that the flux density measurements are not simultaneous and could be affected by the variability of the central AGN.However, considering only the eastern jet/component, Yang et al. (2022) obtained a value of α = −0.64±0.05.This is very close to the theoretical injection spectral index (Kirk et al. 2000).Considering the source to have a spectral index in the range −0.51 to −0.64, the inverse Compton scattering of the CMB photons and radio photons can give rise to a power law X-ray spectrum whose photon index, Γ can be 1.51 to 1.64.This is close to the value of Γ = 1.67 found by Kammoun et al. (2019) from an analysis of XMM data in the 2−10 keV band.
The above considerations all show that the radio emission in the central 10 parsec region in NGC 4395 is from a low-power jet launched by an intermediate-mass black hole.

Radio and [OIII]λ5007 emission
Fig. 11 shows the [OIII]λ5007 map of NGC 4395 over a region of 1 ′′ ×1 ′′ in the total line emission (left panel), the narrow line component (middle panel) and the broad outflowing line component (right panel).Also, overplotted in these figures are the 15 GHz radio contours in green and the [OIII]λ5007 HST emission in black.The broad outflowing component of [OIII]λ5007 emission is brighter in the eastern side, where the radio emission also tends to be brighter.From these figures, it is evident that the total [OIII]λ5007 emission is prevalent over the entire extent of the radio emission, with the peak of the total [OIII]λ5007 emission coinciding with the peak of the central 15 GHz emission and the outflowing component of the [OIII] line emission is displaced towards the east and overlaps with the eastern radio component.We note here that the peak of the [OIII]λ5007 flux from GMOS and that from HST are comparable, with values of (1.56±0.08)×10−15 erg s −1 cm −2 and (1.25±0.06)×10−15 erg s −1 cm −2 respectively.The pixel scales of the [OIII]λ5007 images from GMOS and HST are 0.05 ′′ and 0.04 ′′ , respectively.The HST [OIII]λ5007 emission is asymmetric, has a cone-like structure with a convex shape near the terminal points of the radio jet, resembling bow shocks in FRII radio galaxies (Kaiser & Alexander 1999).The eastern [OIII]λ5007 cone has a narrower opening angle, while the western cone has a wider opening angle.Also, the western [OIII]λ5007 cone has a more pronounced convex shaped morphology compared with the eastern one.The asymmetry in the [OIII]λ5007 emission could be due to the jets passing through an inhomogeneous medium.The eastern jet is brighter in the radio band, travelling in a denser and larger E(B−V) medium (see Fig. 5), could have ionized the gas, leading to dimmer [OIII]λ5007 emission and brighter synchrotron emission because of shock-induced compression.Similarly, the western jet is travelling in a less dense medium, with smaller E(B−V) values, and enhanced [OIII]λ5007 emission.The observed morphology of the source in radio and [OIII]λ5007 is unambiguous evidence for the interaction of the radio jets with the ISM of the host of NGC 4395 and is the first structural evidence of jet−ISM interaction operating on scales ∼ 10 parsec in a triple radio source.

Multi-wavelength structure of NGC 4395
The 15 GHz image (Fig. 9) shows the highly asymmetric triple structure discussed earlier.In luminous radio galaxies, the components seen on the side of the jets that interact with a dense cloud in the ISM are usually nearer and brighter (O'Dea & Saikia 2021; Saikia et al. 2003), as there is greater dissipation of energy on this side and the dense clouds inhibit the advancement of the jets.In the case of NGC 4395, the brighter component is farther from the nucleus, although its high-resolution radio structure and our optical emission line study suggest the interaction of the jet with the ISM.Therefore a degree of intrinsic asymmetry in the radio jets cannot be ruled out.The velocities of the jets in these lowluminosity AGN are small so that relativistic beaming effects are not expected to be important, as also seen in the case of Seyfert galaxies (e.g.Roy et al. 2000;Whittle et al. 2004;Lal et al. 2011).
Fig. 2 shows the [OIII]λ5007 image of the 1 ′′ ×1 ′′ , from HST.We found the [OIII]λ5007 emission to be prevalent over the central 1 ′′ ×1 ′′ .The central component of the 15 GHz radio emission and the peak of the [OIII]λ5007 image from HST, coincides with the optical Gaia position.The Gaia position is thus the AGN core.The righthand panel of Fig. 2 clearly shows that the [OIII] gas distribution is asymmetric around the optical Gaia position (the AGN core).Also, the [OIII] emission exhibits a convex shaped morphology at the terminal points on either side of the AGN core, though less conspicuous on the eastern side, but more prominent on the western side where the radio emission is weaker.These features suggest that the structure seen in [OIII] line emission could be an ionised outflow, which we explore later.Similar conclusion was also arrived at by Woo et al. (2019).
In Fig. 6 we show the 1 ′′ ×1 ′′ map of the source in molecular H 2 at 2.4085 µm, obtained from NIFS on the Gemini telescope.The molecular H2λ2.4085 is also extended, in the East-West direction and spatially coincident with the 15 GHz radio emission.The 4.8 GHz emission (Fig. 9, right panel) is also spatially coincident with the 15 GHz emission and oriented in the East-West direction.The continuum emission at 237 GHz too (the top panels in Fig. 8) coincides with the central radio source at 15 GHz and the optical Gaia position.However, the CO(2−1) line emission is concentrated at a larger distance (∼ 0.9 ′′ ) from the central nuclear emission (see Fig. 8).The X-ray image (Fig. 1) too shows emission centred around the nuclear region and having extended emission along the East-West direction.

BPT analysis
Emission line ratios in the optical are an essential tool to distinguish between star forming galaxies and AGN.Also, they can be used to disentangle processes that lead to the line emission from SF, AGN and shocks.To measure the emission line fluxes, we fitted line profiles of Hα and [NII]λλ 6548,6584, [SII]λλ 6717,6731, [OIII]λ5007 and Hβ in the spectra of each spaxel, using two Gaussian components for narrow lines and three Gaussian components for the broad Balmer lines Hα and Hβ (see Sect. 2.3.1).The extra component in all lines is to represent the contribution from outflowing gas, while other components are for the broad line region (BLR) and narrow line region (NLR).During the fitting of the [NII] and Hα lines, the line widths of the narrow components were tied together, and the peak fluxes were left free.For Balmer lines (Hα and Hβ) we used the same velocity shift for the narrow component and one broad component, which are responsible for the NLR and BLR region, respectively.While fitting the [SII] lines, the width of these two lines was tied together.We used the [OIII]λ5007/Hβ versus [SII]/Hα as well as [OIII]λ5007/Hβ versus [NII]/Hβ diagnostic diagrams to investigate the physical processes causing the emission lines.These diagnostic diagrams are shown in Fig. 12.Each point in these diagrams represents one spaxel in the 1 ′′ ×1 ′′ region centered around NGC 4395.Here, the red star is the AGN, and the blue To characterise the ionization processes that operate in the central 1 ′′ region of NGC 4395, we carried out a comparison of emission line measurements from the observed GMOS spectra with photoionization and shock models from MAPPINGS-III (Sutherland et al. 2013) and implemented in ITERA (Groves & Allen 2013).The emission lines in the spectra of material photoionized by AGN depend on the ionization parameter U, the slope of the ionizing continuum, β (ϕ ν ∝ ν β ), the gas density and its metallicity.We generated output spectra for a range of input parameters with β ranging from −2 to −1.2 and log (U) varying from −4.0 to 0.0.We assumed solar metallicity (Cedrés & Cepa 2002) and a hydrogen density of n H = 1000 cm −3 .
Similarly, to generate the emission line spectra from shocked material, we used the MAPPINGS-III code again implemented in ITERA.We considered shock velocities (v) between 100 and 1000 km s −1 .The metallicity was assumed as solar consistent with that available in the literature (Cedrés & Cepa 2002), and we considered both pure shock and shock plus precursor models.The magnetic parameter B was allowed to vary between 0.01 to 1000 µG.We show in Fig. 13, the comparison between model line ratios and observed line ratios in the log([OIII]λ5007/Hβ) and log([SII]/Hα) plane for photoionization by AGN (left panel) and photoionization by shocks (right panel).The observed line ratios of the spaxels in the central 1 ′′ ×1 ′′ tend to lie in the region predicted by shock models.Thus, the observations analysed in this work show evidence of shocks contributing to the ionization of the gas in the central region of NGC 4395.This is possible with the hypothesis that the expanding radio jets from the central core, on its interaction with the ISM, leads to the shocks in the medium, which dominate the ionization of the gas over other processes, such as photoionization by AGN.

Electron temperature distribution
Knowledge of the electron temperature (T e ) in the central regions of AGN can help one to constrain the contribution of AGN to gas ionization.Shocks from AGN outflows could produce higher values of T e (Riffel et al. 2021).
To better characterise the spatial nature of T e , we used [NII] lines from GMOS spectra to generate a spatially resolved map of T e .Since [OIII]λ4363 is not covered by the GMOS spectra we used the line ratio R N 2 to generate the T e map.For this we considered only those The green solid curve is from Kauffmann et al. (2003), black and red solid lines are from Kewley et al. (2001).The typical error in these plots is 0.01 dex in both axes.spaxels where the S/N ratio (ratio of the peak of the [NII]λ5755 line to the standard deviation of the pixels in the adjacent continuum) is greater than 30.The T e map is shown in Fig. 14.We found T e to have a range of values, with the value increasing from the center of NGC 4395 outwards, both towards the eastern and western terminal points of the radio jet.This increase in temperature towards the eastern and western sides is evident in the temperature difference map shown in the right panel of Fig 14 .This temperature difference map is generated by subtracting each temperature value from the temperature calculated over the central 0.05 ′′ ×0.05 ′′ region.This difference in temperature is significant, as the error in the temperature estimated using Equations 3 and 4 is typically around 6%−8%.The increase of temperature from the centre of NGC 4395 towards the edges (the difference in temperature is larger than the error in the estimated temperature) coinciding with the radio jet, points to the gas being mostly ionised by shocks.Shocks could be produced by the interaction of the ra-dio jet with the ISM, and this increase of T e from the centre towards the edges is a direct evidence of shock ionisation (Riffel et al. 2021).

Warm ionized gas and shock
The availability of gas reservoirs in the few tens of parsec in the central regions of AGN is an important ingredient in the feeding and feedback processes in them.In particular, the presence of ionized gas in the central regions of AGN is believed to be a consequence of SF as well as AGN activity.Such ionised emission could also be produced by shock excitation.The presence of such ionised gas is easily traced in the optical through emission lines and could trace the effect of AGN and the presence of outflows.From recent IFU observations in the optical and infrared of the central 1 ′′ ×1 ′′ region of NGC 4395, Brum et al. (2019) suggest that these may be ionized by the AGN based on the location of these spatially resolved measurements in the BPT diagram (Baldwin et al. 1981) (in the case of optical) and IR line ratio diagram (in the case of infrared).However, in the zoomed in version of the BPT diagram, the eastern component, the core and the western component nicely gets segregated (see Fig. 12).It is thus likely (similar to that seen in a nearby AGN NGC 1068 by D'Agostino et al. 2019b), the emission in the spaxels within the central 1 ′′ ×1 ′′ region of NGC 4395 could have contribution from AGN, besides shock ionisation as discussed earlier.
Outflows can have multiple constituents, such as the hot ionized gas produced at the shock front as well as neutral and molecular gas entrained in the flow.Shocks produced by AGN-driven outflows and/or radio jet-ISM interaction could also provide the possibility of energetic feedback altering the SF characteristics of the ISM.We consider here the possibility of the shocks leading to the observed morphology of the ionized [OIII]λ5007 We calculated the mass of the outflowing ionised hydrogen from the measured luminosity of the Hα emission using the following relation (Cresci et al. 2017).
By considering F broad (Hα) = 7.40×10 −14 erg cm −2 s −1 (integrated flux density over a circular aperture of radius 0.4 ′′ on the extinction corrected outflowing component of Hα line image from GMOS), and mean electron density, n e = 1700 cm −3 , we obtained M ion out ∼ 652M ⊙ .Using a σ of 123 km s −1 (median σ of outflowing component of [OIII]λ5007 line), we calculated the kinetic energy of this ionised mass as E KE = M out ion (σ 2 ) = 1.97×10 50erg.Taking a velocity of 9 km/s (median of velocity shift of outflowing component of [OIII]λ5007 line) and the projected distance of the tip of the eastern jet as 0.3 ′′ (6.3 parsec), the time required to reach the terminal point is 2.16×10 13 s.The power of the outflow is thus P out = E KE /t = 9.14×10 36 erg s −1 .
We calculated the mass and radius of the NLR using the following relations (Peterson 1997).
parsec (8) By considering n e of 1700 cm −3 (obtained from GMOS observations, see Fig. 5) and assuming a filling factor (ϵ) of 10 −2 (typical upper limit;Peterson 1997) we obtained mass and radius of the NLR of NGC 4395 as 282 M ⊙ and 5.35 parsec respectively over a circular region of 0.4 ′′ radius.
The disk accretion rate is generally represented by the Eddington ratio (λ Edd ) and is defined as Here, L Edd is the Eddington luminosity defined as  Using L Bol of (1.95−4.97)×10 41erg s −1 and M BH values of (9.1×10 3 -3.6×10 5 ) M ⊙ (Woo et al. 2019;Peterson et al. 2005) we obtained λ Edd values of 0.004 to 0.044.
Given the jet power and the bolometric luminosity to be larger than the power of the outflowing ionized emission, the outflow seen in this source on the scale of the NLR of the source could be because of either jet-mode or radiative mode process.The optical spectrum from MaNGA for the central region encompassing the complete core−jet structure having an angular size of 0.5 ′′ , shows the presence of the [OIII]λ4363 and HeIIλ4686 lines (see Fig. 7).The logarithm of the ratio between [OIII]λ4363 and [OIII]λ5007 lines is −1.6;HeIIλ4686 and Hβ ratio is −0.72 and [OIII]λ5007 and Hβ ratio is 0.86.These line ratios point to the presence of shocks (Comerford et al. 2017;Moy & Rocca-Volmerange 2002).
A comparison of the emission line ratios obtained from photoionization and shock modelling and observed line ratios also indicates the gas in the central regions of NGC 4395 to be ionised by shocks (see Fig. 13).Assuming a spectral index (α) of −0.64 (Yang et al. 2022) 12).Also, in the asymmetry of the line versus the velocity dispersion diagram (see Fig. 16; left panel), the spaxels in the eastern jet, occupy a region of higher line asymmetry and higher velocity dispersion, while the western jet occupies a region of lower asymmetry index and lower velocity dispersion.High velocity dispersion and high asymmetry of the lines are attributed to shock excitation (D'Agostino et al. 2019a).The eastern jet thus seems to occupy a region that is dominated by shock excitation, while the western jet seems to occupy a region of weaker shocks.Spaxels in the central 1 ′′ ×1 ′′ region show a tight correlation between the velocity dispersion and the shock sensitive line ratio [NII]/Hα (see Fig. 16; right panel) (Ho et al. 2014).Shock models predict an increase in [NII]/Hα with an increase in shock velocity (Annibali et al. 2010).We show in Fig. 15  Photoionization and shock modelling from MAPPINGS-III (Sutherland et al. 2013), the electron temperature distribution and disturbed kinematics point to the gas in the central region of NGC 4935, excited by shocks.From a multitude of arguments, we conclude that shocks are the dominant process contributing to the excitation of the gas, and such shocks could be due to the interaction of the jet with the ISM in the central 10 parsec region of NGC 4395.

A radio jet-ISM interaction on 10 parsec scale in NGC 4395
From an analysis of data in the optical, infrared, radio and sub-mm bands, we have found evidence of a low-luminosity jet interacting with its host on the scale of about 10 parsec.The eastern jet component which is brighter in the radio band, is resolved in the highresolution High Sensitive Array (HSA) image into two components oriented approximately in the North-South direction, which is nearly orthogonal to the source axis (Wrobel & Ho 2006).This indicates interaction of the jet plasma with the ISM, with the plasma following the path of least resistance.On the eastern side, the [OIII]λ5007 line-emitting gas has higher velocity, higher velocity dispersion and higher asymmetry (see Fig. 16, left panel), possibly due to the shocks associated with the interaction of the radio plasma with the [OIII]λ5007 gas.The weakness of the [OIII]λ5007 emission can either be due to the gas being more ionized or larger extinction, E(B−V) (see Fig. 5) or a combination of both.The weaker jet on the western side has a smaller effect on the [OIII]λ5007 gas with a lower velocity dispersion and asymmetry.This suggests that there may be an intrinsic asymmetry in the oppositely-directed jets.The radio emission is found to exist co-spatially with the emission at other wavelengths such as the hot ionised [OIII]λ5007 emission in the optical band, the warm molecular H 2 λ2.4085 in the infrared band and the 237 GHz emission in the sub-mm band.However, the cold CO(2−1) emission is displaced by ≈1 ′′ from the core.The presence of cold molecular gas is conducive for star formation.The displacement of the CO(2−1) gas, and the paucity of cold gas along the source axis, possibly due to interactions by the jet, has led to the conditions less favourable for SF on 10 parsec scale.A schematic of our proposed coherent picture of the central region of NGC 4395 is shown in Fig. 17  In this work we carried out a systematic investigation of the central region of NGC 4395 using imaging and spatially resolved spectroscopic observations.We summarize our main findings below: 1. From VLA images at 15 GHz, NGC 4395 is found to show a triple radio structure having a projected size of ∼ 10 parsec.The central component of the triple structure is found to coincide with the optical Gaia position which we call as the radio core.
The source is also highly asymmetric in brightness with the eastern component brighter than the western one.
2. The triple radio structure in NGC 4395 is reminiscent of bipolar jet ejection in radio-loud AGN and the eastern and western components of this triple structure are formed by the low power jet (P jet = (1.3 ± 0.3)×10 40 erg s −1 ) powered by the intermediate-mass black hole in NGC 4395.
3. From HST observations we found the [OIII]λ5007 emission to be prevalent over the entire extent of the radio emission with the peak of the [OIII]λ5007 emission coinciding with the optical Gaia position and the radio core.The [OIII]λ5007 emission is asymmetric and shows a convex-shaped structure at the terminal points on either side of the core of NGC 4395 indicating an outflow.This asymmetry in the brightness of [OIII]λ5007 emission could be due to intrinsic asymmetries causing different levels of ionisation or differences in the degree of extinction or a combination of both.
4. The X-ray emission in the 0.5−7 keV band is found to overlap with the radio jet.Similarly, the peak of the continuum emission at 237 GHz is spatially coincident with the radio core.Also, the molecular H 2 λ2.4085 is found to be extended along the radio jet direction and have a close correspondence with the radio emission.
5. From AGN photoionization and shock modelling from MAPPINGS-III and the distribution of the electron temperature distribution, we conclude that the gas in the central region of NGC 4395 is excited mostly by shocks and such shocks could be due to the interaction of radio jet with the ISM in the central parsec region of NGC 4395.This is the first detection of radio jet -ISM interaction at such small spatial scales.
6.The cold CO(2−1) emission is found to be displaced northwards of the radio core by about 1 ′′  (∼ 20 pc).The paucity of cold molecular gas in the central region, possibly due to interactions with the jet makes conditions less favourable for star formation on scales of about 10 parsec in NGC 4395.
The detections of AGN and intermediate mass black holes in a number of dwarf galaxies in recent years have opened the possibility of studying feedback processes in dwarf galaxies.Studies of nearby dwarfs also enable us to probe feedback processes on parsec scales.Schutte & Reines (2022) reported a 150 parsec long ionized filament in the dwarf galaxy Henize 2-10 from HST observations, which connect the black hole region with a site of recent SF.Nyland et al. (2017) reported possible evidence of shock excitation in the nearby dwarf AGN galaxy NGC 404 with an amorphous radio outflow extending over ∼ 17 parsec.NGC 4395 is the clearest example of a dwarf AGN with a triple radio structure, where there is clear evidence of jet−ISM interaction on the smallest scale of ∼ 5 parsec on either side of the core.A comprehensive picture depicting this jet−ISM interaction in the central parsec scale region based on the analysis of images from multiple wavelengths is shown in Fig. 18.This finding will bolster the prospect of finding more such instances in dwarf AGN host galaxies, paving the way for a better understanding of the complex interplay between AGN and their hosts on such small scales in these galaxies.(Nandi et al. 2023).This image has a size of 30 ′′ ×30 ′′ .Also, shown in the same image are two square boxes one of size 5 ′′ ×5 ′′ (black colour) and the other of 1 ′′ ×1 ′′ (white colour).On the top panels are the 4.8 GHz image from the VLA, 237 GHz image from ALMA, the CO(2-1) image from ALMA and the X-ray image in the 0.5−7 keV from Chandra.These images have a size of 5 ′′ ×5 ′′ .The bottom panels show the 15 GHz image from the VLA, molecular H2λ2.4085image from Gemini and [OIII]λ5007 image from HST over a 1 ′′ ×1 ′′ region.In each of these images, the core of NGC 4395 taken as the Gaia position, is shown as a cross.

Figure 1 .
Figure 1.X-ray Image of NGC 4395 in the 0.5−7 keV energy range from Chandra.The red cross is the optical Gaia position.

Figure 2 .
Figure 2. Left panel: The narrow-band image from HST in the F502N filter that contains the emission from both the ionised [OIII]λ5007 gas and the continuum.Middle panel: The continuum image from HST in the F547M filter.Right panel: Difference image after subtraction of the scaled continuum F547M filter image from the narrow-band F502N filter image.This difference image reveals an asymmetric biconical [OIII]λ5007 outflow.The white cross in all the figures is the optical core (Gaia position).

Figure 3 .
Figure 3. Gaussian fit to the observed Hβ (left) and [OIII]λ5007 (right) emission lines along with the residuals (lower panel of each spectrum).Here, shaded blue is the Gaussian fit to the outflowing component, red and black are the Gaussian fits to the narrow and broad components, and the dotted black line is the best fit spectrum.The observed spectra are shown as a solid blue line.

Figure 4 .
Figure 4. Kinematics map of [OIII]λ5007 line emission.Left panel: Velocity map.Middle panel: The map of W90 parameter, which is equivalent to velocity dispersion.Right panel: The map of asymmetry index.The red cross represents the core (optical Gaia position).

Figure 5 .
Figure 5. Left panel: The extinction E(B−V) map.Right panel: The map of electron density calculated from [SII]λλ6717,6731 line ratio.The red cross is the core (optical Gaia position).

Figure 6 .
Figure 6.The molecular H2λ2.4085image of NGC 4395.The core of NGC 4395 taken as the optical Gaia position is shown as a white cross.

Figure 7 .
Figure 7.The rest frame spectrum of NGC 4395 from MaNGA covering the central brightest 0.5 ′′ ×0.5 ′′ .The lines used to calculate the line intensity ratios are marked.
(upper left) the continuum image of NGC 4395 at 237.1227 GHz observed on 22 March 2018 having a synthesized beam size of 0.805 ′′ ×0.469 ′′ along position angle (PA) of 356 deg.From two-dimensional Gaussian fits we found the peak and integrated flux densities to be 93±20 µJy beam −1 and 131±45 µJy respectively.Also, the continuum image at 237.1227 GHz, observed on 23 January 2019 is shown in Fig. 8 (upper right).It has a synthesized beam size of 1.935 ′′ ×1.253 ′′ along PA of 3 deg.From Gaussian fits to the data, we found the peak and integrated flux densities to

Figure 9 .
Figure 9. Left: The VLA radio images at 15 GHz with a size of 1 ′′ ×1 ′′ .The contours are at 0.03, 0.05, 0.10 mJy/beam.The rms noise is 11 µJy/beam.Here, C is the radio core, E is the eastern jet component and W is the western jet component.Right: The 4.8 GHz VLA image with a size of 5 ′′ ×5 ′′ .The contours are at 0.15, 0.20, 0.40, 0.50, 0.60 mJy/beam.The rms noise is 48 µJy/beam.The synthesised beams are shown as white ellipses.These have a size of 0.129 ′′ ×0.124 ′′ with PA of −18 deg (left panel) and 1.75 ′′ ×1.19 ′′ with PA of 89 deg (right panel).The red cross in both the panels represents the optical Gaia position.

Figure 10 .
Figure10.Plot of flux density against frequency.The solid lines are linear least squares fits to the data.The orange line is the fit to all the data points, while the green line is the fit excluding the 8.4 GHz measurement.

Figure 11 .
Figure 11.GMOS [OIII]λ5007 image in colour scale along with the radio emission at 15 GHz (green contours) with contour levels of 0.03, 0.08, 0.1 mJy/beam and HST [OIII]λ5007 emission (black contours) with contour levels of (0.05, 0.1, 0.2, 1.0)×10 −15 erg s −1 cm −2 .The left panel shows the GMOS total flux in [OIII]λ5007, the middle panel shows the GMOS flux in the narrow component of [OIII]λ5007 and the right hand panel shows the GMOS flux in the broad component of [OIII]λ5007.The red cross is the core (optical Gaia position).The green circle is the synthesised beam of 15 GHz data with a size of 0.129 ′′ ×0.124 ′′ along PA −18 deg.and cyan triangles represent the spaxels in the eastern and western jet components.Though all the spaxels lie in the AGN region of the Baldwin, Phillips and Terlevich (BPT; Baldwin et al. 1981) diagram, there is a clear segregation between the core, the eastern and the western components.

Figure 12 .
Figure 12.The position of the spaxels belonging to the central 1 ′′ ×1 ′′ region of NGC 4395 in the line ratio diagnostic diagrams.The green solid curve is fromKauffmann et al. (2003), black and red solid lines are fromKewley et al. (2001).The typical error in these plots is 0.01 dex in both axes.

Figure 13 .
Figure 13.Comparison between predictions of the line ratios due to photoionization by AGN (left panel) and shocks (right panel) and the observed line ratios.The clustered black points are the observed data points in central 1 ′′ ×1 ′′ .The model grids are shown for different shock velocities, magnetic fields and ionization parameters.The typical error of the data points in these plots is 0.01 dex in both axes.

Figure 14 .
Figure 14.Left panel: Temperature map from [NII] line ratio.Right panel: Temperature difference map with respect to the central spaxel.The red cross is the optical Gaia position.

Figure 15 .
Figure 15.Temperature plot.Here, the yellow point refers to NGC 4395.The blue curve is the AGN photoionization grid.The black and red points are from Dors et al. (2020) and Riffel et al. (2021) respectively.
for the eastern jet component we derived the Mach number M s = 2α−3 2α+1 (Al Yazeedi et al. 2021) of the shock as M s =3.91.In the line ratios diagnostic diagrams, such as the [OIII]λ5007/Hβ versus [SII]λ6717,6731/Hα and [OIII]λ5007/Hβ versus [NII]λ6584/Hα diagrams, though all the spaxels lie in the region occupied by AGN, the structure is clearly delineated (see Fig. the position of NGC 4395 in the T e[N II] versus T e[OIII] diagram estimated from MaNGA spectrum.In the same diagram, there are measurements for few AGN along with predictions from AGN photoionization from MAPPINGS-III.AGN NGC 4395 lies in a distinct position in this Figure, significantly deviant (inclusive of errors in the temperature measurement) from the postion occupied by sources photoionized by AGN, pointing to such high temperatures being produced by shocks. .

Figure 17 .
Figure 17.Schematic diagram of our proposed scenario in the inner region of NGC 4395.The jet on its travels outwards from the central radio core, interacts with the medium and ionizes the gas via shock excitation.The radio core coincides with the optical Gaia position, the peak of the [OIII] emission and the peak of the 237 GHz emission.Ionised[OIII] has a cone-like structure, with the radio jet along the axis and causing the outflows.The CO(2−1) gas is located northwards by ∼ 1 ′′ from the radio core.While ionised [OIII]λ5007, warm molecular H2λ2.4085emissions are along the jet, there is a lack of cold CO in the vicinity which is possibly due to interactions with the radio jet.As cold gas is needed for SF process, the lack of cold gas naturally leads to conditions less favourable for SF at scales of ∼ 10 parsec close to the center of NGC 4395.

Figure 18 .
Figure18.The comprehensive picture of the jet−ISM interaction in the central ∼ 10 parsec region of NGC 4395.In the center is the image of NGC 4395 in the NUV filter observed with UVIT(Nandi et al. 2023).This image has a size of 30 ′′ ×30 ′′ .Also, shown in the same image are two square boxes one of size 5 ′′ ×5 ′′ (black colour) and the other of 1 ′′ ×1 ′′ (white colour).On the top panels are the 4.8 GHz image from the VLA, 237 GHz image from ALMA, the CO(2-1) image from ALMA and the X-ray image in the 0.5−7 keV from Chandra.These images have a size of 5 ′′ ×5 ′′ .The bottom panels show the 15 GHz image from the VLA, molecular H2λ2.4085image from Gemini and [OIII]λ5007 image from HST over a 1 ′′ ×1 ′′ region.In each of these images, the core of NGC 4395 taken as the Gaia position, is shown as a cross.