CAHA/PPAK Integral-field Spectroscopic Observations of M81. I. Circumnuclear Ionized Gas

CAHA/PPAK Integral-field Spectroscopic Observations of M81. I. Circumnuclear Ionized Gas Zongnan Li 1,2 , Zhiyuan Li 1,2 , Rubén García-Benito 3 , and Shuai Feng 4,5 1 School of Astronomy and Space Science, Nanjing University, Nanjing 210023, Peopleʼs Republic of China; [email protected],[email protected] 2 Key Laboratory of Modern Astronomy and Astrophysics, Nanjing University, Nanjing 210023, Peopleʼs Republic of China 3 Instituto de Astrofísica de Andalucía (CSIC), P.O. Box 3004, E-18080 Granada, Spain 4 College of Physics, Hebei Normal University, 20 South Erhuan Road, Shijiazhuang, 050024, Peopleʼs Republic of China 5 Hebei Key Laboratory of Photophysics Research and Application, 050024 Shijiazhuang, Peopleʼs Republic of China Received 2021 November 15; revised 2022 February 2; accepted 2022 February 17; published 2022 March 30 Abstract Galactic circumnuclear environments of nearby galaxies provide unique opportunities for our understanding of the coevolution between supermassive black holes and their host galaxies. Here, we present a detailed study of ionized gas in the central kiloparsec region of M81, which hosts the closest prototype low-luminosity active galactic nucleus, based on optical integral-field spectroscopic observations taken with the CAHA 3.5 m telescope. It is found that much of the circumnuclear ionized gas is concentrated within a bright core of ∼200 pc in extent and a surrounding spiral-like structure known as the nuclear spiral. The total mass of the ionized gas is estimated to be ∼2×10 5 M e , which corresponds to a few percent of the cold gas mass in this region, as traced by co-spatial dust extinction features. Plausible signature of a biconical outflow along the disk plane is suggested by a pair of blueshifted/redshifted low-velocity features, symmetrically located at ∼120–250 pc from the nucleus. The spatially resolved line ratios of [NII]/Hαand [OIII]/Hβdemonstrate that much of the circumnuclear region can be classified as a LINER. However, substantial spatial variations in the line intensities and line ratios strongly suggest that different ionization/excitation mechanisms, rather than just a central dominant source of photoionization, are simultaneously at work to produce the observed line signatures. Unified Astronomy Thesaurus concepts: LINER galaxies (925);Galaxy nuclei (609);Warm ionized medium (1788) 1. Introduction Galactic circumnuclear environments, in which the multiphase interstellar medium (ISM)and various stellar populations are coupled under the strong influence of a supermassive black hole (SMBH), if present, are crucial for our understanding of the coevolution of SMBHs and their host galaxies. The circumnuclear ISM, itself supplied by externally acquired material and/or local stellar ejecta, is the necessary fuel for the SMBH to become an active galactic nucleus (AGN). Meanwhile, the circumnuclear ISM also provides sensitive probes to nuclear activities, chiefly through their spectroscopic and kinematic behavior. The recent advent of integral-field spectroscopy (IFS) surveys has enabled systematic studies of nearby galaxies based on large sets of optical spectra, providing new insights in particular to the circumnuclear environments. Low ionization nuclear emission-line regions (LINERs; Heckman 1980), which are characterized by prominent low-excitation emission lines of weakly ionized atoms, have been identified in a large number of inactive galaxies by the pioneer Spectrographic Areal Unit for Research on Optical Nebulae survey (Sarzi et al. 2010), and later by the Calar Alto Legacy Integral Field spectroscopy Area (CALIFA; Sánchez et al. 2012), SydneyAAO Multi-object Integral-field spectrograph (SAMI; Bryant et al. 2015)and Mapping Nearby Galaxies at Apache Point Observatory (MaNGA)surveys (Belfiore et al. 2016). A variety of physical mechanisms have long been proposed for the ionization and excitation of LINERs, which include photoionization by a low-luminosity AGN (LLAGN; Ferland & Netzer 1983; Halpern & Steiner 1983), massive young stars (e.g., OB stars and Wolf-Rayet stars; Filippenko & Terlevich 1992;Barth& Shields 2000), hot evolved stars (e.g., planetary nebulae and post-asymptotic giant branch [pAGB]stars; Binette et al. 1994; Stasińska et al. 2008; Yan & Blanton 2012; Singh et al. 2013),as well as shocks (Heckman 1980; Dopita et al. 1997)and cosmic rays (Ferland & Mushotzky 1984). The leading role of AGN photoionization in LINERs is revisited by the IFS observations and often called into question due to insufficient ionizing photons from an LLAGN (Eracleous et al. 2010;Belfiore et al. 2016). More recently, higher-resolution Gemini Multi-Object Spectrograph (GMOS)IFS observations and Hubble Space Telescope (HST)slit observations (Ricci et al. 2015b; Molina et al. 2018)of a small samples of LINERs also suggest that radiation from the central LLAGN is short for the ionization, and that other mechanisms, such as pAGB stars and shocks induced by LLAGN-driven jets/outflows should play a significant role. In addition, gas motions deviating from a pure rotating disk are discovered in LINERs, both by HST observations (Walsh et al. 2008)and by MaNGA on more extended regions (Rodríguez del Pino et al. 2019), suggesting the presence of outflows. Clearly, a spatially resolved view of the spectroscopic and kinematic properties of the circumnuclear ionized gas is key to understanding the universality and relative contributions of various ionization/excitation mechanisms. At a distance of 3.6 Mpc (1″corresponds to 17.5 pc; Freedman et al. 2001), M81 hosts the nearest prototype LLAGN, known as M81 * (Ho 1999), which has a bolometric The Astrophysical Journal, 928:111 (14pp), 2022 April 1 https://doi.org/10.3847/1538-4357/ac56d9 © 2022. The Author(s). Published by the American Astronomical Society. Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s)and the title of the work, journal citation and DOI. 1 luminosity of 2 ×10 41 erg s −1 , corresponding to an Eddington ratio of 3 ×10 −5 (Nemmen et al. 2014)for an estimated SMBH mass of 7 ×10 7 M e (Devereux et al. 2003). Thanks to its proximity, the central region of M81 has been extensively investigated in multiwavelengths and classified as both an LINER (Heckman 1980)and type-1 Seyfert (Peimbert & Torres-Peimbert 1981), the property of which was reviewed in Ho et al. (1996). AGN activities are evidenced by a variable, power-law X-ray continuum (Ishisaki et al. 1996), a featureless, compact UV continuum (Ho et al. 1996; Devereux et al. 1997), optical double-peaked broad emission lines (Bower et al. 1996), and a compact radio core plus a one-sided, precessing jet (Iyomoto & Makishima 2001; Martí-Vidal et al. 2011). M81 thus provides us with one of the best laboratories for studying the interplay between an LLAGN and its close environment. M81 * is found to interact vividly with the its environment on various scales. GMOS IFS observations have found signatures of both an inflow of ionized gas at a rate of 3 ×10 −3 M e yr −1 and an outflow vertical to a rotating ionized gas disk in the central 120 ×250 pc region (Schnorr Müller et al. 2011). This outflow was reinterpreted by Ricci et al. 2015a as in line with the radio jet. More recently, a high-velocity outflow of hot gas has been detected in the central 40 pc region with Chandra X-ray spectroscopy (Shi et al. 2021), which is most likely driven by the hot accretion flow onto the SMBH. Evidence for this hot wind interacting with the circumnuclear ISM was also suggested from detection of diffuse hot gas. On larger scales, M81 resides in a well-known galaxy group, where the main members, M81, M82, and NGC 3077 are actively interacting with each other. High-resolution, wide-field H ɪobservations have found that these galaxies share a common atomic gas envelop (Yun et al. 1994; de Blok et al. 2018), and are connected with massive H ɪfilaments. Arguably undergoing a merger with its satellite M82, M81 is fueled by tidally disrupted materials from both M82 and NGC 3077 (Smercina et al. 2020). Such interactions could have significantly affected the nuclear activity in M81 in the recent past (or may do so in the near future). In the optical band, an Hα-bright gaseous structure across the central few hundred parsecs is the most conspicuous feature in the circumnuclear region of M81 (Devereux et al. 1995). This structure has an organized morphology characterized by spiral-like filaments (hence dubbed a nuclear spiral)extending into the close vicinity of the SMBH, raising the interesting question of whether it could be responsible for fueling the AGN activity. The ionization mechanism of this structure remains uncertain, although massive stars can be safely ruled out due to their current absence in this region (Devereux et al. 1997). The LLAGN M81 * is a promising source of photoionization, especially for the innermost region (Ho et al. 1996). However, it remains to quantify the role of the LLAGN in outer regions and to examine possible contributions from other mechanisms mentioned in the above. In this regard, the case of M81 can be highly instructive when compared to the mysterious case of M31, which hosts the nearest known LINER in the clear absence of either an LLAGN or massive stars (Li et al. 2009). Although high-resolution spectroscopic observations, including those afforded by GMOS and HST (e.g., Devereux et al. 2003; Schnorr Müller et al. 2011; Devereux 2019), have been carried out to resolve the properties of ionized gas in the inner ∼100 pc down to the central 1 pc of M81, sensitive IFS observations covering a wider area would still be critical to reveal the full connection between the SMBH and its circumnuclear environment. In particular, IFS observations hold promise for understanding the ionization mechanism(s)of the nuclear spiral in M81 through emission-line diagnostics, which would help to resolve the enigma of LINERs in general. We are thus motivated to carry out the first optical IFS observations with a full coverage of the central kiloparsec region of M81, utilizing the same instrument and a similar observational setup as the CALIFA survey, which satisfy the above requirements. In this first paper of a series, we provide an overview of the observations and present the properties of the circumnuclear ionized gas as traced by their major emission lines. The properties of the stellar population, and a detailed investigation of the gas ionization mechanism will be presented in subsequent papers. In Section 2, we describe the observations and data reduction procedure. The morphology and kinematics of the ionized gas derived from the emission lines, as well as the gas properties inferred from the line ratios are presented in Section 3. Some interesting implications of the observational results are addressed in Section 4, followed by a summary in Section 5. 2. Observation and Data Reduction The spectroscopic observations of M81 were taken with the PPAK integral field unit (IFU)of the Potsdam Multi-Aperture Spectrograph instrument at the 3.5 m telescope of the Centro Astronómico Hispano Alemán (CAHA)at Calar Alto. The observations were carried out on 2017 November 19 using the medium-resolution V1200 grating (3650–4840 Å; FWHM 2.3 Å)and on 2018 April 19 using the low-resolution V500 grating (3745–7500 Å; FWHM 6 Å), respectively. These settings are equivalent to those of the CALIFA survey. A combined spectral cube (called COMBO)of the two gratings were produced after convolving the V1200 resolution to match that of the V500. The COMBO cube solves the problem of vignetting that otherwise affects the blue end of the V500. The final wavelength coverage ranges from 3700–7300 Å, corrected for the heliocentric velocity and vignetting at the spectral edges. The field-of-view (FoV)is approximately a hexagon of 74″×64″, sampled by 2 7fibers. A three-pointing dithering scheme was adopted to ensure a full FoV coverage of the spectral cube, which was resampled to have a pixel size of 1″ (∼17.5 pc at the assumed distance of M81). We carried out V1200 observations with an exposure time of 3600 s per pointing (split into three individual exposures), while for the V500 setup we obtained observations with a total of 2100 s per pointing (split into seven individual exposures to avoid saturation).TheFoVis illustrated in Figure 1against an optical image of M81 combining HST and Subaru Telescope observations. 6 Data reduction procedure followed the standard automatic pipeline of CALIFA (Sánchez et al. 2012). The main steps are summarized below: (1)CCD frame alignment and cosmic-ray removal; (2)spectral extraction; (3)sky subtraction; (4)flux calibration; (5)spatial rearrangement and image reconstruction; and (6)atmospheric refraction and extinction correction. Both atmospheric and Galactic extinction have been corrected, which is 0.14 and 0.25 mag in the Vband, respectively. For more details of the data reduction procedure readers are 6 Copyright 2016 Robert Gendler, Roberto Colombari, Hubble Legacy Archive, Subaru Telescope. 2 The Astrophysical Journal, 928:111 (14pp), 2022 April 1 Li et al. referred to Husemann et al. (2013), García-Benito et al. (2015), and Sánchez et al. (2016). An absolute flux calibration was not performed, as the flux accuracy is empirically determined to be within ∼8% with respect to the Sloan Digital Sky Survey (SDSS; Sánchez et al. 2012). To determine the properties (e.g., line-of-sight velocity, velocity dispersion, and metallicity)of the stellar populations, the observed spectra were fitted pixel by pixel with the STARLIGHT software (Cid Fernandes et al. 2005,2011a),which utilizes synthesized spectra of single stellar populations (SSPs), following García-Benito et al. (2017). The results were then processed through PyCASSO (the Python CALIFA starlight Synthesis Organizer; Cid Fernandes et al. 2013;deAmorim et al. 2017)to produce a suite of spatially resolved stellar population properties. Expected significant emission lines, as well as bad pixels, were masked during the fitting. The stellar population synthesis model consists of 254 SSPs from the GRANADA library (González Delgado et al. 2005, for stellar populations younger than 60 Myr)and Vazdekis et al. 2015 based on BaSTi isochrones (for older ages). The metallicity ranges from −2.3 to +0.4 (with eight values, log Z/Z e =−2.28, −1.79, −1.26, −0.66, −0.35, −0.06, +0.25, and +0.40),and the age ranges from 0.001–14 Gyr. A Salpeter initial mass function (IMF; Salpeter 1955)was assumed here. A detailed investigation of the stellar populations will be presented in a future work (R. Garcia-Benito et al. in preparation). Figure 2displays the observed spectrum accumulated over all pixels (blue curve), the cumulative modeled stellar spectrum (red curve)and the model-subtracted residual spectrum (black curve). The latter consists primarily of significant emission lines, including [OII]λ3727, Hγ(λ4341),[OIII]λ4363, Hβ (λ4861),[OIII]λ4959,5007 doublet, [OI]λ6300, Hα(λ6563), [NII]λ6548,6584 doublet, and [SII]λ6716,6731 doublet. Absorption features near 5895 and 6290 Åare due to imperfect subtraction of the stellar continuum or telluric line. On a pixelby-pixel basis, we derived the emission-line parameters, including line flux, line centroid (line-of-sight velocity)and width (velocity dispersion)by fitting each line with a single Gaussian profile. The velocity dispersion was corrected for the instrumental broadening. The noise level varies among different lines and different positions; for reference, the median surface brightness value with signal-to-noise ratio (S/N)∼3 for the Hαline is 2.4 ´ --- - 10 erg s cm arcsec 17 1 2 2 . We employ ancillary images from the HST archive to assist our analysis: (1)a Wide Field Planetary Camera 2 (WFPC2) Hαmap of the central ∼1ʹregion of M81, originally presented by Devereux et al. (1997), and (2)a Wide Field Camera 3 (WFPC3)F438W image covering the central 160″×160″of M81 (project ID 11421), chiefly to quantify the starlight distribution and to reveal dust extinction features co-spatial with the ionized gas. 3. Analysis and Results 3.1. Morphology Unless otherwise stated, in the following we will focus on the four familiar strong emission lines, Hα,Hβ,[NII]λ6584, and [OIII]λ5007, which conventionally provide important diagnostics of the ionized gas. An S/N>5 cut is applied to each of these lines, except for the relatively weak Hβ, for which a lower S/N cut >3 is adopted. The surface brightness distribution of the four lines are presented in Figure 3. All four lines show a highly similar morphology, although Hβappears patchy due to its relative dimness. The morphology is characterized by a bright core (within the central ∼200 pc; see the zoom-in view in Figure 6) surrounded by a lopsided, spiral-like structure with an extent of ∼1 kpc. This spiral structure, most prominent to the northeast and southwest of the core, has already been reported by narrowband imaging and dubbed the nuclear spiral (Devereux et al. 1995). We show below that this prominent structure is also conspicuous in dust extinction. The Hαmap in Figure 3is contrasted with intensity contours derived from the HST Hα image of Devereux et al. (1997), which delineate the nuclear spiral at a higher resolution. Nevertheless, our Hαmap reveals the presence of more diffuse ionized gas across almost the Figure 1. Left: FoV (74 ″×64″, black hexagon)of the PPAK observations overlaid on a composite HST and Subaru image of M81. Right: the continuum image integrated over the range of 3700–7300 Åfrom the PPAK observations. The nucleus of M81, marked by the black cross, is slightly offset toward southeast from the image center. 3 The Astrophysical Journal, 928:111 (14pp), 2022 April 1 Li et al. entire FoV, which is not obvious in the HST narrowband image. We note that two additional strong emission lines, [OII] and [SII], share a similar morphology with Hα. For reference, the intensity maps of [OII]and [SII]are displayed in Figure A1. The high-resolution dust extinction map shown in the right panel of Figure 4is produced based on the HST WFC3/ F438W image (left panel of Figure 4), which covers a region a few times larger than the PPAK FoV. This wide-band, blue image is chosen for its sensitivity to uncovering dust extinction features against the bulge starlight. We utilize the imaging fitting algorithm GALFIT (v3.0.5, Peng et al. 2002,2010)to model the intrinsic starlight distribution, adopting three spatially distinct physical components: a bulge, a disk, and a point source standing for the LLAGN M81 * . The point-spread function (PSF)file suitable for the WFC/F438W image is downloaded from the HST website 7 and fed to GALFIT, along with a mask image that accounts for bad pixels and bright, offnuclear sources. We fix the galactic center at the brightest pixel in the image, and allow the bulge Sérsic index n(=4 for classical bulge, 1 for exponential disk; Sérsic 1963)to vary during the fit. The resultant parameters of the three components are listed in Table 1, and the modeled image combining the three components is shown in the middle panel of Figure 4. The Sérsic index nderived from our fit is 3.23 ±0.01, indicating a classic bulge. The corresponding effective radius is 999 ±5 pc, which is roughly covered by the PPAK FoV and well within the WFC3 FoV. These values are in rough agreement with the values of n=4.09 ±0.48 and R e =1258.46 ±977.40 pc reported by Fabricius et al. (2012), which were obtained by fitting one-dimensional surface brightness profile combining multiple data sources. We caution that the disk effective radius may be underestimated due to the limited WFC3 FoV. The relative contribution of different components within the PPAK FoV is given in Table 1, which shows that the bulge is dominant over the disk. Once the starlight model is subtracted from the WFC/F438W image, dust extinction features are readily revealed, which exhibit the familiar morphology of the nuclear spiral (Figure 4). The apparent dust extinction along a given line of sight can be evaluated as ( ) =- ´ lll A 2.5 log F F ,app ,obs ,int ,whereF λ,obs is the observed flux (e.g., in the F438W band)and F λ,int is the intrinsic flux represented by the above model. Assuming a thin screen geometry, which is reasonable as suggested by the spatial coincidence between the dust extinction features and the nuclear spiral, only the bulge starlight from behind this screen would be attenuated. To account for this effect, we follow Dong et al. (2016)to introduce a parameter (f),i.e.,thefractionof obscured starlight along the line of sight, such that = l F,obs [ ()]´+-´ l -´ l ffF10 1 A0.4 ,in t ,abs , where A λ,abs is the absolute extinction. Here, we consider two special cases: (1)f=1, which occurs when the dusty screen is located in front of the bulge, and (2)f=0.5, which occurs when the dusty screen is embedded in the bulge and seen face-on. The latter case might be closer to reality. We thus calculate the absolute extinction in the F438W band (central wavelength 4325 Å)and convert it into the more conventional V-band (5470 Å)extinction A V ,assuming the Galactic extinction law of Cardelli et al. (1989). The resultant V-band extinction maps for the PPAK FoV are shown in left and right panels of Figure 5,forthecaseoff=1 and 0.5, respectively. The f=0.5 case exhibits systematically higher values than the f=1 case, due to the negative correlation between fand A λ,abs for a given A λ,app . The hydrogen column density N H can thus be inferred to range between ∼10 20 cm −2 to afewtimes10 21 cm −2 in the f=1 case, using the relation N H ;2×10 21 A V (Güver & Özel 2009).Thef=0.5 case results in a higher N H ,reaching10 22 cm −2 in some regions. In Figure 6, the central 30″×30″is highlighted in Hα emission and dust extinction. Under this zoom-in view, the bright core is seen to be connected with arm-like features to the southeast and southwest, as indicated by the red arrows in the figure, which have apparent counterparts in the dust extinction map. These arms might be feeding gas to the galactic nucleus. Strong dust extinction features (indicated by red color)are also seen to the north of the nucleus, possibly with co-spatial Hα emission that connects with the bright core. Figure 2. The observed spectrum (blue curve)and the fitted stellar population model (red curve)stacked over all pixels. The emission-line spectrum (black curve)is obtained by subtracting the model from the observed spectrum. Prominent emission lines are labeled. 7 https://www.stsci.edu/hst/instrumentation/wfc3/data-analysis/psf 4 The Astrophysical Journal, 928:111 (14pp), 2022 April 1 Li et al. Figure 3. Surface brightness distribution of Hα,Hβ,[NII]λ6584, and [OIII]λ5007. A cut of S/N>3 is applied to Hβ, while a cut of S/N>5 is applied for the other three lines. The Hαmap is overlaid with Hαintensity contours derived from the HST narrowband image of Devereux et al. (1997). Position of the nucleus is marked by a black cross. Figure 4. Left: HST WFC3/F438W image covering the inner region of M81. The hexagon delineates the PPAK FoV, and the black cross marks the M81 nucleus. Middle: best-fitGALFIT model composed of a bulge, a disk, and a point-like nucleus. The contours indicate the bulge-to-total flux ratio, at levels of 80%, 60%, and 40% from inside out. Right: the residual image after subtracting the model image, in which dust extinction features stand out with negative values (represented by white color). 5 The Astrophysical Journal, 928:111 (14pp), 2022 April 1 Li et al. 3.2. Kinematics 3.2.1. Broad Lines As demonstrated in previous work (e.g., Bower et al. 1996; Ho et al. 1996), the nucleus of M81 exhibits a broad (FWHM 1000 km s −1 )component in the Hαline, which is a typical signature of the broad-line region of an AGN. The PPAK data allows us to probe the existence of a broad line at regions beyond the nucleus. To do so, we fit the Hαline of the central 9″×9″region with a double-Gaussian profile, one for the narrow component and the other for the putative broad component. In the upper panel of Figure 7, we show the fitted spectrum in the interval around Hαof the innermost pixel as an illustration (black symbols and curves). Two additional Gaussians are included to account for the [NII]doublet, as well as a power law to account for the baseline continuum. In this central pixel, the velocity dispersion of Hαis σ Hα =913 ±3kms −1 for the broad component, and ∼100 km s −1 for the narrow component (comparable to the instrumental broadening, which has been subtracted). Errors have been estimated from 200 iterations of Monte Carlo simulations. This result is in good agreement with the value of 862 ±37 km s −1 obtained from the GMOS spectrum of the central 1 5 (Schnorr Müller et al. 2011). It is found that a broad Hαcomponent is prevalent in this central region, with a radial extent of at least 3″(∼50 pc)in all directions, as illustrated in the lower panel of Figure 7, where the spatial distribution of the best-fit line-of-sight velocity and velocity dispersion of the broad component are plotted. We note that a broad component is not required to provide a satisfactory fit to the [NII]doublet. On the other hand, we find that the [OIII]line in this region also exhibits the signature of a broad component, as illustrated in the upper panel of Figure 7(red symbols and curves), again using the innermost pixel as an example. We thus fit the [OIII] line with a double-Gaussian profile. The resultant broad component of [OIII]has a velocity dispersion of 434 km s −1 , which is about half that of the broad Hα, but is significantly higher than that of the narrow component (∼100 km s −1 ).As evident in the lower panel of Figure 7, the broad component of [OIII]shares the broad Hαwith a similar spatial extent and a trend of radially decreasing intensity, which suggest a common origin between the two lines. However, the broad Hαand [OIII]components are distinct in their kinematics. As can be seen in Figure 7, the Hαis collectively redshifted, with a mean line-of-sight velocity v Hα =197 ±5kms −1 , and has a mean velocity dispersion of σ Hα =940 ±14 km s −1 . In contrast, the [OIII]is collectively blueshifted, with v [O III] =−176 ±5km s −1 , and has a substantially lower velocity dispersion of σ [O III] =454 ±9km s −1 . Here, the line-of-sight velocity has not been corrected for the systemic velocity of M81 (−39 km s −1 , de Blok et al. 2008). A redshifted broad Hαline was also reported based on HST/STIS spectra (Devereux et al. 2003; Devereux & Shearer 2007; Devereux 2019), with v=192 km s −1 and Figure 5. Left: absolute extinction derived from the WFC3/F438W residual image, for the f=1 case. Right: similar to the left panel but for the f=0.5 case, overlaid with the HST Hαintensity contours. The nucleus is marked by the cross. The white stripe is due to the chip gap of WFC3. Table 1 GALFIT Results Component P.A. a R eb Axis Ratio c Sérsic Index Relative Flux d (degree)(pc) Bulge 150.28 ±0.02 999 ±5 0.70 3.23 ±0.01 1.0 Disk 149.67 ±0.03 2472 ±5 0.60 1 0.2 Nucleus LLLL0.002 Notes. a Position angle of the fitted component. b Effective radius. c Projected axis ratio. d Flux relative to that of the bulge component, integrated over the PPAK FoV. 6 The Astrophysical Journal, 928:111 (14pp), 2022 April 1 Li et al. σ=930 ±11 km s −1 . Interestingly, from the STIS spectrum covering the central 0 2 ×0 35, Devereux (2019)reported a smaller velocity dispersion (σ≈250 km s −1 )for the [OIII] λ5007 line, which is intermediate between the values of the broad and narrow components found here. Unfortunately, the [OIII]line was beyond the wavelength coverage of the GMOS IFU observation (Schnorr Müller et al. (2011)). The apparent radial extent of the broad Hαand [OIII]lines is substantially larger than previously known, suggesting the presence of a high-velocity and high-excitation plasma within Figure 7. Upper: spectra of Hα(black asterisks)and [OIII](red crosses)of the innermost pixel. The broad and narrow components of Hαand the [NII]doublet are fitted with four Gaussians (black dashed lines), while the [OIII]doublet are each fitted with a broad plus a narrow component (red dashed lines). The total multiGaussian fit is shown by the solid line, and the residual is plotted. Lower: line-of-sight velocity (left)and velocity dispersion (right)maps of the broad component of Hαand [OIII]in the central 9″×9″, overlaid with intensity contours with levels 1, 3, 7 ×10 −14 erg s −1 cm −2 arcsec −2 for Hα, and 3, 7, 15 ×10 −15 erg s −1 cm −2 arcsec −2 for [OIII]. The black cross marks the central pixel (1″size). Figure 6. Left: zoom-in view of the Hαsurface brightness distribution in the central 30″×30″, overlaid with HST Hαcontours (same as in Figure 3). The two fanshaped regions mark a possible outflow as revealed in the Hαvelocity field (Figure 8), and the red arrows mark the two arm-like features connecting with the core. Right: the same region in dust extinction of the f=0.5 case, highlighting the dust lanes close to the nucleus, marked by a cross. 7 The Astrophysical Journal, 928:111 (14pp), 2022 April 1 Li et al. the central 50 pc of M81. As a cautionary note, the effect of PSF scattering (FWHM 24; García-Benito et al. 2015)of the bright nucleus could have led to an artificially large broadline region. Future IFU observations with a higher angular resolution are needed to confirm this case. 3.2.2. Narrow Lines We now turn to the narrow lines that are prevalent throughout almost the entire PPAK FoV, focusing on the kinematics of the Hαand [OIII]lines. The observed line-ofsight velocity of Hαand [OIII]is shown in the left panels of Figure 8, along with the stellar line-of-sight velocity derived from the continuum model. Overall, the stars show a regular and smooth velocity field with an obvious rotation pattern, while the Hαand [OIII]lines follow this rotation pattern but exhibit a more complex velocity field. We apply a simple analytical model of Bertola et al. (1991), which assumes a circular orbit, to fit the line-of-sight velocity of the stars, as a function of the projected radius (R)and position angle (Ψ, east from north), () () ( ) () () {[()()()] ()} Y= +Y-Y Y-Y + Y-Y + 1 VR V AR i i Ri i , cos sin cos sin cos cos C cos , s p p mod 0 22 022 00 222 where V s is the systemic velocity, Ψ 0 is the position angle of the kinematic major-axis, iis the inclination angle of the orbital plane, Ais the maximum orbital velocity, C 0 is the radius at the Figure 8. The velocity field of stars, Hαand [OIII], from top to bottom. Left: the observed two-dimensional distribution of line-of-sight velocity. Middle: the best-fit kinematic model of the stellar velocity field. The same model is adopted as the base model for Hαand [OIII]. Right: the residual velocity map by subtracting the model from the observed velocity field. The dashed line marks the kinematic major-axis, and the cross marks the galactic nucleus. In the middle and bottom right panels, the two fan-shaped regions outline a putative biconical outflow, and the HST Hαintensity contours are overlaid. The small and large black boxes in Hαand [OIII]maps, respectively, outline regions with a significant redshifted component as shown in Figure 9. 8 The Astrophysical Journal, 928:111 (14pp), 2022 April 1 Li et al. maximum velocity, and pis the slope of the rotation curve outside of C 0 . Due to the moderate FoV, the parameters A,C 0 , and pare degenerate. Fortunately, this degeneracy does not have a significant effect on the other parameters; hence, we will not include them in the following discussion. The best-fit parameters are derived using Markov Chain Monte Carlo simulations, which result in =- - + V 56.60 s0.24 0.18 km s −1 ,= - + i34. 8 1.2 1.2, and Y = - + 153. 22 00.15 0.14. The systemic velocity is somewhat more blueshifted than the canonical value of −39 ±3kms −1 , which was obtained from the H ɪ kinematics primarily tracing the outer disk of M81 (de Blok et al. 2008). The inclination angle is also significantly lower than the value of 58°estimated from the axis ratio of the M81 disk. 8 This may be understood as the velocity field in this region being dominated by the bulge rather than the disk (Table 1). Nevertheless, the excellent agreement between the fitted position angle of the kinematic major axis (marked by the dash line in the upper panels of Figure 8)and that of the disk (153°; de Vaucouleurs et al. 1991)suggests that we have arrived at a reasonable phenomenological model of the stellar velocity. This is further supported by the the difference map between the observed and modeled stellar velocity field (the upper right panel of Figure 8), which shows no significant residual patterns. From the optical image of M81 shown in Figure 1, which exhibits more prominent extinction features on the southwestern side of the disk, we can conclude that this side is the nearside. Accordingly, it can be inferred that the disk rotates counterclockwise, which is consistent with a trailing spiral arm pattern. We then examine potential kinematic signatures of the ionized gas beyond the coherent rotation. For this purpose we subtract the above stellar kinematic model from the observed gas velocity field. This implicitly assumes that to zeroth order the bulk of the ionized gas has the same underlying rotation pattern as that of the stars. This is verified by fitting an independent model of Equation (1)to the Hαand [OIII] velocity fields, both of which result in best-fit parameters that are in rough agreement with the stellar velocity field. Nevertheless, we take the latter as the underlying base model because it is free of complex kinematic features. The residual maps are shown in the right panels of Figure 8, which signify the nonrotating components of the velocity field. It is evident that the residual velocity field is highly similar between the Hαand [OIII]lines, and perhaps not surprising because the same base velocity field has been subtracted. Perhaps the most interesting pattern in the residual velocity map is a pair of redshifted/blueshifted arc-shaped features roughly symmetric about the nucleus, as outlined in the middle right and lower right panels of Figure 8. Both features are located at a projected radius of ∼120–250 pc from the nucleus and have an absolute line-of-sight velocity of ∼50 km s −1 with respect to the nucleus. A plausible explanation for these arcs is that they are the manifestation of a biconical outflow from the nucleus, which lies predominantly in or near the disk plane, such that the blueshifted component is found at the southwestern side (nearside)and the redshifted component is found at the northeastern side (farside). It is noteworthy that these two arcs are not part of the nuclear spiral, rather they fall in between the core and the nuclear spiral (Figure 6). Alternatively, a bipolar inflow along the disk rotation axis would produce similar velocity patterns, which, however, appears physically implausible. Noncircular motions caused by a nuclear bar or triaxial bulge could be another possibility, but the observed velocity appears too high to be compatible with typical pattern speed of bars (20–50 km s −1 kpc −1 ; Li et al. 2015). Support for an outflow comes from Schnorr Müller et al. (2011)and Ricci et al. (2015a), who both analyzed the velocity field in the central 10 pc using the principle component analysis tomography (Steiner et al. 2009), and interpreted one of the components (PC 4)as a conical outflow of ionized gas. Also prominent in the residual velocity map are redshifted features at the southwestern and northwestern corners of the FoV (appearing in red color), as well as blueshifted features at the northeastern edge. Judging from the dust extinction map (Figure 5), the redshifted features might be part of the outer nuclear spiral. These features may signify inflowing motions along the disk plane. We caution that the line emission in these regions are relatively weak, which may be more vulnerable to systematic effects. Future IFU observations of a larger FoV are warranted to test and better quantify the putative inflowing motion associated with these features. Figure 9displays the distribution of velocity dispersion of Hαand [OIII]. The two lines show overall smooth and similar distributions, with a comparable level of mean velocity dispersion (∼120 km s −1 ; also comparable to the level of instrumental broadening, which has been subtracted). We note that typical uncertainty of the velocity dispersion is 5kms −1 in the nuclear spiral, and somewhat larger (∼15–20 km s −1 )in the outer regions with lower S/N. There is no significant feature in the central region after excluding the broad component. A region of high-velocity dispersion, with values 250 km s −1 , is seen in the northwestern side (outlined by the black rectangle)of the [OIII]map. The stacked spectra of [OIII]λ5007 and Hα+[NII]of pixels having σ [O III] >250 km s −1 in this region are presented in the lower right panel of Figure 9. Some excess is seen on the red wing of both Hαand [NII]λ6584, possibly in accordance with the broad [OIII]. The physical origin of this feature is unclear and deserves further exploration. A compact feature of high-velocity dispersion is additionally seen at the eastern side of the Hαmap (marked by the small white box). This clump-like feature is also prominent in the Hα velocity map (Figure 8), which is redshifted with respect to its vicinity. The stacked spectra of this region are shown in the lower left panel of Figure 9. It can be seen that there is a second velocity component redshifted by ∼500 km s −1 from the primary component (at ∼−130 km s −1 )of Hα, which is however not seen in the [OIII]line. The location of this clump coincides with a molecular cloud identified by Casasola et al. (2007)with IRAM 30 m CO observations. However, the CO cloud is blueshifted, same as the primary component of the Hα and Hβ. Hence, the redshifted secondary component cannot be attributed to the CO cloud. We speculate that this clump is a low-excitation cloud in the M81 group. 3.3. Line Ratios Physical mechanisms responsible for the gas ionization can in principle be distinguished using line ratio diagnostics, among which the most commonly used is the Baldwin −Phillips–Terlevich (BPT)diagram (Baldwin et al. 1981). We construct a spatially resolved BPT diagram for the circumnuclear region of M81, using the standard pair of 8 NASA/IPAC Extragalactic Database. 9 The Astrophysical Journal, 928:111 (14pp), 2022 April 1 Li et al.