A Deeper Look into eFEDS AGN Candidates in Dwarf Galaxies with Chandra Adonis A. Sanchez1, Amy E. Reines1 , Ákos Bogdán2 , and Ralph P. Kraft2 1 eXtreme Gravity Institute, Department of Physics, Montana State University, Bozeman, MT 59717, USA 2 Center for Astrophysics | Harvard and Smithsonian, 60 Garden Street, Cambridge, MA 20138, USA Received 2024 June 26; revised 2024 August 17; accepted 2024 August 28; published 2024 October 1 Abstract The ability to accurately discern active massive black holes (BHs) in nearby dwarf galaxies is paramount to understanding the origins and processes of “seed” BHs in the early Universe. We present Chandra X-ray Observatory observations of a sample of three local dwarf galaxies (M* � 3× 109 Me, z � 0.15) previously identified as candidates for hosting active galactic nuclei (AGN). The galaxies were selected from the NASA- Sloan Atlas with spatially coincident X-ray detections in the eROSITA Final Equatorial Depth Survey. Our new Chandra data reveal three X-ray point sources in two of the target galaxies with luminosities between log(L2−10 keV [erg s−1])= 39.1 and 40.4. Our results support the presence of an AGN in these two galaxies and an ultraluminous X-ray source (ULX) in one of them. For the AGNs, we estimate BH masses of MBH∼ 105−6Me and Eddington ratios on the order of ∼10−3. Unified Astronomy Thesaurus concepts: Black holes (162); Dwarf galaxies (416); X-ray active galactic nuclei (2035); Active galactic nuclei (16) 1. Introduction It is well-understood and documented that massive black holes (BHs) with 106MBH/Me 109 inhabit the centers of nearly all massive galaxies. A portion of these BHs exhibit accretion, emit light across the electromagnetic spectrum, and are luminescent as active galactic nuclei (AGNs). As we reach toward lower mass scales (MBH 105Me), the fraction of dwarf galaxies with stellar masses Må 109Me hosting massive BHs/AGNs becomes much less certain. However, these systems are paramount, as they may provide meaningful constraints on the seeding model(s) for the formation of the first generation of BHs in the early Universe (e.g., M. Volont- eri 2010; J. E. Greene et al. 2020; A. E. Reines 2022). They also are representative of an early stage in the growth of massive BHs and provide an avenue for studying the effects of BH feedback at low masses (J. Silk 2017). There has been consistent growth in the evidence supporting massive BHs in dwarf galaxies found via various selection techniques, such as optical emission line spectroscopy, (J. E. Greene & L. C. Ho 2004; A. E. Reines et al. 2013), optical variability, mid-IR colors, and radio observations (for a review, see A. E. Reines 2022). However, each of these methods has its limitations. Optical searches tend to be biased toward more massive BHs with higher Eddington fractions and galaxies with low star formation rates (A. E. Reines et al. 2013). Selection based on mid-IR colors suffers from contamination by dwarf starburst galaxies mimicking AGN signatures (K. N. Hainline et al. 2016; L. J. Latimer et al. 2021b). Finally, the sensitivities of current radio surveys may miss a significant fraction of the population (A. E. Reines et al. 2020). X-ray observations can bypass many issues suffered by other techniques, as they are effective at detecting active BHs with low Eddington ratios (E. Gallo et al. 2008; L. C. Ho 2008; R. C. Hickox et al. 2009) or those in galaxies with active star formation (A. E. Reines et al. 2016; R. She et al. 2017; E. Kimbro et al. 2021). X-ray studies have been conducted in the past not only to search for AGNs in dwarf galaxies but also to explore the BH occupation fraction and place AGN candidates in dwarf galaxies under further scrutiny to confirm their presence. For example, S. M. Lemons et al. (2015) found candidate AGNs in local dwarf galaxies (z< 0.055;Må 109.5Me) from archival data within the Chandra Source Catalog, and a similar study was conducted by K. L. Birchall et al. (2020) with the 3XMM catalog. Follow-up observations by E. Thygesen et al. (2023) aimed to verify the nature of the AGNs in three dwarfs from the sample found by S. M. Lemons et al. (2015) using high- resolution Chandra and Hubble Space Telescope observations. In this study, we present follow-up Chandra X-ray observa- tions of three nearby dwarf galaxies, each previously found to host one nuclear or off-nuclear X-ray source. These three targets come from a sample of six originally identified by L. J. Latimer et al. (2021a), who provided a determination of the AGN fraction in dwarf galaxies using X-ray observations from the eROSITA Final Equatorial Depth Survey (eFEDS; H. Brunner et al. 2022). This work provided a meaningful example of what will be possible with the release of the eROSITA All-Sky Survey (eRASS) regarding the search for and understanding of AGNs in dwarf galaxies (A. Sacchi et al. 2024; S. D. Bykov et al. 2024). 2. Sample Selection Our target galaxies were selected from the study by L. J. Latimer et al. (2021a). They began with a parent sample of 63,582 local dwarf galaxies (z� 0.15, Må� 3× 109Me) from the NASA-Sloan Atlas (NSA; v1_0_1) and found that 495 of these galaxies fall within the ∼140 deg2 (H. Brunner et al. 2022) eFEDS field. They cross-matched the ∼28,000 X-ray sources in the eFEDS main catalog with the 495 NSA dwarf galaxies in the eFEDS footprint, resulting in a final sample of six dwarf galaxies with candidate massive BHs, each The Astrophysical Journal, 974:3 (7pp), 2024 October 10 https://doi.org/10.3847/1538-4357/ad7588 © 2024. 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 https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0002-0765-0511 https://orcid.org/0000-0002-0765-0511 https://orcid.org/0000-0002-0765-0511 http://astrothesaurus.org/uat/162 http://astrothesaurus.org/uat/416 http://astrothesaurus.org/uat/2035 http://astrothesaurus.org/uat/2035 http://astrothesaurus.org/uat/16 https://doi.org/10.3847/1538-4357/ad7588 https://crossmark.crossref.org/dialog/?doi=10.3847/1538-4357/ad7588&domain=pdf&date_stamp=2024-10-01 https://crossmark.crossref.org/dialog/?doi=10.3847/1538-4357/ad7588&domain=pdf&date_stamp=2024-10-01 http://creativecommons.org/licenses/by/4.0/ with one associated eFEDS X-ray source. We refer to L. J. Latimer et al. (2021a) for further details on the sample. All six of the galaxies exhibited moderately enhanced X-ray emission compared to their expected galaxy-wide contribution from X-ray binaries (XRBs) based on the stellar mass and star formation rate of each galaxy, suggestive of accreting massive black holes (see Section 3.3 of L. J. Latimer et al. 2021a). They found X-ray source luminosities of L0.5−8 keV∼ 1039−40 erg s−1 across the galaxies in their sample. Of the six dwarf galaxies, we have selected three (IDs 1, 2, and 3) to obtain new Chandra X-ray Observatory observations as part of the Chandra Guaranteed Time Observer (GTO) program (PI: Kraft). Chandra supplies us with a subarcsecond spatial resolution for targets located at the aimpoint, whereas the eROSITA resolution is ∼16″ (P. Predehl et al. 2021). The much better resolution of Chandra allows us to precisely constrain the locations and positional uncertainties of the X-ray sources previously detected in these galaxies, more accurately describe their X-ray emission, as well as fit spectra to determine further characteristics. We adopt star formation rates from L. J. Latimer et al. (2021a), based on the far-UV and mid- infrared luminosities via the relationships of L. C. Kennicutt & N. J. Evans (2012) and C.-N. Hao et al. (2011). The properties of the three dwarf galaxies are summarized in Table 1, and images with the positions of their associated X-ray source are in Figure 1. 3. Observations and Data Reduction The observations were taken with Chandra between 2022 March 6, and 2023 February 27, with exposure times ranging between 11.4 and 22.8 ks. Each observation was taken with the galaxy centered on the S3 chip of the Advanced CCD Imaging Spectrometer (ACIS). A summary of the Chandra observations is provided in Table 2. We carried out the data reduction using CIAO v.4.14 (A. Fruscione et al. 2006) software with the calibration data set, CALDB v4.9.8. We first reprocessed and calibrated the data using the chandra_repro task, which resulted in new level 2 event files. Based on these event files, we generated images in the 0.5–2 keV (soft), 2–7 keV (hard), and 0.5–7 keV (broad) bands. To identify and remove any time intervals with high back- ground periods, we used the deflare script and excluded time intervals where the background rate was >3σ above the mean rate. For galaxy ID 1, two observations are available, which were merged using the merge_obs task. To identify and create a list of X-ray point sources in the Chandra images, we used the wavdetect tool in CIAO on the ACIS-S3 chip. We adopted wavelet scales of 1.0, 1.4, 2.0, 2.8, and 4.0 to search for point sources, setting the significance threshold to 10−6, which approximately results in one false source detection over the area of the S3 chip. While we detected multiple sources in each image, in the immediate vicinity of the galaxies, we detected zero, one, and two sources for galaxies ID 1, ID 2, and ID 3, respectively. The location, characteristics, and nature of these X-ray sources are further discussed in Section 4. 4. Analysis and Results 4.1. X-Ray Sources We begin by determining if any X-ray sources are associated with the target galaxies by searching the regions within 3r50, where r50 is the half-light radius provided by the NSA. For galaxy ID 1, the detected X-ray sources lie well beyond the optical extent of the galaxy in both the individual and merged observations. For galaxies ID 2 and ID 3, we detect X-ray sources within the optical extent of the galaxies with one source in ID 2 and two sources in ID 3. Based on the locations of the X-ray sources, we find that their positions are consistent with the X-ray source positions and positional uncertainties of the sources detected by eROSITA (L. J. Latimer et al. 2021a). Figure 1 shows the DECaLS optical images and the Chandra X-ray images of the three galaxies. This reveals that the X-ray source detected in ID 2 is relatively central, with an offset of ≈1″ from its poorly defined center. For galaxy ID 3, one of the sources, X2, is roughly associated with the nucleus of the dwarf galaxy, with an offset of ≈2 8, while the offset between the northern source, X1, and the galaxy centroid is ≈5″. To derive the number of X-ray counts associated with the sources, we applied circular apertures with a 2″ radius, which encompass 90% of the enclosed energy fraction at 4.5 keV. The background counts were derived from annuli with inner and outer radii of 2″–24″. We derived the number of net counts for each source region and applied a 90% aperture correction. We used the CIAO srcflux tool to derive the X-ray fluxes of the sources in the 0.5−2 and 2−10 keV bands. We assumed a power-law spectral model with a photon index of Γ= 1.8, which is common for low-luminosity AGN (L. C. Ho 2008, 2009), and ultraluminous X-ray sources in this luminosity range (D. A. Swartz et al. 2008). We use the Galactic column density maps from J. M. Dickey & F. J. Lockman (1990). We note that potential absorption intrinsic to each source is not accounted for, and therefore the observed fluxes should be considered as lower limits. Table 1 Galaxy Properties ID NSAID R.A. Decl. NH z r50 Distance log M*/M☉ Star Formation Rate (deg) (deg) (1020 cm−2) (kpc) (Mpc) (M☉ yr−1) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) 1 82162 140.942967 2.753302 3.58 0.0177 8.24 73 9.08 0.16 2 623354 145.180051 3.958864 3.64 0.0051 0.91 21 8.40 0.04 3 648474 141.572049 3.134800 3.82 0.0149 3.24 61 9.44 0.27 Note. Table 1 information is taken from L. J. Latimer et al. (2021a). Column (1): identification number used in L. J. Latimer et al. (2021a). Column (2): NSAIDs from v1_0_1 of the NSA. Columns (3) and (4): R.A. and decl. of the galaxy. Column (5): Galactic neutral hydrogen density (retrieved via https://cxc.harvard.edu/toolkit/ colden.jsp). Column (6): redshift, specifically the zdist parameter from the NSA. Column (7): Petrosian 50% light radius. Column (8): galaxy distance. Column (9): galaxy stellar mass. Column (10): estimated star formation rates from L. J. Latimer et al. (2021a). The values given in columns (6)–(9) are from the NSA and we assume h = 0.73. 2 The Astrophysical Journal, 974:3 (7pp), 2024 October 10 Sanchez et al. https://cxc.harvard.edu/toolkit/colden.jsp https://cxc.harvard.edu/toolkit/colden.jsp To derive the luminosities of the sources, we use the distances in Table 1. The obtained 2−10 keV band luminosities are in the range of ( -Llog 2 10 keV/erg s−1)= 39.20–40.43. We report the number of counts, the unabsorbed fluxes, and luminosities in Table 3. We note that ID 3-X1 has a luminosity ∼5× higher than the eFEDS source detected in ID 3, while the other detected X-ray sources are consistent with the eFEDS luminosities. In the absence of a detection in the galaxy ID 1, we place an upper limit on the luminosity of any potential source at the location of the eFEDS detection. The upper limit from Chandra is Figure 1. Left: three-color images of our target galaxies retrieved from the Dark Energy Camera Legacy Survey (DECaLS; A. Dey et al. 2019) with the z, r, and g bands depicting red, green, and blue, respectively. Right: Chandra X-ray images of the galaxies in the 0.5–7 keV band. We show the positions and the corrected combined positional uncertainty (based on the RADEC_ERR_CORR parameter from the eFEDS main catalog) in magenta. We also overlay the positions of our detected X-ray sources as red crosses. The green circles represent the ∼2″ apertures around the sources. The red circle shows the 95% positional uncertainties of the eROSITA- detected X-ray sources. Each Chandra-detected source falls within the positional uncertainty reported in the eFEDS main catalog, except for ID 1, which does not have a detectable X-ray source in the Chandra image. Table 2 Chandra Observations ID Date Observed Obs ID Exp. Time Nbkg (ks) 1 2022 Oct 13 26034 18.5 1.3840 2022 Oct 16 27479 11.4 0.8680 2 2022 Mar 07 26035 12.9 0.1456 3 2023 Feb 27 26036 22.8 0.3561 Note. Nbkg is the expected number of hard band N (>S) background sources within our galaxy region (3r50) using A. Moretti et al. (2003). 3 The Astrophysical Journal, 974:3 (7pp), 2024 October 10 Sanchez et al. L0.5−2 keV< 2.4× 10−15 erg s−1 cm−2 and the eFEDS source in ID 1 has L0.5−2keV= 2.3× 10−15 erg s−1 cm−2 (L. J. Latimer et al. 2021a). Given the consistency of these values, the source could be relatively soft and not detected by Chandra due to decreased sensitivity in the soft band, or the source could be variable. While spectral analysis could help differentiate between these scenarios, the eROSITA-detected source is too faint to carry out such an analysis. 4.2. Spectral Properties We derived the X-ray hardness ratios using the Bayesian Estimation of Hardness Ratios (BEHR) code (T. Park et al. 2006). BEHR is a robust tool for estimating hardness ratios for sources within the Poisson regime of low counts, which also provides uncertainties if a source is detected in only one of the two bands. We define the hardness ratio as HR= (H− S)/ (H+ S), where H and S are the numbers of detected counts in the hard (2−7 keV) and soft (0.5–2 keV) bands, respectively. We note that source X2 in galaxy ID 3 is not detected in the hard band; hence, we apply a correction factor of ∼1.6 to the soft band flux using the spectral models introduced below. The resulting hardness ratios are in the range of HR=−0.9−0 and are shown in Figure 2. To place these hardness ratios in context, we derived the expected values for a range of spectral models. To this end, we used the Portable, Interactive Multi-Mission Simulator (PIMMS)3 tool and derived hardness ratios for unabsorbed power-law models with Γ= 1.8, Γ= 2.0, and Γ= 2.5. Additionally, we derived the expected hardness ratios assuming a power-law model with a slope of Γ= 1.8 and intrinsic absorption of NH= 1022 cm−2, NH= 1023 cm−2, and NH= 1024 cm−2. These are shown as solid and dashed lines, respectively, in Figure 2. We find that the observed hardness ratios are broadly consistent with unabsorbed power-law models with Γ= 1.8−2.5, although source X2 in galaxy ID 3 exhibits a significantly softer spectrum, which is also indicated by its nondetection in the hard band. We also modeled and fitted the spectra of two of our X-ray sources, the source in ID 2 and source X1 in ID 3, with CIAOʼs built-in modeling and fitting package, SHERPA v.4.14 (D. Burke et al. 2021). The extracted source and background apertures were the same as described earlier (see Section 3). We used the CIAO specextract tool to extract the spectra and the response files. Since our sources contain a relatively low number of photon counts, we used the C-statistic to calculate the best-fitting parameters. We modeled the back- ground along with our source spectra. Given the possibility of both Galactic absorption and absorption intrinsic (NH,int) to the target galaxies, we fitted the spectra using two models: (i) an absorbed power-law model without intrinsic absorption (xsphabsxspowerlaw) and (ii) an absorbed power-law with intrinsic absorption (xsphabsx- sphabs∗xspowerlaw). For both models, we used the Galactic column densities reported by J. M. Dickey & F. J. Lockman (1990), which we kept fixed. We allowed all other parameters to vary and estimated all uncertainties to 68% confidence. Performing the fit for each case on our sources, we found that the spectrum of ID 2 has a best-fit power-law index of G = - +2.37 0.32 0.33. There was no significant difference between the two models for the spectrum of ID 2, as it yielded a negligible value for the best-fit intrinsic absorption. For ID 3, the X1 X-ray source was the only detection with sufficient counts for spectral fitting. We found a similar situation, with both models yielding a best-fit power-law index of G = - +1.66 0.19 0.25, with the second model arriving at a negligible intrinsic absorption. We show the spectrum for each galaxy and their best-fit power-law models in Figure 3. We calculated unabsorbed 2−10 keV X-ray source fluxes and luminosities Table 3 X-Ray Sources ID R.A. Decl. Net Counts Flux (10−15 erg s−1 cm−2) Luminosity (log(erg s−1)) (deg) (deg) 0.5–2 keV 2–7 keV 0.5–2 keV 2–10 keV 0.5–2 keV 2–10 keV (1) (2) (3) (4) (5) (6) (7) (8) (9) 1 L L L L <2.40 <5.23 <39.2 <39.5 2 145.179013 3.959214 28.5 ± 9.7 - +15.3 6.1 8.7 27.43 29.72 39.16 39.20 3-X1 141.571720 3.136256 57.3 ± 13.5 58.1 ± 13.8 32.58 59.89 40.16 40.43 3-X2 141.571383 3.135414 - +10.3 4.8 7.3 L 5.88 L 39.42 L Note. Column (1): galaxy ID. Columns (2) and 3: R.A. and decl. of the X-ray sources. Columns (4) and (5): net counts after applying a 90% aperture correction. Error bars represent 90% confidence intervals. Columns (6) and (7): fluxes corrected for Galactic absorption. Columns (8) and (9): luminosities corrected for Galactic absorption; calculated assuming a photon index of Γ = 1.8. Figure 2. Hardness ratio vs. 2–10 keV band X-ray luminosity for our two galaxies with a detected X-ray source in at least one of the two bands. The hardness ratio was calculated using BEHR (see Section 4.1). The error bars are the 68% confidence intervals. Hardness ratios for unabsorbed power-law models with Γ = 1.8, 2.0, and 2.5 are depicted as solid gray lines, while the absorbed power laws with Γ = 1.8 are shown with gray dashed lines. 3 https://heasarc.gsfc.nasa.gov/cgi-bin/Tools/w3pimms/w3pimms.pl 4 The Astrophysical Journal, 974:3 (7pp), 2024 October 10 Sanchez et al. https://heasarc.gsfc.nasa.gov/cgi-bin/Tools/w3pimms/w3pimms.pl based on the modeled spectra and summarized the results in Table 4. We note that the best-fit power-law indices for the source in ID 2 and source X1 in ID 3 are consistent with the hardness ratios calculated above, with Γ≈ 2.4 and Γ≈ 1.7, respectively (see Figure 2). The X-ray source luminosities found, whether via srcflux or fitting the spectra, return values that are also consistent with values found in other studies concerning AGNs in dwarf galaxies (e.g., A. E. Reines et al. 2014; V. F. Baldas- sare et al. 2017; M. Mezcua et al. 2018). 5. Discussion 5.1. Possible Origins of the X-Ray Emission Below we consider a number of possible origins for the X-ray sources we detect toward our target dwarf galaxies. Based on the eFEDS detections, L. J. Latimer et al. (2021a) found luminosities above the galaxy-wide expected contrib- ution from XRBs. Here we consider contamination from the cosmic X-ray background (CXB) and ultraluminous X-ray sources (ULXs). 5.1.1. Contamination from CXB Sources We first estimate the number of hard/soft X-ray sources from the CXB that we expect may reside within 3r50 of the target galaxies. Using the -N Slog log function established by A. Moretti et al. (2003), we derive the number of expected CXB sources per unit area given a limiting flux. Based on the source detection threshold achieved by the Chandra observa- tions for our target galaxies, the faintest detectable sources are in the range of ( )= - ´ - - -S 1.2 2 10 erg s cmmin,soft 15 1 2 and ( )= - ´ - - -S 2.3 4 10 erg s cmmin,hard 15 1 2 for the 0.5−2 keV and 2−10 keV bands, respectively. Based on these fluxes, the -N Slog log function, and the area of the 3r50 for the galaxies, we expect to detect 0.085−0.525 and 0.147−1.38 CXB sources in the soft and hard bands, respectively. Therefore, it is unlikely that all detected sources are resolved CXB sources. The location of source X1 in ID 1 lies far to the northwest of the target galaxy (see Figure 1) yet lies within 3r50 (∼24 kpc; Table 1). It is therefore a prime candidate for a background source. 5.1.2. ULXs Next, we take into consideration the likelihood of ULXs as the origin of the X-ray sources detected in our sample. ULXs are described as being off-nuclear sources with X-ray luminosities above 1039 erg s−1 (see P. Kaaret et al. 2017 for a review). This population is consistent with the extension of the bright end tail of the luminosity function of high-mass X-ray binaries. Our source positions and their luminosities fall squarely into that definition. We refer to the analysis done by L. J. Latimer et al. (2021a), who consider various relations to estimate the expected number of ULXs in their sample of six dwarf galaxies. Ultimately, they conclude that ∼1 of the X-ray sources could be a ULX. Taking into account our smaller sample size, and hence lower combined star formation rate, we expect that our sample contains ∼0.5 ULXs. Given that we detect two X-ray sources in ID 3 with our Chandra observations, it is plausible that one of the sources is a ULX and one is an AGN. While we cannot definitively say which source is which, our favored interpretation is that X1 is a ULX while X2 is an AGN. First, in contrast to X2, X1 is more offset with respect to the previous eFEDS detection and has a position just barely consistent with the optical center of the galaxy. It is also highly variable with a Chandra luminosity ∼5× higher than the eFEDS luminosity (Section 4.1), whereas source X2 has a Chandra luminosity consistent with the eFEDS Figure 3. X-ray energy spectra for the X-ray source detected in galaxy ID 2 (left panel) and for source X1 in ID 3 (right panel). The spectra were fit with absorbed power-law models, resulting in best-fit values of G = - +2.37 0.32 0.33 and G = - +1.66 0.19 0.25 for the sources in ID 2 and ID 3, respectively. Table 4 X-Ray Source Properties from Fitting ID C-stat/degrees of freedom Γ F2−10 keV log L2−10 keV (10−15 erg s−1 cm−2) (erg s−1) (1) (2) (3) (4) (5) 2 24.6/30 - +2.37 0.32 0.33 - +21.97 8.98 15.16 39.1 ± 0.2 3-X1 50/48 - +1.66 0.19 0.25 - +60.27 22.52 34.80 40.4 ± 0.2 Note. Column (1): identification number associated with the X-ray source. Column (2): goodness-of-fit/degrees of freedom. Column (3): best-fit photon index. Column (4): fluxes in the 2−10 keV band found from the modeled spectra corrected for absorption. Column (5): luminosity in the 2−10 keV band. The Galactic NH values given in Table 1 provided the best fit for both sources. 5 The Astrophysical Journal, 974:3 (7pp), 2024 October 10 Sanchez et al. luminosity. While variability is not inconsistent with AGN activity, this behavior is quite common for stellar mass BHs. The photon index (Γ∼ 1.7) derived from the spectrum of X1 is also entirely consistent with known ULXs (D. A. Swartz et al. 2004). Moreover, the softer value of the hardness ratio for X2 is consistent with previous X-ray studies of AGNs in dwarf galaxies (A. E. Reines et al. 2016; V. F. Baldassare et al. 2017). 5.2. Massive BHs If the X-ray luminosities we have measured are due to accreting massive BHs, we can estimate their Eddington ratios. The Eddington ratio of an accreting BH is found via ( ) ( ) ( )k=f L L , 1Edd X Edd where κ is the hard (2−10keV) bolometric correction, LX is the hard X-ray luminosity of the BH, and LEdd is the Eddington luminosity of the BH, which is given by ( )» ´ -L M M 1.26 10 erg s . 2Edd 38 BH 1  The BH masses are estimated using Equation (4) from A. E. Reines & M. Volonteri (2015), which gives log(MBH) as a function of total host-galaxy stellar mass, log(M*). The relation for local AGNs is given by ( ) ( ) ( )a b= +M M M Mlog log 10 , 3BH 11 *  where α= 7.45± 0.08 and β= 1.05± 0.11. The resultant BH mass estimates are log(MBH/Me) ≈4.7 and 5.8 for the AGNs in ID 2 and ID 3, respectively. These BH masses carry uncertainties of ∼0.55 dex (A. E. Reines & M. Volonteri 2015). To apply the hard band bolometric correction, we invoke F. Duras et al. (2020), which returns a mean Lbol/L2−10 keV≈ 15.3. Combining these values, we find Eddington fractions of ∼0.3% and ∼0.08% for the candidate AGNs in ID 2 and ID 3 (X2), respectively. 5.3. Comparison with Latimer et al. (2021a) To compare the X-ray luminosities from eROSITA pre- sented by L. J. Latimer et al. (2021a) with our values from Chandra, we must adopt the same parameters in the analysis. To be consistent with L. J. Latimer et al. (2021a), we re- estimate source fluxes using a photon index of Γ= 2.0 and Galactic absorption of 3.0× 1020 cm−2. We also run through our analysis in the 0.5–8 keV band. To compare ID 3-X2, we used the above parameters and analyzed them to obtain soft band fluxes, which we convert to 0.5–8 keV using a correction factor of ∼2.1 found through PIMMS. The luminosities of the X-ray source in ID 2 and the source X2 in ID 3 are consistent with the sources L. J. Latimer et al. (2021a) originally detected in the eROSITA data. However, we also detect an additional source in ID 3 (X1) that is a bit offset from the nucleus with a luminosity of log (L2−10keV/erg s−1) ≈40.4. Given the 16″ resolution of eROSITA (P. Predehl et al. 2021), X1 and X2 in ID 3 would not have been resolved in the eFEDS catalog since we find a spatial offset between X1 and X2 of ∼3″(∼1 kpc projected). It may be that X2 was the source originally detected by L. J. Latimer et al. (2021a) and we have detected a separate, new source that could be a ULX (see Section 5.1.2). 6. Summary and Conclusions We analyze new Chandra X-ray Observatory observations of three AGN candidates in dwarf galaxies identified by L. J. Lat- imer et al. (2021a) from the eFEDS catalog. Our main findings are summarized below: 1. X-ray emission is detected in two of our three target galaxies (one source in ID 2 and two sources in ID 3) with the X-ray sources being located within the positional uncertainty reported by eFEDS. We do not detect a luminous X-ray source in ID 1 within the positional uncertainty reported by eFEDS nor the optical extent of the galaxy itself. 2. The X-ray source in ID 2 has a luminosity consistent with the eFEDS source (L2−10 keV∼ 1× 1039 erg s−1) and has a position that is roughly central within the target galaxy, although there does not seem to be a well-defined center. Given the luminosity is also well above that expected from XRBs based on the host-galaxy stellar mass and star formation rate (L. J. Latimer et al. 2021a), the X-ray source may indeed be an accreting massive BH. The X-ray spectrum is well-described by an absorbed power- law model with Γ= 2.4 and negligible intrinsic absorp- tion. The soft spectrum is consistent with other X-ray studies of AGNs in dwarf galaxies (e.g., A. E. Reines et al. 2016; V. F. Baldassare et al. 2017). 3. Our Chandra observations resolved the previous eFEDS source in ID 3 into two individual sources. X2 has a luminosity consistent with that in eFEDS (L2−10 keV∼ 4× 1039 erg s−1), which is higher than that expected from XRBs (L. J. Latimer et al. 2021a). X2 also has a position approximately consistent with the optical nucleus of the galaxy, strongly suggesting this X-ray source is due to an active massive BH. X1 has a luminosity ∼5× higher than the eFEDS source detected in ID 3 (L2−10 keV∼ 2× 1040 erg s−1) and is offset from the nucleus. The X-ray spectrum of X1 is best fit by a power-law model with a photon index of Γ= 1.7 and negligible intrinsic absorp- tion. Given its position and high variability, this source is potentially a ULX. 4. For ID 2 and ID 3-X2, BH masses are estimated using the scaling between BH mass and host-galaxy stellar mass for local AGNs from A. E. Reines & M. Volonteri (2015). The BH mass estimates are log(MBH/Me) ≈ 4.7 and 5.8 for ID 2 and ID 3-X2, respectively, and carry uncertainties of ∼0.55 dex. The corresponding Eddington ratios are ∼0.3% for ID 2 and ∼0.08% for ID 3-X2. Through this work, we have strengthened the case for the presence of massive BHs in two of the dwarf galaxies (IDs 2 and 3) identified by L. J. Latimer et al. (2021a), while weakening the case for a highly offset AGN in ID 1. Additional follow-up observations at radio wavelengths would help confirm the sources in ID 2 and ID 3 are indeed AGNs since massive BHs are much more luminous than XRBs in the radio. Acknowledgments We thank the reviewer for the constructive comments. A.E.R. gratefully acknowledges support for this work provided by the NSF through CAREER award 2235277 and NASA through EPSCoR grant No. 80NSSC20M0231. Á.B. and R.P.K. acknowledge support from the Smithsonian Institution 6 The Astrophysical Journal, 974:3 (7pp), 2024 October 10 Sanchez et al. and the Chandra Project through NASA contract NAS8-03060. This paper employs a list of Chandra data sets, obtained by the Chandra X-ray Observatory, contained in doi:10.25574/ cdc.282. ORCID iDs Amy E. Reines https://orcid.org/0000-0001-7158-614X Ákos Bogdán https://orcid.org/0000-0003-0573-7733 Ralph P. Kraft https://orcid.org/0000-0002-0765-0511 References Baldassare, V. F., Reines, A. E., Gallo, E., & Greene, J. E. 2017, ApJ, 836, 20 Birchall, K. L., Watson, M. G., & Aird, J. 2020, MNRAS, 492, 2268 Brunner, H., Liu, T., Lamer, G., et al. 2022, A&A, 661, A1 Burke, D., Laurino, O., Wmclaugh, et al. 2021, sherpa/sherpa: Sherpa v4.14.0, Zenodo, doi: 10.5281/zenodo.5554957 Bykov, S. D., Gilfanov, M. R., & Sunyaev, R. A. 2024, MNRAS, 527, 1962 Dey, A., Schlegel, D. J., Lang, D., et al. 2019, AJ, 157, 168 Dickey, J. M., & Lockman, F. J. 1990, ARA&A, 28, 215 Duras, F., Bongiorno, A., Ricci, F., et al. 2020, A&A, 636, A73 Fruscione, A., McDowell, J. C., Allen, G. E., et al. 2006, Proc. SPIE, 62701V Gallo, E., Treu, T., Jacob, J., et al. 2008, ApJ, 680, 154 Greene, J. E., & Ho, L. C. 2004, ApJ, 610, 722 Greene, J. E., Strader, J., & Ho, L. C. 2020, ARA&A, 58, 257 Hainline, K. N., Reines, A. E., Greene, J. E., & Stern, D. 2016, ApJ, 832, 119 Hao, C.-N., Kennicutt, R. C., Johnson, B. D., et al. 2011, ApJ, 741, 124 Hickox, R. C., Jones, C., Forman, W. R., et al. 2009, ApJ, 696, 891 Ho, L. C. 2008, ARA&A, 46, 475 Ho, L. C. 2009, ApJ, 699, 626 Kaaret, P., Feng, H., & Roberts, T. P. 2017, ARA&A, 55, 303 Kennicutt, R. C., & Evans, N. J. 2012, ARA&A, 50, 531 Kimbro, E., Reines, A. E., Molina, M., Deller, A. T., & Stern, D. 2021, ApJ, 912, 89 Latimer, L. J., Reines, A. E., Bogdan, A., & Kraft, R. 2021a, ApJL, 922, L40 Latimer, L. J., Reines, A. E., Hainline, K. N., Greene, J. E., & Stern, D. 2021b, ApJ, 914, 133 Lemons, S. M., Reines, A. E., Plotkin, R. M., Gallo, E., & Greene, J. E. 2015, ApJ, 805, 12 Mezcua, M., Civano, F., Marchesi, S., et al. 2018, MNRAS, 478, 2576 Moretti, A., Campana, S., Lazzati, D., & Tagliaferri, G. 2003, ApJ, 588, 696 Park, T., Kashyap, V. L., Siemiginowska, A., et al. 2006, ApJ, 652, 610 Predehl, P., Andritschke, R., Arefiev, V., et al. 2021, A&A, 647, A1 Reines, A. E. 2022, NatAs, 6, 26 Reines, A. E., Condon, J. J., Darling, J., & Greene, J. E. 2020, ApJ, 888, 36 Reines, A. E., Greene, J. E., & Geha, M. 2013, ApJ, 775, 116 Reines, A. E., Plotkin, R. M., Russell, T. D., et al. 2014, ApJL, 787, L30 Reines, A. E., Reynolds, M. T., Miller, J. M., et al. 2016, ApJL, 830, L35 Reines, A. E., & Volonteri, M. 2015, ApJ, 813, 82 Sacchi, A., Bogdan, A., Chadayammuri, U., & Ricarte, A. 2024, arXiv:2406. 01707 She, R., Ho, L. C., & Feng, H. 2017, ApJ, 835, 223 Silk, J. 2017, ApJL, 839, L13 Swartz, D. A., Ghosh, K. K., Tennant, A. F., & Wu, K. 2004, ApJS, 154, 519 Swartz, D. A., Soria, R., & Tennant, A. F. 2008, ApJ, 684, 282 Thygesen, E., Plotkin, R. M., Soria, R., et al. 2023, MNRAS, 519, 5848 Volonteri, M. 2010, A&ARv, 18, 279 7 The Astrophysical Journal, 974:3 (7pp), 2024 October 10 Sanchez et al. https://doi.org/10.25574/cdc.282 https://doi.org/10.25574/cdc.282 https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0001-7158-614X https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0003-0573-7733 https://orcid.org/0000-0002-0765-0511 https://orcid.org/0000-0002-0765-0511 https://orcid.org/0000-0002-0765-0511 https://orcid.org/0000-0002-0765-0511 https://orcid.org/0000-0002-0765-0511 https://orcid.org/0000-0002-0765-0511 https://orcid.org/0000-0002-0765-0511 https://orcid.org/0000-0002-0765-0511 https://doi.org/10.3847/1538-4357/836/1/20 https://ui.adsabs.harvard.edu/abs/2017ApJ...836...20B/abstract https://doi.org/10.1093/mnras/staa040 https://ui.adsabs.harvard.edu/abs/2020MNRAS.492.2268B/abstract https://doi.org/10.1051/0004-6361/202141266 https://ui.adsabs.harvard.edu/abs/2022A&A...661A...1B/abstract https://doi.org/10.5281/zenodo.5554957 https://doi.org/10.1093/mnras/stad3355 https://ui.adsabs.harvard.edu/abs/2024MNRAS.527.1962B/abstract https://doi.org/10.3847/1538-3881/ab089d https://ui.adsabs.harvard.edu/abs/2019AJ....157..168D/abstract https://doi.org/10.1146/annurev.aa.28.090190.001243 https://ui.adsabs.harvard.edu/abs/1990ARA&A..28..215D/abstract https://doi.org/10.1051/0004-6361/201936817 https://ui.adsabs.harvard.edu/abs/2020A&A...636A..73D/abstract https://doi.org/10.1117/12.671760 https://ui.adsabs.harvard.edu/abs/2006SPIE.6270E..1VF/abstract https://doi.org/10.1086/588012 https://ui.adsabs.harvard.edu/abs/2008ApJ...680..154G/abstract https://doi.org/10.1086/421719 https://ui.adsabs.harvard.edu/abs/2004ApJ...610..722G/abstract https://doi.org/10.1146/annurev-astro-032620-021835 https://ui.adsabs.harvard.edu/abs/2020ARA&A..58..257G/abstract https://doi.org/10.3847/0004-637X/832/2/119 https://ui.adsabs.harvard.edu/abs/2016ApJ...832..119H/abstract https://doi.org/10.1088/0004-637X/741/2/124 https://ui.adsabs.harvard.edu/abs/2011ApJ...741..124H/abstract https://doi.org/10.1088/0004-637X/696/1/891 https://ui.adsabs.harvard.edu/abs/2009ApJ...696..891H/abstract https://doi.org/10.1146/annurev.astro.45.051806.110546 https://ui.adsabs.harvard.edu/abs/2008ARA&A..46..475H/abstract https://doi.org/10.1088/0004-637X/699/1/626 https://ui.adsabs.harvard.edu/abs/2009ApJ...699..626H/abstract https://doi.org/10.1146/annurev-astro-091916-055259 https://ui.adsabs.harvard.edu/abs/2017ARA&A..55..303K/abstract https://doi.org/10.1146/annurev-astro-081811-125610 https://ui.adsabs.harvard.edu/abs/2012ARA&A..50..531K/abstract https://doi.org/10.3847/1538-4357/abec6a https://ui.adsabs.harvard.edu/abs/2021ApJ...912...89K/abstract https://ui.adsabs.harvard.edu/abs/2021ApJ...912...89K/abstract https://doi.org/10.3847/2041-8213/ac3af6 https://ui.adsabs.harvard.edu/abs/2021ApJ...922L..40L/abstract https://doi.org/10.3847/1538-4357/abfe0c https://ui.adsabs.harvard.edu/abs/2021ApJ...914..133L/abstract https://doi.org/10.1088/0004-637X/805/1/12 https://ui.adsabs.harvard.edu/abs/2015ApJ...805...12L/abstract https://doi.org/10.1093/mnras/sty1163 https://ui.adsabs.harvard.edu/abs/2018MNRAS.478.2576M/abstract https://doi.org/10.1086/374335 https://ui.adsabs.harvard.edu/abs/2003ApJ...588..696M/abstract https://ui.adsabs.harvard.edu/abs/2003ApJ...588..696M/abstract https://doi.org/10.1086/507406 https://ui.adsabs.harvard.edu/abs/2006ApJ...652..610P/abstract https://doi.org/10.1051/0004-6361/202039313 https://ui.adsabs.harvard.edu/abs/2021A&A...647A...1P/abstract https://doi.org/10.1038/s41550-021-01556-0 https://ui.adsabs.harvard.edu/abs/2022NatAs...6...26R/abstract https://doi.org/10.3847/1538-4357/ab4999 https://ui.adsabs.harvard.edu/abs/2020ApJ...888...36R/abstract https://doi.org/10.1088/0004-637X/775/2/116 https://ui.adsabs.harvard.edu/abs/2013ApJ...775..116R/abstract https://doi.org/10.1088/2041-8205/787/2/L30 https://ui.adsabs.harvard.edu/abs/2014ApJ...787L..30R/abstract https://doi.org/10.3847/2041-8205/830/2/L35 https://ui.adsabs.harvard.edu/abs/2016ApJ...830L..35R/abstract https://doi.org/10.1088/0004-637X/813/2/82 https://ui.adsabs.harvard.edu/abs/2015ApJ...813...82R/abstract http://arxiv.org/abs/2406.01707 http://arxiv.org/abs/2406.01707 https://doi.org/10.3847/1538-4357/835/2/223 https://ui.adsabs.harvard.edu/abs/2017ApJ...835..223S/abstract https://doi.org/10.3847/2041-8213/aa67da https://ui.adsabs.harvard.edu/abs/2017ApJ...839L..13S/abstract https://doi.org/10.1086/422842 https://ui.adsabs.harvard.edu/abs/2004ApJS..154..519S/abstract https://doi.org/10.1086/587776 https://ui.adsabs.harvard.edu/abs/2008ApJ...684..282S/abstract https://doi.org/10.1093/mnras/stad002 https://ui.adsabs.harvard.edu/abs/2023MNRAS.519.5848T/abstract https://doi.org/10.1007/s00159-010-0029-x https://ui.adsabs.harvard.edu/abs/2010A&ARv..18..279V/abstract 1. Introduction 2. Sample Selection 3. Observations and Data Reduction 4. Analysis and Results 4.1. X-Ray Sources 4.2. Spectral Properties 5. Discussion 5.1. Possible Origins of the X-Ray Emission 5.1.1. Contamination from CXB Sources 5.1.2. ULXs 5.2. Massive BHs 5.3. Comparison with Latimer et al. (2021a) 6. Summary and Conclusions References