Nustar view of the central region of the Perseus cluster

Nustar view of the central region of the Perseus cluster


Located at the center of the Perseus cluster, 3C 84 is an extremely bright and nearby radio galaxy. Because of the strong diffuse thermal emission from the cluster in X-rays, the detailed properties and the origin of a power-law component from the central active galactic nucleus (AGN) remains unclear in the source. We report here the first NuSTAR observations of 3C 84. The source was observed for 24.2 and 32 ks on February 01 and 04, 2018, respectively. NuSTAR observations spectrally decompose the power-law AGN component above 10 keV. The power-law component dominates the spectrum above 20 keV with a photon index 1.9 and an energy flux F = 1.0 10 erg cm s, corresponding to an isotropic luminosity, = 7.410 erg s. We discuss possible emitting sites for the power-law component. The expected thermal emission from the accretion disk is not hot enough to account for the hard X-rays detected from the source. Similar X-ray and -ray photon indices and long-term flux variations, the absence of cutoff energy in the hard X-ray spectrum of the source, correlated hard X-ray flux and hardness ratio variations, and the similarity of optical-X-ray slope to blazar rather than Seyfert galaxies supports the hard X-ray power-law component originating from the jet.

Subject headings:
galaxies: active – quasars: individual: 3C 84 – X-rays

1. Introduction

The Perseus cluster of galaxies is one of the X-ray brightest and nearest clusters in the sky. At the center of the Perseus cluster, 3C 84 is a nearby (z=0.0172) radio loud active galactic nucleus (AGN) with a black hole mass (3.4–8.0)10 M (Wilman et al., 2005; Scharwächter et al., 2013) and a relatively low Eddington ratio, 3 (Sikora et al., 2007). The source is classified as a Seyfert Type 1.5 galaxy (Ho et al., 1997). Being at the center of a bright cluster, the X-ray properties of 3C 84 have been extensively studied. Fabian et al. (2015) compared the long-term X-ray variations with 90 GHz radio flux and found similar variations in the two. XMM-Newton observations taken in 2001 suggest the presence of a narrow Fe-K line in the spectrum (Churazov et al., 2003). The recent observations by the Hitomi/Soft X-ray telescope (SXS) confirm the detection of the 6.4 keV Fe-K line at the 5.4 level, and suggest that the line originates within 1.6 kpc of the 3C 84 core (Hitomi Collaboration et al., 2018).

Figure 1.— NuSTAR image of the Perseus cluster in the 3-70 KeV Kev band taken on Feb. 04, 2017. The bright emission at the center of the image is 3C 84.

The source has been detected at GeV energies since the beginning of Fermi Large Area Telescope (LAT) observations (Abdo et al., 2009). Being relatively inactive throughout the 1990s, there has been an increase in -ray flux and cm-radio activity, which appears to have begun in 2005 (Abdo et al., 2009; Nagai et al., 2010; Suzuki et al., 2012). From December 2016 to early January 2017, the source was detected at TeV energies (Mukherjee & VERITAS Collaboration, 2017; Mirzoyan, 2017; MAGIC collaboration et al., 2018). An increase in the radio, soft X-ray and -ray flux was also detected during this flaring activity period (Hodgson et al., 2018; Fukazawa et al., 2018).

The X-ray emission of the source is strongly dominated by the thermal diffuse radiation from the cluster and a power-law component associated with the central AGN (Hitomi Collaboration et al., 2018; Ajello et al., 2009). However the physical origin this power-law component, e.g., whether it is more similar to the jet-like origin of the X-ray continuum in blazars or the thermal Comptonization component in Seyfert galaxies, is not clear. Similarities in the X-ray and radio/-ray flux variations (Fabian et al., 2015; Fukazawa et al., 2018) clearly hint in favor of the presence of a jet-based emission component. We present here the first NuSTAR observations of the Perseus cluster, which spectrally decompose the power-law emission component from the central AGN, and present the Swift-BAT data revealing variability of the hard X-ray flux. This paper is structured as follows. Observations and data reduction are described in Sect. 2. Analysis and results are given in Sect. 3, and Sect. 4 and 5 presents the discussion and conclusions, respectively.

2. Observations and data reduction

At the end of 2016, 3C 84 went through an extreme flaring activity at GeV/TeV energies (Mirzoyan, 2017; Hodgson et al., 2018). To follow up the flaring activity at hard X-rays, we requested two NuSTAR (Harrison et al., 2013) observations of the source. NuSTAR observations were taken in February 2017 (observations IDs : 90202046002, 90202046004). The pointing on February 01, 2017, resulted in 24.2 ks of observing time, while on February 04, 2017, 32 ks were obtained. We also use the archival 105 months of Swift/BAT data, obtained from Oh et al. (2018); we detail the spectral fitting of those data in Section 4.2.

Data were analyzed using the NuSTAR Data Analysis Software (NuSTARDAS) package v.1.3.1. After processing the raw data via nupipeline, we used nuproducts to extract higher level products, i.e. source spectra and light curves. A hard X-ray image of the source can be seen in Fig. 1. Since the Perseus cluster is a spatially extended source, we used different approaches to extract the flux and spectrum of the central AGN 3C 84. Data for the central AGN component were extracted from a region of radius, centered on the position of 3C 84 (RA=3:19:48.38,Dec.=+41:30:42.10). Background was extracted from a radius region roughly SW of the source location. We also cross-check the results by selecting an annular background region centered at the source position with an inner radius of and an outer radius of . We binned the spectra in order to have at least 30 counts per rebinned channel. For spectral fitting, we considered the channels corresponding nominally to the 3–70 keV energy range, where the source was robustly detected. In either case, we treat the thermal cluster emission as a background, and do not care about the residual thermal cluster flux in the source extraction region, only about the hard X-ray emission dominating at higher energies. Our implicit assumption is that the residual cluster emission at the 3C84 extraction region and at the region used by us as background, have roughly the same temperature (Schmidt et al., 2002).

Figure 2.— Fitted NuSTAR spectrum of 3C 84 – data taken on Feb. 01, 2017 (middle) and Feb. 04, 2017 (bottom). Plus symbols are the observed data while the lines represent the fitted models.
Model Parameter Period A Period B
nH (column density in units of cm) 0.14 (frozen)
KT (plasma temperature in keV) 3.47 3.70
Abun (metal abundances) 0.39 0.39
z (redshift) 0.017 (frozen)
Norm (apec normalization) 0.044 0.039 0.039 0.034
(power-law index) 1.85 1.90
K (power-law normalization) 0.00974 0.0114 0.0106 0.0120
/dof (prob.) reduced /degrees of freedom (a corresponding probability) 0.89/454 (0.96) 1.03/564 (0.28)
F AGN flux in units of 10 erg cm s 1.01 0.99 0.99 0.98
F AGN flux in units of 10 erg cm s 3.36 3.5 3.20 3.40
Model1: Parameter NuSTAR data BAT data
apec kT (keV) 3.56 4.02
apec norm 0.031 0.82
powerlaw PhoIndex 1.96
powerlaw norm 0.013 0.008
/dof (prob.) 578/577 (0.20)
Model2: Parameter NuSTAR data BAT data
apec kT (keV) 3.45 3.84
apec norm 0.031 97
cutoffpl PhoIndex 1.94
cutoffpl norm 0.013 0.008
cutoffpl HighECut (keV) 100 100
/dof (prob.) 577/558 (0.21)
Model3: Parameter NuSTAR data BAT data
apec kT (keV) 3.93 4.95
apec norm 0.028 16
pexrav PhoIndex 2.08
pexrav foldE (keV) 100 100
pexrav rel_refl 0.50
pexrav Redshift 0.017 (frozen)
pexrav abund 1.0 (frozen)
pexrav Fe_abund 1.0 frozen)
pexrav cosIncl 0.45 (frozen)
pexrav norm 0.016 0.011
/dof (prob.) 575/558 (0.22)

The following parameters are Fixed in Model1, Model2, and Model3: nH/wabs (10 cm) = 0.14, Abundanc/apec = 0.50, Redshift = 0.017. We used the AtomDB (version 3.0.9) and the default abundance table is from Anders & Grevesse (1989).
Note: The uncertainties on the parameters are 90 confidence limits.

Table 1Parameters of the spectral fits

3. Analysis and Results

3.1. Nustar data

We used  V15.4.0 to model the source spectrum. In addition to an absorption model (), which describes the absorption in the AGN by cold absorbing material in our Galaxy, we used a combined power-law, , and an emission spectrum from hot diffuse gas, , to fit the source spectrum, i.e. . In , we fixed the column density () to the Galactic value of 1.410 cm. The redshift () of the source is fixed to 0.017. The plasma temperature (), metal abundance (), and normalization () are kept free while fitting the spectrum. Both the photon index () and normalization () of the are also set free. The model fit parameters and the reduced values along with the degrees of freedom and the corresponding probability are given in Table 1. As the statistics suggest, this simple model well describes the source spectrum over the energy range form 3.0 to 70.0 keV.

Motivated by the detection by the Hitomi satellite of the Fe K line emission from the AGN (Hitomi Collaboration et al., 2018), we searched for a presence of such a line in the NuSTAR data. The flux of the sum of the two components of the line measured by Hitomi is modest, at photons cm s. We added a narrow Gaussian emission component to our combined model, and note a significant improvement in (with of 10 for either observation). We note that the best-fit of the line flux is significantly greater than that measured by Hitomi (by about a factor of 5, at photons cm s). However, the fitted energy (as observed) is not well determined (values between 6.0 and 6.4 keV are allowed). In conclusion, we consider as a possibility to account for this finding the fact that the weak Fe-K line is sitting on the wing of the very strong (but intrinsically narrow) 6.7 keV (rest) emission line from the hot cluster gas, and our detection is an artifact of slight miscalibration of NuSTAR’s spectral response (specifically, the low-energy wing) to a narrow line.

A significant dominance of the thermal diffuse emission (see Fig. 2) in the source spectrum is seen up to 10 keV. The power-law AGN component clearly takes over above 10 keV. Most of the radiation we see above 20 keV is contributed by the central AGN, 3C 84. In both epochs of our observations, the photon index is determined to be 1.8–1.9, which is quite similar to the -ray photon index during that period, 2.0 (Hodgson et al., 2018). The 20-30 keV energy flux range is 0.9–1.1 10 erg cm s, corresponding to an apparent isotropic X-ray luminosity, , of (7.350.66) 10 erg s. We do not observe any variation in the photon index or flux across the 4 day gap of the two NuSTAR observations (see Table 1).

The estimated soft X-ray, 2-10 keV, flux of the central region (including both the AGN and the residual cluster emission in the 30 arcsec radius extraction region) is 610 erg cm s. Soft X-ray flux of the AGN component is 3.310 erg cm s ( = 2.310 erg s), which is similar to the AGN continuum flux, F (1.96–4.36)10 erg cm s, measured by the Hitomi/SXS observations in February, 2016 (Hitomi Collaboration et al., 2018). The photon index is also similar for the two observations. Hitomi/SXS measured the power-law index in the range 1.45-2.06, and for the NuSTAR observations, it is 1.9. The improved constraints relative to Hitomi result from NuSTAR’s superior hard X-ray coverage (with sensitivity out to 70keV) and spatial resolution.

3.2. Swift/BAT data

We used the archival Swift/BAT data over the first 105 months of observations from December 2004 to August 2013 (Oh et al., 2018) to investigate the hard X-ray spectral properties of the source. The detailed analysis of the BAT data can be found in Oh et al. (2018). We performed a combined fit of the NuSTAR and BAT data, which we discuss in Section 4.2. Note that the NuSTAR and BAT observations are not close in time and the BAT data are averaged over 9 years.

4. Discussion

Given its high sensitivity and angular resolution at hard X-rays, NuSTAR observations spectrally decompose the power-law component in the 3C 84 spectrum. The power-law component of the spectrum could either be from the immediate vicinity, i.e. disk/corona, of the central black hole or from the jet. In the following sub-sections we discuss these possibilities in detail.

4.1. Accretion disk

For a 3-810 M black hole accreting at 10 of its rate, the peak temperature at 1  is 2.510 K, corresponding to 5.210 Hz, which falls in the ultraviolet (UV) band. Of course for 3C 84, the accretion rate is way below 10 and (); so, the disk would be even cooler. Moreover, the thermal radiation from the disk does not follow a power law.

4.2. Corona

Inverse-Compton scattering of photons from the accretion disk by the corona could emit in hard X-rays (Reis & Miller, 2013; Fabian et al., 2015). For radiatively compact emission regions (like corona, 2-30 , Fabian et al., 2009; Wilkins & Fabian, 2011; Sanfrutos et al., 2013), pair production naturally occurs. At extremely high temperature, pair production can be a runway process, limiting any further rise in the temperature and a producing a cutoff in the spectrum, which for Seyfert galaxies is in the 50-200 KeV range (Fabian et al., 2015). However at low mass accretion rates () the compactness of the corona will be low and so very high temperatures can be achieved before the pair-production condition is met.

In the NuSTAR spectrum extending up to 70 keV, we do not see any hint of a cutoff in the spectrum. To further examine the possibility of a cutoff at higher energies, we performed a combined fit of the NuSTAR and Swift/BAT observations. We noticed a clear disagreement between NuSTAR and BAT data between (15-40) keV, which could be anticipated as the data are not simultaneous. Moreover, the BAT spectrum is extracted from a larger region (19.5 arcmin) compared to NuSTAR and is dominated by cluster emission up to 50 KeV (Ajello et al., 2009), which could be possibly responsible for the discrepancy between the two. Because of this reason, we ignore the BAT data below 50 KeV in the joint NuSTAR-BAT spectral fitting.

We tried several models to fit the combined NuSTAR and BAT data. At first, we used the simple model () and froze the model parameters to their respective NuSTAR model-fit values except the plasma temperature () and normalization and the power-law index and its normalization. While we present here the spectral-fit results using the NuSTAR data taken on February 04, 2017, we obtained similar results for the first NuSTAR observation. The model-fit parameters are given in Table 1, and the fitted spectrum is plotted in Fig. 3 (top). As the statistics suggest, well describes the source spectrum up to 200 KeV. Moreover, the power-law index for the combined NuSTAR-BAT data is same as the power-law index for the NuSTAR spectral fits.

We next replaced the power-law component by a cutoff power-law () (), and compared the test statistics of the two fits. We noticed no significant change in the after adding the cutoff energy. Figure 4 (top) illustrates the contour plots between the power-law photon index and cutoff energy, which constrain the cutoff energy to be above 100 keV for values of the index between 1.7-2.1. We further tested the reflection signatures in the spectrum ( Magdziarz & Zdziarski, 1995). Again no substantial improvement is noticed in the spectral fitting. The best fit parameters are listed in Table 1, and the fitted spectrum is plotted in Fig. 3. The index and cutoff energy contour plots in the reflection model are shown in Fig. 4 (bottom). The reflection model constrained the cutoff energy to be above 100 keV for index 1.9-2.2. The absence of any significant improvement in the spectral fits by adding either cutoff energy or reflection component over a simple power-law component suggests that if there is a cutoff energy in the hard X-ray spectrum of the source, it is beyond the BAT energy range. Both spectral models constrain the cutoff energy to be above 100 keV, which argues in favor of the jet-based origin of the hard X-ray emission in the source.

Figure 3.— Combined NuSTAR and BAT spectrum fitted with Model1: (, top), Model2: (, middle), and Model3: (, bottom). The best fit parameters are listed in Table 1.
Figure 4.— Top: Confidence contours of cutoff energy versus photon index parameters in the (cutoff power-law, top) and (reflection, bottom). The best fit parameter (“+” symbol) and 68, 90, and 99 levels are shown.

4.3. Jet

Since the photon index and flux behavior is quite similar in the NuSTAR X-ray and Fermi-LAT -ray observations, we used the NuSTAR spectral information to extrapolate and predict the -ray flux in the Fermi energy range using the 7 routine in . The task returns a flux value of 1.910 erg cm s corresponding to an isotropic luminosity of 1.1410 erg s in the 0.3–1.0 GeV energy range. The estimated isotropic -ray luminosity using NuSTAR observations is similar to the observed -ray luminosity of the source using the Fermi-LAT, i.e. L 10 erg s (Fukazawa et al., 2018; Hodgson et al., 2018). This suggests that the hard X-ray power-law spectrum extends up to rays. If this is true, the flux variations at the two energy bands should also have similar behavior. In fact, in a recent study Fukazawa et al. (2018) found a similar variability trend in the long-term 5-10 keV X-ray and -ray flux variations. Moreover, the X-ray emission can be described as the low-energy tail of the inverse-Compton scattering in case of a one-zone synchrotron self-Compton model, suggesting a jet-based origin of the X-ray variations (Fukazawa et al., 2018). A significant correlation between -ray and radio flux variations suggests multiple high-energy emission sites located within a distance of 2.0 parsecs from the central black hole (Hodgson et al., 2018). Correlated X-ray and -ray variations in the source imply the same site for the X-ray emission, i.e. the power-law component of the X-ray emission is located within 2 parsecs of the central black hole.

Another check for the jet-based origin of the hard X-ray emission is the spectral slope from optical to X-rays (). Seyfert galaxies (where most of the X-ray emission is from the immediate vicinity of the central black hole) have 1.4 (Lusso et al., 2010). The slope could be a bit smaller for low-luminosity AGN like 3C 84 as is a function of luminosity (see Donato et al., 2001, for more details). For blazars (where X-ray emission is mostly from the jet), is 1.0-1.2 (Donato et al., 2001). We compared the optical continuum flux at 1500 from Evans & Koratkar (2004) to 2-10 KeV X-ray flux, and the estimated value for 3C 84 is 1.12, which is more similar to blazars than Seyfert galaxies favoring the jet-based origin of the X-ray emission.

Using the argument that X-ray and -ray emission in 3C 84 are co-spatial, we derive a lower limit on the Doppler factor using (Dondi & Ghisellini, 1995),


where is the X-ray frequency in keV and is the corresponding flux density in Jy, is the X-ray spectral index, is the -ray frequency, and is the variability timescale at high-energies (mostly used to constrain the size of the high-energy emission region). From the NuSTAR observations, we have = 0.16 Jy (flux density calculated from the X-ray flux measured in the 20-30 KeV band) and = 1.9. The source exhibits variations on multiple timescales at GeV energies, i.e. the rapid flares have doubling timescales of 10 hr (Tanada et al., 2018) while the long-duration outburst lasts for 8-12 months (250-350 days) (Hodgson et al., 2018). Using = 25 keV, = 0.9 ( 1), and = 10 hr, we obtained 1.4 and 3.9. Using = 300 days, we have 0.4 and 1.3. The estimated Doppler factor values () are similar to those obtained using the broadband spectral energy distribution and jet kinematics (, Fukazawa et al., 2018; Hodgson et al., 2018).

5. Conclusion

We present analysis of the NuSTAR observations of the central nucleus (3C 84) of the Perseus cluster. Because of the strong dominance of the diffuse thermal emission from the cluster, the true nature and origin of the power-law component from the central AGN remains under debate. The greater sensitivity at hard X-rays offered by NuSTAR allowed us to probe the AGN component in the source spectrum. A strong dominance of the thermal diffuse emission from the cluster is seen below 10 keV, while the power-law component takes over above 20 keV. No flux and spectral variations were observed in the two NuSTAR pointings during 4 days. The 20-30 keV energy flux range is  1.010 erg cm s ( = 7.410 erg s) with a photon index of 1.9.

We discussed the possibility of different AGN components contributing to the power-law spectrum. An accretion disk of a (3-8)10 M black hole does not shine in X-rays. The following arguments suggest that the hard X-ray and -ray radiation is emitted by the same population of relativistic electrons from the jet: (1) X-ray and -ray photon indices are quite similar ( 2), (2) -ray flux is just an extension of the X-ray flux, (3) the source exhibits similar long-term flux variations at X-rays and -rays, (4) no indication of cutoff energy is seen in the combined NuSTAR and BAT spectrum, and (5) the optical-to-Xray spectral slope is more similar to that of blazars rather than Seyfert galaxies. Our study therefore supports that the hard X-ray power-law component dominating the source spectrum above 20 KeV originates from the jet.

This research was supported by an appointment to the NASA Postdoctoral Program at the Goddard Space Flight Center, administered by Universities Space Research Association through a contract with NASA. BR thanks Erin Kara, Mihoko Yukita, Dave Thompson, Stefan Walker, Amy Lien, Hans Krimm, and Chris Shrader, for interesting discussions about AGN disk and corona.


  1. affiliationmark:
  2. affiliationmark:
  3. affiliationmark:
  4. affiliationmark:
  5. affiliationmark:
  6. thanks: NPP Fellow


  1. Abdo, A. A., Ackermann, M., Ajello, M., et al. 2009, ApJS, 183, 46
  2. Ajello, M., Rebusco, P., Cappelluti, N., et al. 2009, ApJ, 690, 367
  3. Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197
  4. Churazov, E., Forman, W., Jones, C., & Böhringer, H. 2003, ApJ, 590, 225
  5. Donato, D., Ghisellini, G., Tagliaferri, G., & Fossati, G. 2001, A&A, 375, 739
  6. Dondi, L., & Ghisellini, G. 1995, MNRAS, 273, 583
  7. Evans, I. N., & Koratkar, A. P. 2004, ApJS, 150, 73
  8. Fabian, A. C., Walker, S. A., Pinto, C., Russell, H. R., & Edge, A. C. 2015, MNRAS, 451, 3061
  9. Fabian, A. C., Zoghbi, A., Ross, R. R., et al. 2009, Nature, 459, 540
  10. Fukazawa, Y., Shiki, K., Tanaka, Y., et al. 2018, ApJ, 855, 93
  11. Harrison, F. A., Craig, W. W., Christensen, F. E., et al. 2013, ApJ, 770, 103
  12. Hitomi Collaboration, Aharonian, F., Akamatsu, H., et al. 2018, PASJ, 70, 13
  13. Ho, L. C., Filippenko, A. V., & Sargent, W. L. W. 1997, ApJS, 112, 315
  14. Hodgson, J. A., Rani, B., Lee, S.-S., et al. 2018, MNRAS, 475, 368
  15. Lusso, E., Comastri, A., Vignali, C., et al. 2010, A&A, 512, A34
  16. Magdziarz, P., & Zdziarski, A. A. 1995, MNRAS, 273, 837
  17. MAGIC collaboration, Ansoldi, S., Antonelli, L. A., et al. 2018, ArXiv e-prints, arXiv:1806.01559
  18. Mirzoyan, R. 2017, The Astronomer’s Telegram, 9929
  19. Mukherjee, R., & VERITAS Collaboration. 2017, The Astronomer’s Telegram, 9931
  20. Mushotzky, R. F., Done, C., & Pounds, K. A. 1993, ARA&A, 31, 717
  21. Nagai, H., Suzuki, K., Asada, K., et al. 2010, PASJ, 62, L11
  22. Oh, K., Koss, M., Markwardt, C. B., et al. 2018, ApJS, 235, 4
  23. Reis, R. C., & Miller, J. M. 2013, ApJ, 769, L7
  24. Sanfrutos, M., Miniutti, G., Agís-González, B., et al. 2013, MNRAS, 436, 1588
  25. Scharwächter, J., McGregor, P. J., Dopita, M. A., & Beck, T. L. 2013, MNRAS, 429, 2315
  26. Schmidt, R. W., Fabian, A. C., & Sanders, J. S. 2002, MNRAS, 337, 71
  27. Sikora, M., Stawarz, Ł., & Lasota, J.-P. 2007, ApJ, 658, 815
  28. Suzuki, K., Nagai, H., Kino, M., et al. 2012, ApJ, 746, 140
  29. Tanada, K., Kataoka, J., Arimoto, M., et al. 2018, ApJ, 860, 74
  30. Wilkins, D. R., & Fabian, A. C. 2011, MNRAS, 414, 1269
  31. Wilman, R. J., Edge, A. C., & Johnstone, R. M. 2005, MNRAS, 359, 755
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description