AMUSE-VIRGO. III. MID-INFRARED PHOTOMETRY OF EARLY-TYPE GALAXIES AND LIMITS ON OBSCURED NUCLEAR EMISSION

First we performed aperture photometry to identify detections for further study and determine upper flux limits for the non-detections. We used an aperture with a diameter of 12'' (which corresponds to about twice the size of the FWHM of MIPS at 24 μm) and a sky annulus between 20''and 32''. The noise level was determined by the distribution of fluxes in 1000 apertures randomly placed in empty parts of the sky.

The flux distribution was then fitted by a Gaussian. The resulting standard deviation, appropriately corrected for point-spread function (PSF) losses, was taken as the 1σ photometric uncertainty for this mosaic. Sources detected at less than 3σ significance are considered non-detections (with typical 3σ upper limits of ∼0.3 mJy). In total we detect 53/95 sources observed at 24 μm. Upper limits for the non-detected sources are given in Table 2. We note that the upper limits determined here are by design point-source upper limits (i.e., for a nuclear MIR excess, see below) rather than upper limits for the total galaxy emission.

Figure 1 shows that all the objects undetected at 24 μm have stellar masses log M_*/M_☉ < 10. Stellar masses were taken from Gallo et al. (2010, see also Table 2).

Zoom In Zoom Out Reset image size

One of the goals of this project is to detect possible nuclear MIR emission which may point toward obscured low-level AGN activity. However, in such a scenario, diffuse dust in the host galaxy becomes a significant source of contamination. In order to isolate any faint excess of nuclear dust emission, we must first account for possible diffuse dust emission originating from the host galaxy itself.

Some galaxies in our sample have extended (kpc-scale) dust features clearly detectable from the optical images (hereafter often called "optical dust," Ferrarese et al. 2006). The spatial resolution of Spitzer is not sufficient to resolve MIR emission from this dust in most of our objects. In such cases we cannot disentangle how much of the unresolved MIR flux comes from diffuse dusty structures or from a putative nuclear source. Therefore, we exclude any objects with optically detected dust in the host galaxy from further analysis. For reference, there are 18 objects with optically detected dust in our sample and we detect most of them at 24 μm, with the exception of VCC 1422 and VCC 1779. Thus, for this sample sources with optical dust have a higher incidence of MIR detections than sources lacking optical dust features. The MIR images of the 16 detected sources are shown in Figure 2.

Zoom In Zoom Out Reset image size

For 11 of the 18 sources with optical dust, MIR spectra are available in the literature (Bressan et al. 2006; Bregman et al. 2006; Smith et al. 2007; Leipski et al. 2009) or in the Spitzer archive. Six of these show clear signs for recent (star formation) activity as inferred from the presence of polycyclic aromatic hydrocarbon (PAH) features (VCC 0759, VCC 0763, VCC 1030, VCC 1154, VCC 1535, and VCC 1619). In addition, and conceivably as a result of the on-going star formation, those sources are typically also detected at FIR wavelengths (e.g., Temi et al. 2009). One additional source is M 87 (VCC 1316) for which the MIR is dominated by non-thermal emission (e.g., Perlman et al. 2007; Buson et al. 2009). The remaining four sources do not show any particular features hinting at activity in their MIR spectra. We note that the fraction of X-ray detections for MIR-detected objects is roughly comparable for sources with and without optical dust (∼60%, see Table 1).

After excluding the 16 sources mentioned above, the remaining 37 sources with clear MIR detections show no indications of host galaxy dust in the optical from their HST images. Twenty of these also show a nuclear X-ray source as detected by Chandra (Gallo et al. 2010, see also Table 1).

It has been shown that for early-type galaxies the total luminosity at 24 μm correlates well with the optical luminosity (e.g., Temi et al. 2007). Furthermore, the mid-infrared radial light profiles follow the optical light profiles very closely (e.g., Knapp et al. 1992; Temi et al. 2007, 2008). In addition, the MIR spectra of quiescent early-type galaxies reveal emission in excess of photospheric Rayleigh–Jeans predictions (e.g., Bressan et al. 2006). This is generally interpreted as emission from dust produced in the photospheres and circumstellar regions of evolved mass-losing stars, mainly on the asymptotic giant branch (AGB). These findings argue that most of the MIR emission in early-type galaxies is directly associated with the stellar population. This provides a valuable tool in this study, because the optical light profiles (scaled appropriately) can be used as a proxy for the bulk of MIR emission which is associated with the stellar population.

For the 37 MIR-detected objects without optical dust detection we produced and analyzed azimuthally averaged radial surface brightness profiles at 24 μm. Radial light profiles were created using the task ellipse within IRAF. The central question is whether these galaxies require an additional nuclear component to explain their MIR light profiles or if the observed MIR emission is in accordance with emission from the host galaxy (stellar light and dust emission associated directly with the stellar population).

In order to use the optical light profiles as a proxy for the MIR emission expected from the stellar populations, we also created radial profiles from the re-reduced ACS z-band images, which we convolved with an observed MIPS PSF, after aligning them to the MIPS images to ACS-pixel accuracy. Because only few MIR maps included a sufficiently bright star to serve as a MIPS PSF for our analysis, we chose to create an average PSF from 12 individual stars on our maps which were reasonably bright, not close to the edges of the maps, and also fairly isolated. The radial surface brightness profile for the MIR PSF was derived in the same manner as for the ACS and MIPS images.

The observed MIR surface brightness profiles were then fitted as the linear combination of the radial profile of the MIR-PSF convolved optical images (i.e., stellar host galaxy contributions) and the radial profile of the average MIR PSF (i.e., a possible nuclear excess of unresolved dust). We show the radial profile fits in Figure 3. In most objects an additional PSF component is required to match the optical light profile with the MIR light profile. This nuclear MIR excess can be small (e.g., VCC 1537), of medium strength (e.g., VCC 0731), or quite substantial (e.g., VCC 2092). We will discuss the likely origin of this nuclear excess emission in Section 4.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

As a second step in our analysis we modeled the observed MIR images using GALFIT (Peng et al. 2002). As we have seen in the previous section, a two-component model is well suited to describe the MIR radial profiles of the early-type galaxies in our sample. Therefore, we chose to fit the two-dimensional images also with two components: one component representing the host contributions, plus a PSF component representing a possible nucleus. As tracers for the host galaxy light distribution we again used the HST z-band images, which have been modeled previously by Ferrarese et al. (2006). We fixed the shape parameters (e.g., effective radius, Sersic index, axis ratio, and orientation) describing the host component to those determined by Ferrarese et al. (2006), leaving the absolute scaling and centroid as the only free parameters. For the PSF component we provided the same average PSF used in Section 3.3. For the vast majority of objects this two-component model provided a very good description of the MIR images. In a few individual cases we added an additional PSF to account for nearby stars. In the cases of VCC 1125 and VCC 2095 an additional disk component was needed to archive a satisfactory model of the galaxy. The model for VCC 1938 was more complex and required a second Sersic profile and an additional (off nuclear) PSF. The MIPS 24 μm images and the residuals after subtracting the best-fit model for all 37 sources for which the fitting was performed are presented in Figure 4.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The significance of PSF components in the individual objects was tested using similar techniques as described earlier for the aperture photometry: we placed 1000 random apertures in the residual images (i.e., after subtraction of the best-fit model) and fitted the resulting distribution of flux values with a Gaussian. The value of three times the sigma of this Gaussian (including an aperture correction) was converted into an AB magnitude limit for the PSF component. For six sources the fitted PSF component was fainter than the 3σ magnitude limit for the respective image and, thus, the limiting magnitude was taken as the bright limit for the PSF component in these cases. For the remaining 31 objects the nuclear PSF component is significant at a >3σ level.

The additional nuclear PSF required to fit the 24 μm radial profiles and images could be due to processes like star formation or low-luminosity AGN activity. However, radial variations in the age and/or metallicity of the stellar population can also result in a modified optical/MIR light ratio toward the center of the galaxies (e.g., Bressan et al. 1998; Piovan et al. 2003; Temi et al. 2005; Bressan et al. 2006). Since our method assumes that the stellar host component has a fixed z/24 μm ratio throughout the galaxy, part of the nuclear PSF component could in fact be due to a variation in the central properties of the stellar populations. In the following we will discuss such effects.

4.1.1. Residual Star Formation

One possible process to boost the MIR emission in the core would be residual star formation. In fact, the UV/optical colors of z < 0.1 early-type galaxies indicate that 10%–30% of them have experienced recent star formation with 1%–3% of the stellar mass having ages ⩽ 1–2 Gyr (e.g., Treu et al. 2005; Yi et al. 2005; Kaviraj et al. 2007). However, if this is the case for our objects, then the accompanying dust has to be located on very small scales (or be very diffuse) because it is not detected in the high-resolution ACS optical images.

This aspect can be explored further by studying MIR spectra of sufficient quality, which are available for 12 out of 37 objects in which no dust is visible from optical imaging. Genuinely quiescent early-type galaxies usually do not show aromatic features (like PAHs, strong indicators of recent activity) in their MIR spectra and have very low levels of FIR emission (e.g., Bressan et al. 2006; Temi et al. 2009; Clemens et al. 2010). This is consistent with the findings for the objects in our sample (e.g., Bressan et al. 2006; Bregman et al. 2006; Temi et al. 2009).

Clemens et al. (2009) report a number of "MIR excess" sources in their color–magnitude diagram (K_s−16 μm versus $M_{{\rm K}_s}$, their Figure 3) of early-type galaxies, which they interpret as objects being younger (or rejuvenated). We find only a few sources showing significant offset from most other galaxies in our sample for a comparable diagram in z−24 μm versus M_z (Figure 5). From the four sources falling noticeably above the rest of the sample, two show dust in the optical images which possibly boosts the 24 μm emission.

Zoom In Zoom Out Reset image size

While residual star formation may contribute some nuclear flux in a few individual objects, it appears unlikely to be responsible for the bulk of the nuclear MIR excess observed (to varying degrees) in almost all of our MIR-detected objects.

4.1.2. Stellar Age Gradients

4.1.3. Stellar Metallicity Gradients

Unlike age gradients, negative metallicity gradients are commonly reported for early-type galaxies (e.g., Mehlert et al. 2003; Sánchez-Blázquez et al. 2007; Spolaor et al. 2010; Pipino et al. 2010), but the SSP models mentioned above show only a mild dependence of the MIR emission on metallicity. In Figure 6, we present an optical/MIR color–color diagram which shows the colors predicted by the GRASIL SSP models (Silva et al. 1998; Bressan et al. 1998) for our filters in an age–metallicity grid. For most objects in our sample the host metallicities range from about solar to half-solar values. In addition, the coding of the symbols reveals the well-known mass–metallicity relation where more massive galaxies show higher (average) metallicities than less massive objects (e.g., Spolaor et al. 2010).

Zoom In Zoom Out Reset image size

We can use these models to predict the color changes owing to metallicity gradients and compare them with the observed nuclear colors. Using the median metallicity gradient for the high-mass objects in Spolaor et al. (2010) of ${\rm grad}\,[{\rm {\rm \textit {Z}}/{\rm \textit {H}}}]$ = −0.23 (and using their definition of the gradient), we estimate an increase of the metallicity Z by a factor of ∼1.5 between 15'' and 2'' (using the relation between [Z/H] and Z given in Cassisi & Salaris 1997). These radii were chosen for the calculation because at 15'' distance from the center we exclude any significant contributions from a nuclear PSF component and the colors are dominated by the host. At 2'' we are able to sample nuclear PSF emission (if present), while avoiding the very centers of the radial profiles, which are also not sampled by the optical spectra (Spolaor et al. 2010). For the models shown in Figure 6, an increase in metallicity by a factor of 1.5 results in a z−24 μm color change of about 0.3 mag in the metallicity range ∼0.008–0.02.

To estimate the observed nuclear color excess we use our radial profile fits presented in Section 3.3. The integration of the scaled z-band profile and the 24 μm profile in the inner 3'' in radius (which corresponds to the approximate size of the MIPS PSF at 24 μm) yields a measure of the color excess in the nucleus over the host galaxy z−24 μm color. The results are presented in Figure 7 in the form of a histogram of the nuclear color excess caused by the additional point source. In Table 2, we collect the fractional contribution of this additional nuclear point source to the flux within the inner 3'' radius. The histogram (Figure 7) reflects the impression from the radial profile plots and demonstrates that the nuclear MIR excess can change the nuclear colors significantly. The median of this (fairly broad) distribution falls close to the value we estimate from the models for a median metallicity gradient. Thus, the nuclear MIR excess for many of our objects with less prominent nuclear MIR PSFs is consistent with the metallicity gradient derived from optical observations. We emphasize that we only used a median value for the metallicity gradient to calculate the typical color change of about 0.3 mag. However, the large dispersion in measured metallicity gradients can in principle also explain the objects with more prominent nuclear MIR excess emission. At higher galaxy masses, metallicity gradients in a range of approximately 0.0 to −0.4 are observed (Spolaor et al. 2010, their Figure 4). This converts into a nuclear color change between zero, and ∼0.4 or up to a magnitude for lower metallicities (∼0.008) or higher metallicities (∼0.02), respectively, in the models shown in Figure 6. In addition, we see no dependency of the nuclear color excess on the stellar mass of the host galaxy.

Zoom In Zoom Out Reset image size

Table 2. AMUSE-Virgo III: Source Properties

VCC	ID	d	log M*	log MBH	z	log LX	$m_{\rm {\rm S}ersic}$	mPSF	fPSF	F24 μm
		(Mpc)	(M☉)	(M☉)	(mag)	(erg s−1)	(mag)	(mag)		(mJy)
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
0009	53	17.14	9.7	7.20	12.43	<38.15	...	...	...	<0.31
0021	73	16.50	9.0	6.70	13.74	<38.11	...	...	...	<0.33
0033	70	15.07	8.9	6.60	13.93	<38.25	...	...	...	<0.31
0140	58	16.37	9.4	7.00	12.86	<38.13	...	...	...	...
0200	71	18.20	9.2	6.80	13.55	<38.42	...	...	...	<0.30
0230	85	17.78	8.9	6.50	14.34	<38.62	...	...	...	<0.31
0355	28	15.42	10.3	8.00	10.59	38.77	...	...	...	12.4a
0369	16	15.85	10.4	7.80	10.66	39.26	13.80	16.10	0.28	12.28
0437	68	17.14	9.6	7.20	12.52	<38.17	...	...	...	<0.45
0538	90	22.91	8.9	6.50	14.74	<38.41	...	...	...	<0.28
0543	63	15.70	9.4	4.10	12.83	<38.27	...	...	...	<0.29
0571	72	23.77	9.4	7.10	13.63	<38.46	...	...	...	3.57
0575	54	22.08	10.8	6.60	10.10	<38.39	14.52	>18.02	0.40b	5.87
0654	23	16.50	10.4	7.00	10.54	<38.46	...	...	...	...
0685	21	16.50	10.6	8.20	10.09	39.14	...	...	...	29.8a
0698	49	18.71	10.0	6.40	11.69	<38.44	15.47	18.04	0.05	2.58
0731	7	23.33	11.7	8.80	8.15	39.00	11.31	16.06	0.14	110.01
0751	88	15.78	9.4	6.70	13.12	<38.29	...	...	...	...
0759	17	16.98	10.8	7.30	9.79	<38.18	...	...	...	13.6a
0763	6	18.45	11.7	9.10	7.59	39.73	...	...	...	66.6a
0778	34	17.78	10.2	7.50	11.27	38.56	14.73	16.94	0.00c	5.26
0784	32	15.85	10.3	6.80	10.77	38.62	14.55	17.06	0.23	6.04
0798	5	17.86	11.6	8.30	7.68	<38.43	...	...	...	52.4a
0828	36	17.95	10.2	6.70	11.25	<38.63	14.59	>17.56	0.00b	5.64
0856	57	16.83	9.5	4.50	12.91	<38.16	...	...	...	<0.35
0881	4	16.83	11.9	8.60	6.97	<38.64	...	...	...	27.5a
0944	24	16.00	10.4	7.20	10.39	<38.53	13.80	17.18	0.28	13.85
1025	40	22.44	10.4	7.10	11.22	38.92	14.65	17.20	0.28	5.49
1030	19	16.75	10.8	7.90	...	38.72	...	...	...	111a
1049	56	16.00	9.0	6.70	13.71	<38.08	...	...	...	<0.35
1062	14	15.28	10.7	8.20	9.62	38.47	12.78	17.46	0.19	28.43
1075	81	16.14	9.2	6.60	13.54	<38.12	...	...	...	<0.31
1087	60	16.67	9.6	4.80	12.51	<38.15	...	...	...	<0.33
1125	44	16.50	9.9	7.00	11.69	<38.46	16.11	>18.11	0.29b	4.22
1146	39	16.37	10.0	7.00	11.46	<38.33	15.14	16.89	0.24	3.83
1154	13	16.07	10.9	7.90	9.31	39.03	...	...	...	107a
1178	46	15.85	9.9	7.20	11.67	38.66	15.50	17.47	0.41	2.66
1185	96	16.90	9.1	6.50	13.81	<38.36	...	...	...	<0.29
1192	79	16.50	9.5	6.40	12.98	<38.68	...	...	...	...
1199	93	16.50	9.0	6.30	14.18	<38.14	...	...	...	...
1226	1	17.14	12.0	9.10	6.79	<38.49	...	...	...	747a
1231	11	15.28	10.8	8.10	9.38	38.60	12.67	15.66	0.16	33.02
1242	31	15.56	10.3	6.60	10.73	<38.50	14.19	15.68	0.14	9.60
1250	37	17.62	10.2	7.80	11.17	38.73	...	...	...	35.7a
1261	48	18.11	9.8	5.10	12.12	<38.42	...	...	...	<0.31
1279	26	16.98	10.5	7.60	10.47	<38.73	13.71	17.30	0.19	12.35
1283	47	17.38	10.1	6.30	11.53	38.54	15.06	>17.96	0.22b	3.67
1297	61	16.29	9.7	7.80	12.34	38.42	16.21	16.43	0.45	2.16
1303	41	16.75	10.1	6.90	11.35	<38.15	14.96	17.36	0.36	4.18
1316	2	17.22	11.8	9.40	7.19	41.20	...	...	...	154a
1321	35	15.42	10.1	7.70	11.05	<38.33	14.82	16.59	0.36	5.12
1327	43	18.28	10.1	7.60	11.53	38.68	...	...	...	6.06
1355	59	16.90	9.4	6.90	13.02	38.58	...	...	...	<0.27
1407	91	16.75	9.1	6.60	13.74	<38.35	...	...	...	<0.27
1422	50	15.35	9.6	7.20	12.24	<38.08	...	...	...	<0.30
1431	64	16.14	9.5	6.90	12.86	<38.66	...	...	...	<0.35
1440	84	16.00	9.2	6.70	13.40	<38.27	...	...	...	<0.29
1475	45	16.60	9.9	6.60	11.73	<38.34	15.48	17.75	0.24	2.62
1488	74	16.50	9.0	4.10	13.78	<38.14	...	...	...	<0.20
1489	99	16.50	8.7	6.30	14.73	<38.14	...	...	...	<0.31
1499	77	16.50	8.8	6.70	14.18	38.42	...	...	...	<0.29
1512	98	18.37	9.2	7.30	13.24	<38.23	15.32	17.84	0.42	2.97
1528	65	16.29	9.3	6.90	13.06	<38.13	...	...	...	<0.27
1535	8	16.50	11.0	9.20	...	<38.21	...	...	...	267a
1537	33	15.85	10.1	7.00	11.12	38.52	14.53	16.60	0.12	6.43
1539	95	16.90	8.9	6.60	14.06	<38.16	...	...	...	<0.27
1545	78	16.83	9.2	6.80	13.45	<38.16	...	...	...	<0.51
1619	29	15.49	10.2	6.60	10.81	38.68	...	...	...	9.8a
1627	83	15.63	9.1	6.60	13.64	<38.34	...	...	...	<0.58
1630	38	16.14	10.2	6.80	11.03	<38.29	14.57	>17.13	0.03b	5.91
1632	10	15.85	11.3	8.70	8.36	39.58	...	...	...	58.5a
1661	100	15.85	9.0	6.80	13.63	<38.12	...	...	...	<0.29
1664	22	15.85	10.6	7.70	10.08	39.95	13.43	15.77	0.36	17.20
1692	18	17.06	10.6	8.00	10.15	38.45	13.36	16.72	0.28	17.19
1695	66	16.52	9.5	7.00	12.83	<38.14	...	...	...	<0.26
1720	27	16.29	10.4	7.70	10.41	<38.54	13.99	16.29	0.34	10.31
1743	94	17.62	8.9	6.50	14.31	<38.20	16.75	17.76	0.69c	1.01
1779	75	16.50	9.0	6.70	13.75	<38.14	...	...	...	<0.31
1826	97	16.22	8.8	6.40	14.37	<38.12	...	...	...	<0.32
1828	89	16.83	9.1	6.60	13.72	<38.16	...	...	...	<0.32
1833	67	16.22	9.3	6.90	13.21	<38.11	...	...	...	<0.34
1857	80	16.50	9.0	6.70	13.81	<38.14	...	...	...	<0.30
1861	62	16.14	9.5	6.90	12.74	<38.12	...	...	...	<0.26
1871	52	15.49	9.5	6.90	12.68	<38.08	16.66	17.93	0.00	1.03
1883	30	16.60	10.4	6.80	10.40	38.35	13.68	16.20	0.39	13.45
1886	92	16.50	8.8	6.50	14.21	<38.14	...	...	...	<0.29
1895	76	15.85	9.0	6.60	13.73	<38.10	...	...	...	<0.27
1903	9	14.93	11.3	8.50	8.18	39.11	11.42	15.43	0.26	100.62
1910	55	16.07	9.5	7.00	12.75	<38.30	...	...	...	<0.60
1913	42	17.38	10.1	6.50	11.48	<38.46	15.24	>18.05	0.17b	3.13
1938	25	17.46	10.5	7.30	10.38	38.97	13.49	16.85	0.00	19.24
1948	82	16.50	8.8	6.40	14.34	<38.14	...	...	...	<0.27
1978	3	17.30	11.7	9.40	7.45	39.05	10.37	15.62	0.07	342.10
1993	87	16.52	9.0	6.40	14.11	<38.14	...	...	...	<0.25
2000	20	15.00	10.4	8.30	10.27	38.65	13.43	15.45	0.04	17.82
2019	69	17.06	9.4	6.90	13.01	<38.17	...	...	...	<0.28
2048	51	16.50	9.6	6.30	12.56	<38.12	...	...	...	<0.29
2050	86	15.78	9.0	6.50	13.88	<38.10	...	...	...	<0.27
2092	15	16.14	10.9	8.20	9.44	38.59	12.73	15.60	0.33	31.47
2095	12	16.50	10.6	7.60	10.17	38.71	13.73	15.75	0.41	24.89

Notes. Columns: (1) VCC source name; (2) ACS/VCS target number; (3) distance; (4) stellar mass of the host galaxy, in M_☉; (5) black hole mass; (6) HST/ACS z-band model AB magnitudes (taken from Ferrarese et al. 2006); (7) nuclear luminosity between 0.3 and 10 keV, corrected for absorption; (8) and (9) Sersic and PSF AB magnitudes from the image fitting at 24 μm; (10) fractional contribution of the additional PSF component to the nuclear MIR flux within the inner 3'' in radius as determined from the radial profiles; (11) total 24 μm flux, calculated as the sum of all the fitted components. Upper limits were determined as described in the text. Data for (3), (4), (5), and (7) taken from Gallo et al. (2010). ^aMIR aperture fluxes for dusty sources taken from Temi et al. (2009). ^bThese fractional contributions should be considered upper limits as their significance is low (see Figure 3) and the presence of a nuclear MIR excess is not supported from the two-dimensional images at a 3σ level. ^cSpurious result due to the influence of a strong, nearby point source (see Figures 3 and 4).

Download table as:  ASCIITypeset images: 1 2

In order to explore further the distinction between color gradients and nuclear point sources, we re-fitted the MIR images treating the effective radius as a free parameter and not including a point source. The effective radii from the new fits turn out to be smaller on average than in the optical, which is not surprising given the additional MIR light in the center of the objects. The visual inspection of the residuals from these fits compared with the results from the fixed R_eff (plus an additional nuclear PSF) fitting does not reveal significant differences in most cases. The model chi-square values for the fits also remain virtually unchanged. We conclude that the resolution of the data is insufficient to distinguish a true point source from a color gradient. However, the known stellar population gradients appear to be a more natural explanation than true MIR nuclear emission.

As mentioned in the previous section, our findings are consistent with those of Clemens et al. (2011) who also argue (based on a color–magnitude diagram) that the stronger concentration of the MIR light compared to the NIR or optical light is due to metallicity gradients.

Gallo et al. (2010) detect nuclear X-ray emission in 32 out of the 100 VCS objects (with a limiting 3σ flux limit of 3.7 × 10³⁸ erg s⁻¹ over 0.5–7 keV) and, after accounting for contamination from low-mass X-ray binaries, conclude that for most of them the X-ray emission originates from a low-luminosity AGN. SED studies find that weak-line galaxies and LINERs typically have (in νL_ν) MIR/X-ray ratios ∼1 (Ho 1999; Nemmen et al. 2010). For type-1 AGNs this ratio is typically around 5 (Elvis et al. 1994; Silva et al. 2004; Richards et al. 2006; LaMassa et al. 2009). For type-2 and increasingly obscured sources the ratio can go up to 10^3–4 (e.g., Silva et al. 2004; Zakamska et al. 2008; LaMassa et al. 2009).

In Figure 8, we compare the nuclear X-ray luminosity of the MIR-detected sources with the luminosity of the 24 μm PSF component. The left panels shows that for almost all the sources the MIR component is more luminous than the X-ray emission by a factor of ∼5–10. Assuming that both originate from nuclear activity, this ratio would require an SED typical of largely unabsorbed (or only very mildly absorbed) luminous AGNs. In contrast, allowing for a (possibly dominating) contribution of the stellar population to the nuclear MIR excess (see Section 4.1.3) would move the objects closer to the 1:1 line which is expected for low-luminosity AGNs. In addition, we cannot identify any correlation between the nuclear color excess and the X-ray luminosity of the sources. Thus, the observed MIR/X-ray ratio adds further circumstantial evidence in favor of metallicity gradients as the dominant contribution to the excess of unresolved MIR emission.

Zoom In Zoom Out Reset image size

More robustly, however, the observed MIR/X-ray ratio points to an important conclusion. There are no heavily obscured AGNs lurking at the center of these Virgo early-type galaxies which would exhibit large MIR/X-ray ratios. The black holes found by our survey in the X-rays have truly low bolometric Eddington ratios arising from low accretion rates and/or highly radiatively inefficient accretion.