NGC 2207/IC 2163: A GRAZING ENCOUNTER WITH LARGE-SCALE SHOCKS

The spiral galaxies NGC 2207 and IC 2163, at a distance of 35 Mpc (1'' =170 pc; Elmegreen et al. 1995b) and partially overlapping in projection, are involved in a nearly grazing encounter with closest approach 200–400 Myr ago. We have studied this pair extensively (Elmegreen et al. 1995a, 1995b, 1998, 2000, 2001, 2006; Thomasson 2004; Struck et al. 2005) throughout the electromagnetic spectrum with Hubble Space Telescope (HST) WFPC2 observations in UBVI bands, ground-based Hα, Spitzer/IRAC (3.6–8 μm), and MIPS (24–160 μm) observations, Very Large Array (VLA) H i and radio continuum, and ¹²CO J = 1 → 0 (Swedish ESO Submillimeter Telescope (SEST)) observations and reproduced many of the observed features with N-body and smoothed particle hydrodynamics encounter simulations. Relative to IC 2163, the encounter is prograde and nearly in-plane, producing the observed eye-shaped (ocular) oval and two long tidal arms in IC 2163. Relative to NGC 2207, the encounter is retrograde with IC 2163 moving behind NGC 2207 toward the east. The short-lived ocular phase and other features of this system set strict constraints on the numerical model for the encounter. Along the rim of the ocular oval, there is a large-scale shock front caused by the inflow of gas responding to tidal torques. This observed shock is a signature of the early stages of prograde grazing encounters. The models in Struck et al. (2005) predict that disk or halo scraping between the companion sides of the two galaxies would push shocks at ∼200 km s⁻¹ into each other across a front 30''–60'' in length, with a mass transfer stream from IC 2163 impinging on NGC 2207. Evidence for this may be seen in the radio continuum image as enhanced radio emission from the outer part of the companion sides of NGC 2207 and IC 2163. According to the models in Struck et al. (2005), the two galaxies in this system will eventually merge.

To extend our previous studies of this galaxy pair, we observed NGC 2207/IC 2163 in X-rays and the ultraviolet with XMM-Newton and made new radio continuum observations with the VLA⁷ at λ6 cm at a resolution of 25, comparable to that of the ultraviolet and Spitzer/IRAC images. The goals of the X-ray observations were (1) to detect the predicted soft X-ray emission from diffuse hot plasma at the large-scale shock fronts produced by the grazing encounter and (2) to determine the number, location, and nature of bright, discrete X-ray sources in this pair. The XMM-Newton observations presented here are the first deep X-ray observations of NGC 2207/IC 2163. Note that this galaxy pair was not detected in the ROSAT All-Sky Survey (Voges et al. 1999).

We use our radio continuum, Spitzer infrared, optical, and XMM-Newton X-ray, UVW1 (effective λ = 2910 Å) and UVM2 (effective λ = 2310 Å) observations to study the large-scale shock fronts, the young star complexes, the NGC 2207 nucleus, and various discrete X-ray sources in these galaxies. In an outer spiral arm on the anti-companion side of NGC 2207, there is a morphologically peculiar star-forming region (called feature i by Elmegreen et al. 2000) which is the most luminous Hα, radio continuum, 8 μm, 24 μm, and 70 μm source in NGC 2207/IC 2163. At 24 μm, it accounts for ≃ 12% (Elmegreen et al. 2006) of the total emission from the galaxy pair. Feature i contains an opaque dust cone (400 pc in projected length) aligned nearly parallel to the minor axis of the projection of NGC 2207 into the sky plane. We present new results on feature i and its environs.

IC 2163 and NGC 2207 each have an SFR/M(H i) typical of normal spiral disks, with a star formation rate (SFR) deduced from Hα emission. The radio continuum flux density of NGC 2207/IC 2163 is about three times higher than expected from the IRAS far-infrared flux (Elmegreen et al. 1995b), yet neither galaxy contains a radio-loud active galactic nucleus (AGN). The global value of the Helou q_FIR parameter (the logarithm of the ratio of FIR to λ20 cm radio continuum flux density) is 1.81 for NGC 2207/IC 2163, whereas Condon (1992) finds the median value of q_FIR for galaxies that are not radio-loud AGNs to be ≃ 2.3 with an rms scatter of ⩽0.2. Other galaxies with a similarly low global value of the Helou q_FIR parameter are NGC 2276 (Hummel & Beck 1995) and the Taffy pairs UGC 12914/15 and UGC 813/6 (Condon et al. 1993, 2002; Peterson et al. 2012). The Taffy pairs, NGC 2207/IC 2163 and NGC 2276, represent three different types of interactions. (1) In the Taffy pairs, the bridge between the galaxies is prominent in the radio continuum, H i, H₂, and ¹²CO J = 1 → 0 emission. Condon et al. (1993, 2002) and Peterson et al. (2012) conclude that this results from a face-on collision between the two galaxies, and Lisenfeld & Volk (2010) suggest that 10%–30% of the collisional kinetic energy of the two colliding gas disks has been converted into the energy of relativistic particles in the bridge. (2) The collision between NGC 2207 and IC 2163 is a grazing collision with IC 2163 moving behind NGC 2207, not a face-on collision. (3) Like NGC 2207, the spiral galaxy NGC 2276 has enhanced radio emission in the outer part on one side of the galaxy. The lopsided appearance of NGC 2276 has been attributed either to ram pressure from hot intragroup gas (Rasmussen et al. 2006) or to a tidal interaction with the elliptical galaxy NGC 2300 (Hummel & Beck 1995; Davis et al. 1997). Rasmussen et al. (2006) observed the interacting galaxy pair NGC 2276/NGC 2300 with Chandra and found a shock-like feature in X-rays. NGC 2276 had already been observed in X-rays with the ROSAT High-Resolution Imager (Davis et al. 1997). We shall compare NGC 2207/IC 2163 with NGC 2276 and with the Taffy pairs. Explaining why the radio continuum emission from the galaxy pair is enhanced without a commensurate effect on star formation is important for understanding the conditions necessary for star formation in general.

Section 2 describes our new observations (X-ray and ultraviolet from XMM-Newton and λ6 cm radio continuum from the VLA) and the data reductions. Section 3 presents an overview of the system. Section 4 discusses star-forming clumps prominent in the Spitzer 8 μm (IRAC 4) image and/or in the XMM-Newton ultraviolet images (UVM2 and UVW1). Values of the flux density ratio S_ν(8 μm)/S_ν(λ6 cm) for these kiloparsec-size clumps are compared with those of giant H ii regions in M81 as an example of what is normal for an OB association. Section 5 presents our results on the large-scale shock fronts and comments on infrared-to-radio continuum ratios. Section 6 describes our X-ray results. Section 7 compares NGC 2207/IC 2163 with NGC 2276 and with the Taffy pairs. Section 8 is devoted to feature i and its environs. Section 9 summarizes our conclusions.

For this galaxy pair, we adopt the distance of 35 Mpc as in Elmegreen et al. (1995b), who used H₀ = 75 kpc s⁻¹ Mpc⁻¹.

2.1. XMM-Newton Observations

2.2. Radio Observations at λ6 cm

With the VLA, we observed NGC 2207/IC 2163 in the radio continuum at a central frequency of 4860.1 MHz for 92 minutes (on the target) in B configuration on 2001 April 14 and for 50 minutes (on the target) in D configuration on 1995 May 13. The observations were made with one intermediate frequency pair at 4885.1 MHz with a 50 MHz bandwidth and the other at 4835.1 MHz with a 50 MHz bandwidth. The phase center was R.A., decl. (2000) = 06 16 22.665, −21 22 06.87. For the B configuration (high-resolution) observations, the phase calibrator was 0606-223, the flux calibrators were 3C 286 and 3C 147, and the polarization calibrators were 3C 138 and 3C 286. No significant polarization was detected. For the D configuration (low-resolution) observations, the phase calibrator was 0607-157 and the flux standard was 3C 286. Our D configuration observations were not appropriate for a polarization calibration.

The AIPS software package was used for the data reduction. After calibrating the uv data from each of the VLA configurations separately and checking the separate maps, we combined the uv data sets from the two configurations and ran the AIPS task IMAGR with ROBUST = −2 to make and clean a map with a synthesized beam of 248 × 13 (HPBW) and BPA = 8°. After convolution to a circular beam of 25 (HPBW) and correction for primary beam attenuation, this became our final λ6 cm high-resolution image. In the field of interest, the maximum correction for primary beam attenuation was a factor of 1.2. In this image, which is displayed in Figure 1, a surface brightness of 1 mJy beam⁻¹ corresponds to T_b = 8.279 K and the rms noise is 0.016 mJy beam⁻¹, equivalent to T_b = 0.13 K. We find a total flux density from the galaxy pair in this image S_ν(4.86 GHz) = 0.132  ±  0.001 Jy, with about 20% of this from IC 2163. The single-dish observations of the Parkes–MIT–NRAO survey(Griffith et al. 1994) list S_ν(4.85 GHz) = 0.10 ± 0.01 Jy for NGC 2207; it is not clear whether the latter includes IC 2163.

Zoom In Zoom Out Reset image size

2.3. Additional Data

Other images of this galaxy pair that we use here are the WFPC2 HSTB-band image from Elmegreen et al. (2000), the Spitzer/IRAC and MIPS images from Elmegreen et al. (2006), the Hα image from Elmegreen et al. (2001), the VLA H i and line-free λ20 cm radio continuum images from Elmegreen et al. (1995b), and a radio continuum image at 8.46 GHz (λ3.5 cm) from the VLA public archives⁸ (Program AK 509) from 2003 January 14 observations. Table 2 lists the FWHM of the PSFs of the images we use. The VLA H i and radio continuum images have Gaussian synthesized beams (PSFs). The other images do not, and some of the non-radio images, such as the Spitzer 8 μm and 24 μm images and the XMM-Newton X-ray images, have significant side lobes.

Table 2. Images of NGC 2207/IC 2163

Image	FWHM of PSF	Reference
HST B WFPC2	∼018a	Elmegreen et al. (2001)
UVM2	18	This paper
UVW1	20	This paper
MOS X-ray	∼5''	This paper
MOS + PN X-ray	∼9''	This paper
Swift/XRT	18''	This paper
Spitzer 8 μm	24	Elmegreen et al. (2006)
Spitzer 24 μm	6''	Elmegreen et al. (2006)
Spitzer 70 μm	18''	Elmegreen et al. (2006)
Spitzer 160 μm	40''	Elmegreen et al. (2006)
λ3.5 cm radio continuum	117 × 35	VLA public archives
λ6 cm radio continuum	248 × 13	This paper
λ6 cm radio continuum	25	This paper
λ6 cm radio continuum	6''	This paper
λ20 cm radio continuum	135 × 12''	Elmegreen et al. (1995b)
H i	135 × 12''	Elmegreen et al. (1995b)
SEST 12CO J = 1 → 0	43''	Thomasson (2004)
Hα	42 × 36	Elmegreen et al. (2001)

Note. ^a∼018 for wide field and ∼009 for planetary camera.

Download table as:  ASCIITypeset image

The NASA Gamma-Ray Burst explorer mission Swift (Gehrels et al. 2004) observed the field of NGC 2207/IC 2163 as part of a monitoring campaign of SN 2010 jp (Smith et al. 2012) 12 times between 2010 November 15 and December 9 for a total of 26.2 ks (target ID 31869). The Swift X-ray Telescope (XRT; Burrows et al. 2005) operated in Photon Counting mode (Hill et al. 2004) and the UV/Optical telescope (Roming et al. 2005) obtained data in all six filters. In general, we use the Swift/XRT data only for consistency checks on the XMM-Newton pn data. However, for the X-ray source X3 and the extended X-ray emission from the NE radio ridge, we use the Swift/XRT spectrum in place of the XMM-Newton spectrum. For these two sources, the Swift/XRT data turned out have a better signal-to-noise ratio than the pn data because of the much lower detector background of Swift/XRT. Due to the low number of counts in the Swift/XRT data, we applied Cash statistics (Cash 1979) in doing spectral fits to the Swift spectra.

Figure 1 displays our λ6 cm radio continuum image and the WFPC2 HST B-band image of this galaxy pair, and Figure 2 displays our UVM2 image and the Spitzer 8 μm image overlaid with contours of the line-of-sight H i column density N(H i) associated with each galaxy. The two galaxies partially overlap in projection, with NGC 2207 in front. IC 2163 has an eye-shaped oval midway out in the disk and a tidal tail on the anti-companion side. The H i image of IC 2163 reveals a symmetric tidal bridge arm on the companion side; it is harder to discern in the infrared or optical because it lies behind the central disk of NGC 2207 but can be faintly traced in the HST B band (see Figure 3 in Elmegreen et al. 2001) and Spitzer/IRAC images. The rim of the eye-shaped oval (which we call the eyelids) in IC 2163 is outlined by optical, radio continuum, H i, and infrared emission and is particularly bright in the Spitzer 8 μm and 24 μm images. The eyelids are one of the large-scale shock fronts that we investigate in Section 5. They are produced by radial streaming and convergence of orbits due to tidal forces (see the velocity vectors of the model displayed in Figure 2 of Elmegreen et al. 2000), which concentrate old stars as well as young stars in the eyelids (Elmegreen et al. 1995b). The other large-scale shock is the long ridge of enhanced radio continuum emission on the companion (eastern) side of NGC 2207; in Figure 1, a tilted box 54'' (9 kpc) long is drawn around it. We call it the NE radio ridge. Aside from the highly luminous feature i, the companion side of NGC 2207 is substantially brighter in the radio continuum than its anti-companion side, with the brightest large-scale radio emission coming from the NE radio ridge. The radio continuum emission from the adjacent companion side of IC 2163 is also enhanced. This is evidence of either disk or halo scraping between the two galaxies.

Zoom In Zoom Out Reset image size

The NE radio ridge overlaps the optical spiral arms of NGC 2207 that are visible in the HST B-band image. These are not the outermost spiral arms of NGC 2207 on the companion side. The bottom panel in Figure 2 displays an outer H i arm of NGC 2207 cutting in front of the western part of the northern eyelid of IC 2163 and the eastern part of the southern eyelid. The extinction due to this outer arm of NGC 2207, which is seen backlit by IC 2163 in the HST image (Elmegreen et al. 2000), partly explains the faintness of the western part of the northern eyelid and the eastern part of the southern eyelid in the UVM2 image. This arm has values of N(H i) ⩾3 × 10²¹ atoms cm⁻². Assuming the outer disk of NGC 2207 is somewhat metal poor, Elmegreen et al. (2001) adopt for the relation between extinction and H i column density, A_v = (0.35 ± 0.18) × 10⁻²¹ N(H i). For a foreground dust screen, the ultraviolet extinction A(UVM2) = 2.6 A_v and A(UVW1) = 2.0 A_v (Savage & Mathis 1979). Thus, A(UVM2) on this arm ⩾2.7 ± 1.4 mag. In the λ6 cm image in Figure 1, another outer arm of NGC 2207 is visible cutting across just north of the nucleus of IC 2163.

There is also significant gas in the eyelids of IC 2163. The top panel of Figure 2 displays the N(H i) associated with IC 2163. Using the SEST, Thomasson (2004) detected ¹²CO J = 1 → 0 emission from both disks; the brightest ¹²CO J = 1 → 0 emission is from the central part of IC 2163 and has an integrated intensity, averaged over the beam, of 6 K km s⁻¹, equivalent to 10²¹ H₂ cm⁻² if we use the Milky Way conversion factor X_CO = N(H₂)/I_CO = 1.8 ± 0.3 × 10²⁰ from Dame et al. (2001). The resolution of SEST (43'' HPBW) is too low to tell if this ¹²CO J = 1 → 0 emission is mainly from the eyelids although the double-peaked nature of the ¹²CO J = 1 → 0 line profile of IC 2163 suggests this may be the case (Struck et al. 2005). Except for a bright ¹²CO J = 1 → 0 source on the northwest inner arm of NGC 2207, the SEST ¹²CO J = 1 → 0 image closely resembles the Spitzer/MIPS image at 160 μm of this galaxy pair (Elmegreen et al. 2006). Both have about the same resolution, and the close correspondence tells us that the cooler dust measured by the 160 μm emission has about the same distribution globally as the molecular gas, which is not surprising. We can infer the distribution of gas from the distributions of cooler and warmer dust, as indexed by the 70 μm and 24 μm emission, respectively. Aside from feature i, the brightest 70 μm and 24 μm emission in this galaxy pair is from the eyelids (see the figure in Section 5.1 below and the HiRes deconvolution of the 70 μm and 24 μm images in Velusamy et al. 2008), and thus the highest concentration of gas in IC 2163 is in the eyelids.

Elmegreen et al. (1995b) identified 11 unusually massive (10⁸–10⁹ M_☉) H i clouds in NGC 2207/IC 2163. Most of these are not sites of active star formation. Clouds N1, N5, and N6 (three of the six massive H i clouds associated with NGC 2207) are labeled in Figure 2. The center of H i Cloud N6 is at R.A., decl. (2000) = 06 16 25.560, −21 22 19.08. The brightest λ6 cm radio continuum source on the NE radio ridge and the prominent Spitzer infrared clump IR 12 (see Section 4) are 3'' west, 1'' south of the center of H i cloud N6; SN 2003H is 2'' east, 5'' south of the center of H i Cloud N6. H i cloud N5 obscures the ultraviolet emission from the eastern part of the southern eyelid of IC 2163. At the center of Cloud N5, N(H i) = 4.8 × 10²¹ atoms cm⁻², which corresponds to an A(UVM2) of 4.4 ± 2.2 mag. Elmegreen et al. (1993) discuss the formation of massive H i clouds and tidal dwarf galaxies by large-scale gravitational instabilities in the gas and suggest that H i Cloud N1 in the outer part of NGC 2207, where large z motions are creating a warp may be in the process of forming a tidal dwarf galaxy. The only stellar emission detected from Cloud N1 forms a bow-shaped arc in the northwestern part of the cloud visible, for example, in the Digitized Sky Survey image, in the blue-band plate in Elmegreen et al. (1995b), and in the UVM2 image in Figures 2 and 3. It appears that star formation has commenced in only this part of Cloud N1.

Zoom In Zoom Out Reset image size

Figure 3 displays the UVM2 image overlaid with Spitzer 8 μm contours and with Hα contours, respectively. Aside from the effects of extinction, this demonstrates the generally good correspondence between these three tracers of recent star formation. We use photometry in λ6 cm radio continuum, 8 μm, UVM2, UVW1, and Hα bands to study star-forming clumps in this galaxy pair at a resolution of 20–25 (0.3–0.4 kpc).

Elmegreen et al. (2006) did photometry in the Spitzer/IRAC bands of 225 bright clumps in this galaxy pair by using phot in IRAF with an aperture radius of 36 = 0.61 kpc, and a local background annulus with inner radius = 96 and outer radius = 156 concentric with the source aperture. The clumps are star complexes. In general, the source aperture includes collections of OB associations and older star clusters (see examples in Elmegreen et al. 2006). After registering the 8 μm and ultraviolet images to the same coordinate grid as the high-resolution λ6 cm image, we chose 28 prominent clumps in the 8 μm and/or UVM2 images; these are labeled in Figure 4, which displays the 8 μm emission. For each clump, Table 3 lists the 8 μm flux density from Elmegreen et al. (2006), the Hα flux from Elmegreen et al. (2001), and the λ6 cm flux density, UVM2 magnitude, and UVM2-UVW1 color measured with the same choice of source aperture and local background annulus as for the IRAC measurements. The uncertainty in the λ6 cm flux density of each clump is 0.04–0.05 mJy. In addition to free–free radio emission from the H ii regions, the source aperture is likely to include non-thermal radio emission from the spiral arms, some of which is removed by the local background subtraction. For UVM2 and UVW1, we took from the XMM-Newton User's Handbook the zero points for the magnitudes (defined such that Vega = 0.025 mag) and the conversion factors to get from the count rates to flux densities in mJy. The conversion factors are for white dwarfs and thus the values of the UVM2 flux density used in Table 3 for the ratio of 8 μm to UVM2 flux density are rough estimates.

Zoom In Zoom Out Reset image size

The numbering of the clumps is the same as in Elmegreen et al. (2006) except for the added clumps u1 at R.A., decl. (2000) = 06 16 20.298, −21 22 26.47 and rc1 at R.A., decl. (2000) = 06 16 17.996, −21 22 04.16. In Section 6, we find that clumps IR 11, IR 21, and rc1 coincide with discrete X-ray sources. The first 11 clumps in Table 3 are in IC 2163; the rest are in NGC 2207.

The values of A_v in Table 3 for the NGC 2207/IC 2163 clumps are upper limits obtained from the λ6 cm flux density and Hα flux for case B recombination with T_e = 10⁴ K by assuming all of the λ6 cm flux density after subtracting the local background to be optically thin free–free emission. For a number of the clumps, the A_v upper limits are quite large; this leads us to suspect that some clumps include significant non-thermal radio emission at λ6 cm. Within a given clump, the extinction is far from uniform as the HST observations found lots of dust features on very small spatial scales near star clusters, whose blue colors indicate little extinction (Elmegreen et al. 2001).

As an example of what is expected for the flux density ratio S_ν(8 μm)/S_ν(6 cm) of an OB association, we present data in Table 4 on the giant H ii regions in M81. We chose the 11 giant radio H ii regions in M81 which have the highest signal to noise in the λ6 cm radio continuum observations of Kaufman et al. (1987) and a radio spectral index α consistent with optically thin free–free emission. For these M81 H ii regions, Table 4 uses the λ6 cm radio continuum flux densities from Kaufman et al. (1987) and the Spitzer 8 μm, 24 μm, and GALEX NUV flux densities from Pérez-González et al. (2006) and finds the mean value of the ratio S_ν(8 μm)/S_ν(6 cm) = 19 with the standard deviation σ of the sample = 5, and the mean value of the ratio S_ν(24 μm)/S_ν(6 cm) = 36 with σ = 9. (The average measurement uncertainties are ±3 and ±5, respectively). Comparison of these ratios with global values for entire galaxies is given in Section 5.3. The M81 H ii regions have relatively low extinction. Two estimates of the extinction A_v are listed for each H ii region: A_v(K) from Kaufman et al. (1987) is derived from S_ν(6 cm)/S(Hα); A_v(PG) from Pérez-González et al. (2006) is obtained from the line ratios Hα/Hβ and Hα/Paα. Except for one H ii region, these two methods give the same values for A_v within the uncertainties of the radio data.

The 8 μm emission from star-forming regions is generally attributed to young massive stars exciting aromatic hydrocarbon (polycyclic aromatic hydrocarbon, PAH) emission bands or heating very small grains. Far ultraviolet emission from B stars as well as O stars can produce significant PAH emission via fluorescence (Peeters et al. 2004). Emission at 8 μm depends on mechanisms involved in the formation and destruction of PAH molecules as well as the local SFR (Calzetti et al. 2005; Pérez-González et al. 2006; Dale et al. 2007).

From IRAC color–color plots, Elmegreen et al. (2006) find that the 8 μm emission of most of the IRAC clumps in NGC 2207/IC 2163 is PAH dominated.

The mean value of the ratio S_ν(8 μm)/S_ν(6 cm) for the 15 prominent star-forming clumps in NGC 2207 is 18 with standard deviation σ of the sample = 8. Within the uncertainties, this mean value is the same as that for the giant radio H ii regions in M81. We excluded the clump rc1 since it is probably either a radio supernova or a background quasar (see discussion below and in Section 6). In contrast, the mean value of this ratio for the 10 prominent clumps on the IC 2163 eyelids is 34 with σ = 16. The clumps IR 5, IR 8, and IR 10 on the eyelids and IR 6 on the inner spiral arm of IC 2163 have values of S_ν(8 μm)/S_ν(6 cm) ⩾40. The clumps on the eyelids are more luminous at 8 μm but have the same mean S_ν(6 cm) as the clumps in NGC 2207 (aside from feature i). As the 10 eyelid clumps are detected in Hα, they contain OB associations, but they may in addition contain somewhat older stars with ages up to several tens of million years that were collected by the ocular compression front, and thus could be dominated at 8 μm by PAH excitation involving B stars rather than O stars. We discuss this further in Section 5.1.

Although the source aperture used in Table 3 to measure the clumps is much greater in linear diameter (1.2 kpc versus 300 pc) than for the measurements of the M81 giant H ii regions and thus may include emission, such as non-thermal radio emission from the spiral arm, unrelated to the OB associations, a number of clumps in NGC 2207/IC 2163 have a value of the 8 μm to λ6 cm flux density ratio similar to those of the M81 giant H ii regions. Either subtraction of the local background has sufficiently removed the emission unrelated to the OB associations, or these are examples at 8 μm analogous to the usual FIR–radio continuum correlation for galaxies (see Section 5.3)

The three brightest discrete λ6 cm sources in NGC 2207/IC 2163 are the clumps IR 20 (which is feature i), rc1, and IR 12 in massive H i cloud N6. All three have low values of the 8 μm to λ6 cm flux density ratio compared to the M81 giant H ii regions. Non-thermal radio emission is significant in these three sources. Feature i is the most luminous radio continuum source and contains the most luminous Hα region in the galaxy pair. VLA snapshots at λ6 cm and λ20 cm by Vila et al. (1990) indicate that feature i is dominated by non-thermal radio emission, with a radio spectral index α (where S_ν∝ν^−α) ranging from 0.7 in its 1'' radio core to 0.9 averaged over a 75 × 75 box. Clump rc1, the second-brightest discrete radio continuum source in Figure 1, is unresolved in that image. Clump rc1 is reasonably bright in UVM2 but does not appear as a significant clump in Hα. The faintness in Hα is not the result of extinction, since rc1 is also rather faint at 8 μm. Thus, rc1 is a non-thermal radio source. It may be a supernova remnant (SNR), a radio supernova, or a background quasar. Because we detect rc1 as an X-ray source, we defer further discussion of it to Section 6. Clump IR 12 lies in the region of brightest non-thermal radio emission on the NE radio ridge.

Table 3 also lists the flux density ratio S_ν(8 μm)/S_ν(UVM2), which is sensitive to the amount of extinction, to the distribution of extinction, to age, and to star formation history. For the star-forming clumps that are bright in the ultraviolet, low values of this ratio indicate relatively low extinction. Very high values of this ratio may indicate lots of dust and considerable absorption. From the values of the S_ν(8 μm)/S_ν(UVM2) ratio, it appears that clumps IR 21, IR 26, u1, IR 138, and rc1 have relatively low extinction (e.g., values of A_v similar to the M81 H ii regions in Table 4), and that most of the clumps on the eyelids and feature i suffer high extinction. The four brightest clumps in UVM2 are u1, IR 21, IR 19, and IR 26. Relative to the Hα and radio continuum emission, the ultraviolet emission from u1 is displaced toward the outer edge of the arm. IR 138 is a bright UVM2 clump that is faint in Hα and 8 μm emission and not detected at λ6 cm. On the HST image, it coincides with a short, thin string of sources in the NW interarm of NGC 2207. Clumps u1 and IR 138 are probably slightly older star complexes with little dust and with ultraviolet emission mainly from B stars rather than O stars. SN 1999ec at R.A., decl. (2000) = 06 16 16.18, −21 22 10.1 (Van Dyk et al. 2003) is 12 E, 03 S of the center of IR 21. IR 21 could be bright in the ultraviolet due to shock excitation of circumstellar gas by SN 1999ec (see comments about the corresponding X-ray source in Section 8). IR 21 resembles IR 19 and IR 26 in being fairly bright in Hα, particularly prominent in UVM2, and in having values for the 8 μm to λ6 cm flux density ratio in the same range as the M81 giant radio H ii regions. A simple interpretation is that these three clumps are OB associations with less extinction than most of the other clumps and that the Hα, and most of the λ6 cm emission from IR 21 is from the H ii region, not from the Type Ib SN 1999ec.

For 11 of the clumps, the value of the 8 μm to UVM2 flux density ratio is greater than or equal to 120. These include feature i, 8 of the 10 clumps on the eyelids, and the one clump on the inner spiral of IC 2163. It is not surprising to find that feature i and the eyelid clumps suffer high extinction. HST observations (Elmegreen et al. 2006) reveal a large opaque dust cloud occulting part of feature i. The eyelids contain a large concentration of gas; this is in addition to the extinction from the outer arm of NGC 2207 cutting in front.

We consider the two large-scale shock fronts in this system, i.e., the eyelids of the eye-shaped oval in IC 2163 and the NE radio ridge in NGC 2207. Figure 5 shows the difference between the two shock fronts. The top panel in this figure displays the ratio of the 8 μm surface brightness I_ν(8 μm) to the λ6 cm radio continuum surface brightness I_ν(6 cm), and the bottom panel, the ratio of the 24 μm surface brightness I_ν(24 μm) to λ6 cm radio continuum surface brightness. To match the resolution of the 24 μm image, we made a λ6 cm radio continuum image with a synthesized beam of 6'' from the λ6 cm radio observations described in Section 2.2.

Zoom In Zoom Out Reset image size

Two pixels in feature i were hard saturated and thus blanked in the 24 μm basic calibrated data. For these two pixels, we substituted twice the 24 μm surface brightness at the FWHM of the PSF. The donut-shaped appearance of feature i and its environs in the I_ν(24 μm)/I_ν(6 cm) ratio image (lower panel of Figure 5) results because the radio images have been cleaned of side lobes (the diffraction patterns), but the Spitzer images have not.

The differences between the two large-scale shock fronts are also apparent from the comparison in Figure 6 between the Spitzer 24 μm and 70 μm emission, and the radio continuum emission at λ20 cm and λ3.5 cm. The eyelids are brighter than the NE radio ridge at 24 μm and 70 μm, whereas the NE radio ridge is appreciably brighter than the eyelids at λ20 cm. The value of the ratio of FIR to λ20 cm emission varies with location in the galaxy pair. The radio continuum emission from NGC 2207/IC 2163 is strongly non-thermal with a spectral index α = 0.92 (where S_ν∝ν^−α; Condon 1983), and thus non-thermal radio emission makes a smaller contribution at λ3.5 cm than at λ20 cm. At λ3.5 cm, the NE radio ridge is a bit brighter than the eyelids.

Zoom In Zoom Out Reset image size

5.1. The Eyelids

5.2. NE Radio Ridge

5.3. Comments about IR-to-Radio Continuum Ratios

The top panel in Figure 7 displays an XMM-Newton X-ray image in the 0.5–10 keV range obtained by combining data from the MOS camera with data from the pn camera and smoothing the image to get greater sensitivity. The same screening criteria as for the X-ray spectra (see Section 2) were applied. The bottom panel displays the 8 μm Spitzer image with contours of X-ray emission from the MOS camera data only and with the source extraction boxes for the X-ray spectral analysis overlaid. The X-ray image used in the bottom panel has better spatial resolution (FWHM of the PSF ∼5'') than the X-ray image in the top panel but lower sensitivity. For the discrete X-ray sources, we subtracted the local background by collecting background counts from a nearby region with the same area as the source box. With the MOS cameras, the entire field of interest fits onto a single CCD chip. With the pn camera, part of the southern eyelid of IC 2163 and parts of the discrete sources X2 and X7 in NGC 2207 fell in a gap between two CCDs

Zoom In Zoom Out Reset image size

The X-ray images in Figure 7 clearly show the nucleus of NGC 2207 and nine other discrete X-ray sources labeled in the figure, as well as extended X-ray emission. No X-rays are detected at the position of the nucleus of IC 2163. There is little activity in the IC 2163 nucleus; it is faint (or very faint) in the radio continuum, ultraviolet, X-rays, 8 μm, and 24 μm images. The nucleus of NGC 2207 is prominent in soft X-rays, hard X-rays, radio continuum, Hα, UVW1, and 8 μm images, but rather faint in the UVM2 image (as a result of extinction). The NGC 2207 nucleus is the only source in this galaxy pair that is bright in hard X-rays with E > 5 keV.

6.1. Extended X-Ray Emission from NGC 2207/IC 2163

Extended X-ray emission, which may be from hot galactic gas, is mainly concentrated in NGC 2207. The encounter models for this galaxy pair (Struck et al. 2005) predicted soft X-ray emission from diffuse hot plasma at the large-scale shock fronts. One goal of our XMM-Newton observations was to detect such emission. We find that neither the large-scale shock front along the eyelids nor the NE radio ridge appears enhanced in extended X-ray emission relative to the rest of this galaxy pair. X-ray absorption due to the large concentration of gas in the eyelids plus gas in the outer arm of NGC 2207 cutting in front of IC 2163 may explain why we do not detect significant, extended, soft X-ray emission from the eyelids. Part of the southern eyelid lies in the gap between two CCDs of the pn camera, but this is also where we find very high extinction in the ultraviolet for the star-forming clumps (Section 4). Extinction does not account for the absence of enhanced X-ray emission from the NE radio ridge as the NE radio ridge generally does not have high extinction.

In the models of Struck et al. (2005), the NE radio ridge is attributed to disk or halo scraping at a relative speed of ∼200 km s⁻¹. The post-shock sound speed $a = \sqrt{(}5/16) v = 112$ km s⁻¹, and thus the shock would heat the gas to a plasma temperature T = 1.5 × 10⁶ K, suitable for emitting soft X-rays. With this value for the shock speed v, we use the observed X-ray flux of the NE radio ridge to obtain an upper limit to the halo X-ray emission measure n²L, where n is the halo gas density and L is the line-of-sight path length through the shocked gas in the halo.

We use the 2010 Swift/XRT data to measure the X-ray flux of the extended X-ray emission from the NE radio ridge. Its spectrum is displayed as the bottom right-hand panel in Figure 8. With the foreground absorption column density fixed at the Galactic value from Dickey & Lockman (1990), an absorbed power-law model gives an energy spectral index α_X = 1.26 ± 0.60, an absorption-corrected flux F_0.3–10.0 = (7.3 ± 1.6) × 10⁻¹⁴ erg s⁻¹ cm⁻² (with little emission above 2 keV), and C-statistic per dof = 22/23 (where dof = degrees of freedom). The NE radio ridge is not unusual either in terms of its X-ray brightness or its X-ray spectrum. If the NE radio ridge were twice as bright in X-rays as other extended emission from NGC 2207, this would have been noticeable in the XMM-Newton X-ray image. Also, the X-ray spectrum of the NE radio ridge suggests that the emission could be mainly from a collection of X-ray binaries. We take as an upper limit to the halo X-ray flux from the NE radio ridge F_0.3–10.0 = 3.6 × 10⁻¹⁴ erg s⁻¹ cm⁻².

Zoom In Zoom Out Reset image size

To calculate a corresponding upper limit to the emission measure from shocked gas in the halo of NGC 2207, we integrate the X-ray cooling curve in Raymond & Smith (1977) over the range 0.3–2 keV for T = 1.5 × 10⁶ K and get a volume emissivity = 1.23 × 10⁻²⁴ erg s⁻¹ cm³ × (n²/4π). The flux is then 1.23 × 10⁻²⁴(n²L/4π)Ω erg s⁻¹ cm⁻², where the solid angle Ω subtended by the NE radio ridge is 1.83 × 10⁻⁸ sr. Equating this to the above upper limit of the halo contribution to F_0.3–10.0 from the NE radio ridge gives an emission measure n² L  ⩽  6.5 cm⁻⁶ pc. If the path length L through the shocked gas in the halo is 1–5 kpc, then the shocked halo gas density n ⩽ 0.081–0.036 cm⁻³, respectively. This is not an unreasonable gas density for a galaxy halo, but it may be low for the shocked part of a halo. However, the above calculation assumed that the post-shock temperature was given by the orbit speed. This temperature is probably comparable to halo gas temperature before the shock, which also comes from the depth of the galaxy potential well. In that case, the Mach number of the halo shock will be of order unity and the compression will not be large. The primary effect could be a compressed magnetic field and associated cosmic-ray acceleration, which produces the enhanced radio continuum, with a low level of X-ray from hot gas that is not bright enough to see here.

The observation that the X-ray emission from the NE radio ridge is not enhanced relative to the rest of the galaxy pair supports the suggestion that most of the bright radio continuum emission here arises in a region of much lower density than the thin disk. There is no evidence that a shock due to halo scraping has heated large quantities of gas to X-ray temperatures.

6.2. Discrete X-Ray Sources in NGC 2207/IC 2163

Like the extended X-ray emission, the discrete X-ray sources are mainly concentrated in NGC 2207. Aside from the NGC 2207 nucleus, most of the discrete X-ray sources lie on the spiral arms. Only a few of these correspond to the prominent IR or UV clumps discussed in Section 4.

The X-ray spectrum of each discrete source in Figure 7 plus the NE radio ridge (denoted RR in Table 7) were fitted by an absorbed power-law model with the absorption column density by the Milky Way fixed to the Galactic value from Dickey & Lockman (1990) and the absorption column density of the absorber at the location of NGC 2207 (z = 0.00941) left as a free parameter. The results of these fits are summarized in Table 7. The EPIC pn X-ray spectra of sources X1, X4, X6, X8, X9, and X10, and the Swift/XRT spectra of source X3 and the NE radio ridge are displayed in Figure 8. The MOS spectra of sources X2 and X7, which lie at the chip edge in the pn observations, are displayed in Figure 9. The pn and MOS spectra of X5 (the nucleus of NGC 2207) are displayed in Figure 10. Most of these X-ray spectra can be fitted with a power law with an energy spectral index of about α_X = 1.0 (equivalent to a photon index Γ of 2.0); source X2 has a significantly steeper index, and X5, has a more complicated spectrum. For source X3 and the NE radio ridge, we considered the Swift/XRT data to be more reliable than the XMM pn data because of the lower detector background of Swift. For the other sources with more than 10 counts in the 2010 Swift data, the results with Swift were usually consistent, with the results from XMM, given the uncertainties.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 7. X-Ray Spectral Analysis of X-Ray Sources Displayed in Figure 7^a

Source	R.A.2000b	Decl.2000c	NH, intrd	αXe	χ2/dof	F0.3–10.0f	N(H i)g
X1	06 16 17.94	−21 22 04.5	0.12+0.21−0.12	1.06+0.47−0.59	20/19	4.70	0.32 ± 0.05
X2	06 16 15.86	−21 22 08.5	0.00	2.17+1.62−0.56	10/6	5.40	0.32 ± 0.08
X3	06 16 18.83	−21 22 30.5	...	1.07+0.82−0.72	27/20h	8.10	0.28 ± 0.05
X4	06 16 20.34	−21 22 16.5	0.12+0.21−0.12	0.93+0.49−0.50	16/17	4.03	0.12 ± 0.03
X5i	06 16 21.89	−21 22 25.5	...	...	...	...	0.14 ± 0.04
X6	06 16 15.84	−21 22 34.6	0.15+0.15−0.13	0.97+0.46−0.30	11/10	8.82	0.23 ± 0.11
X7	06 16 17.29	−21 22 53.2	0.27+0.42−0.24	0.86+0.58−0.41	20/13	10.4	0.27 ± 0.10
X8	06 16 23.47	−21 22 18.5	0.03+0.12−0.03	1.13+0.51−0.13	20/17	4.30	0.27 ± 0.10
X9	06 16 24.95	−21 22 28.9	0.04+0.15−0.04	1.05+0.63−0.43	15/17	3.44	0.41 ± 0.06
X10	06 16 26.50	−21 22 13.4	0.35+0.27−0.15	1.03+0.40−0.27	18/24	9.80	0.44 ± 0.08
RRj	06 16 25.13	−21 22 23.4	...	1.26+0.60−0.59	22/23h	7.3	...

Notes. ^aThe EPIC pn data were used for sources X1, X4, X6, X8, X9, and X10. For sources X2 and X7, only the MOS data were used. For X5 (NGC 2207 nucleus), the pn plus MOS data were fitted simultaneously in XSPEC (as shown in Figure 10). For X3 and RR (the NE radio ridge), the Swift/XRT data were used due to better signal to noise. All fits include absorption by neutral gas in our Galaxy with an absorption column density N_{H, gal} = 1.13 × 10²¹ cm⁻² (Dickey & Lockman 1990). ^bR.A. in h m s. ^cDecl. in °, ', and ''. ^dIntrinsic absorption column density at the location of NGC 2207/IC 2163 (z = 0.00941) in units of 10²² cm⁻². ^eEnergy spectral index α_X for a single, absorbed, power-law model. ^fFlux in 0.3–10.0 keV band corrected for Galactic and intrinsic absorption, in units of 10⁻¹⁴ erg s⁻¹ cm⁻². ^gN(H i) of the galaxy pair in units of 10²² atom cm⁻² from the 21 cm line data of Elmegreen et al. (1995b). ^hDue to the small number of counts, Cash statistics were applied (Cash 1979). ⁱX5 = nucleus of NGC 2207. The simple absorbed power-law model does not represent the data. More complicated spectral models are listed in Table 8. ^jRR is the NE radio ridge, not a discrete source.

Download table as:  ASCIITypeset image

If the 10 discrete sources in Table 7 are at the distance of IC 2163/NGC 2207 and emitting isotropically, then each has an X-ray luminosity L_x ⩾ 5 × 10³⁹ erg s⁻¹. Thus, each of the nine discrete sources in the disks of this galaxy pair is a ultraluminous X-ray source (ULX) candidate. Since each could be a collection of sources, e.g., X-ray binaries, rather than a single object, Chandra high-resolution observations and re-observing to look for variablility are necessary to check on this. From Chandra data on 32 nearby galaxies, Colbert et al. (2004) find that most of the discrete X-ray sources in the disks of spiral galaxies have spectra that fit an absorbed power law with Γ ≈ 1–2, appropriate for high-mass X-ray binaries associated with accretion-powered black holes, low-mass X-ray binaries, and ULXs, and that only the Antennae merger pair (NGC 4038/39) contains more than three ULXs. From a study of ULXs in 82 galaxies, Swartz et al. (2004) conclude that 14% of ULX candidates in spiral galaxies are probably background sources. Only the following four galaxies in their sample have more than five ULXs: M82, M51, NGC 4038, and NGC 4486. From Chandra observations of the interacting starburst pair NGC 7714/7715, Smith et al. (2005) identify 11 candidate ULXs, only two of which are more luminous than the faintest discrete X-ray source listed in Table 7 for NGC 2207/IC 2163. If most of the candidates in NGC 2207 are ULXs that would be a greater number than in a typical galaxy.

Note that Chandra observed NGC 2207/IC2163 on 2010 July 18 for a total of 13 ks. While most sources listed in Table 7 appear to be point sources, sources X2 and X8 appear to be diffuse and source X10 clearly consists of at least two sources.

For comparison with the intrinsic absorption column densities N_{H, intr} obtained by fitting the X-ray continuum, Table 7 lists the column density N(H i) of the galaxy pair as measured in the 21 cm line VLA observations (Elmegreen et al. 1995b), averaged over the X-ray extraction aperture and not corrected for helium. In addition to N(H i), there is a significant H₂ column density in much of this galaxy pair (Thomasson 2004). The discrete sources, X2, X8, and X9 have no significant intrinsic absorption along the line of sight, which places them in the layers of the NGC 2207 gas disk closest to the observer. Source X1 may lie close to the midplane of the NGC 2207 gas disk. Most of the other discrete sources are more deeply embedded or located toward the farther side of the NGC 2207 gas disk.

We have the following identifications or possible identifications of the discrete X-ray sources (see Figure 4 for the labeling of the 8 μm or ultraviolet clumps): X5 is the nucleus of NGC 2207, X10 is in the star-forming clump IR 11 on the eyelid of IC 2163, X1 coincides with clump rc1, and X2 corresponds with clump IR 21, which contains SN 1999ec. The value of N_{H, intr} for source X10 is consistent with its location in clump IR 11 on the eyelid of IC 2163 behind an outer arm of NGC 2207; the H i column density of NGC 2207 at X10 is 0.32 × 10²² atom cm⁻² and thus the fitted value of N_{H, intr} places X10 behind NGC 2207 but on the nearer side (relative to us) of the gas layer in the eyelids. The coincidence between X10 and star-forming clump IR 11 suggests that X10 contains a high-mass X-ray binary. Alternatively, Smith et al. (2005) note that some discrete X-ray sources in star-forming regions may be due to SNRs with high-mass progenitors, rather than high-mass X-ray binaries. Source X5 (the nucleus of NGC 2207) is the brightest X-ray source in the entire field. As a matter of fact it is the only X-ray source in the hard X-ray band above 5 keV. We shall discuss sources X5 and X1 in Sections 6.3 and 6.4, respectively, and source X2 in Section 8.

6.3. The Nucleus of NGC 2207

6.4. Source X1

Figure 11 provides a detailed view of X-ray source X1 with X-ray contours from the MOS data overlaid on the HST B band in the left panel and overlaid on the UVM2 image in the right panel. The plus sign marks the location of the unresolved, non-thermal radio continuum clump rc1, and the width of the plus sign is the HPBW of the λ6 cm synthesized beam. The main X-ray emission is centered on a collection of blue star clusters, the most prominent of which lies close to the discrete radio source. With A_v/N_gas = 0.53 × 10⁻²¹ mag atom⁻¹ cm⁻² from Bohlin et al. (1978) for solar neighborhood metallicity, the intrinsic X-ray absorption column density of X1 is equivalent to A_v = 0.6^+1.1_−0.6 mag. The value of the 8 μm to UVM2 flux density ratio (see Section 4) suggests that the clump rc1 suffers little extinction. Other evidence of low extinction is the close resemblance between the B band and UVM2 images. The H i line profiles obtained by Elmegreen et al. (1995b) show no evidence for an absorption feature at rc1, but the H i data has low spatial resolution of 135 × 12''.

Zoom In Zoom Out Reset image size

Fitting a two-dimensional Gaussian plus a flat baseline to the unresolved source rc1 on our λ6 cm radio continuum image with 25 resolution gives S_ν(6 cm) = 1.12 ± 0.03 mJy. If the radio source rc1 is in NGC 2207 (rather than a background quasar) and is isotropic, it has a λ6 cm radio continuum luminosity L_ν(6 cm) = 1.6 × 10²⁰ W Hz⁻¹, which is 230 × L_ν(6 cm) of Cas A (taking L_ν(6 cm) of Cas A as 7 × 10¹⁷ W Hz⁻¹ from Weiler et al. 1989). This λ6 cm luminosity lies in the range of radio supernovae and in the range of the very brightest SNRs. Neff & Ulvested (2000) find the three brightest, discrete, non-thermal radio sources in the Antennae merger pair have L_ν(6 cm) in the range 8 × 10¹⁹ to 2 × 10²⁰ W Hz⁻¹ and attribute these to SNRs. Each is slightly extended in their high-resolution radio continuum observations. The brightest of these Antennae SNRs is listed as a ULX candidate by Swartz et al. (2004) with a L_x in the 0.5–8 keV band of 18 ± 11 × 10³⁹ erg s⁻¹ and Γ = 2.54, i.e., similar in X-ray spectrum and somewhat more luminous in X-rays than our source X1, which has L_x = 7 × 10³⁹ erg s⁻¹.

Possible interpretations of the source X1/rc1 are a radio supernova, an SNR, a background quasar, or a chance superposition of a ULX in NGC 2207 with a background radio quasar. It is clear from Figure 6 that rc1 was brighter on 2003 January 14 when the λ3.5 cm (8.46 GHz) radio continuum observations were taken than on 1990 October 11 when the line-free λ20 cm radio continuum observations were made: in the λ3.5 cm image, rc1 is much brighter than every other source in the galaxy pair except feature i, whereas no local surface brightness peak is seen at the location of rc1 in the λ20 cm radio continuum image. Also, no local surface brightness peak at rc1 is visible in the figure in Condon (1983), which displays a λ20 cm radio continuum image of NGC 2207/IC 2163 with 76 × 6'' resolution. For a quantitative comparison, we measured the flux density of a 21'' × 21'' box centered on rc1 in the above λ3.5 cm image and in a λ6 cm image made from the UV data used in Figure 1, i.e., B configuration (high-resolution) observations on 2001 April 14 plus D configuration (low-resolution) observations on 1995 May 13. For the 21'' box, S_ν(6 cm) = 3.38 mJy and S_ν(3.5 cm) = 3.26 mJy. Part of this flux density is from emission unrelated to rc1. The λ6 cm flux density unrelated to rc1 in this box equals 3.38 mJy–1.12 mJy = 2.26 mJy (where S_ν(6 cm) = 1.12 mJy for rc1, as measured on the high-resolution image). If the spectral index α of the unrelated emission lies in the range 0.1–0.9, then scaling the 2.26 mJy from λ6 cm to λ3.5 cm gives S_ν(3.5 cm) in the range 2.14–1.37 mJy for the unrelated emission. Subtracting this from the measured S_ν(3.5 cm) = 3.26 mJy of the 21'' box in 2003 gives S_ν(3.5 cm) for rc1 in 2003 in the range 1.12–1.69 mJy. As this is greater than or equal to the λ6 cm flux density of rc1 in the high-resolution observations in 2001, either rc1 had increased in radio luminosity by 2003 and/or there is synchrotron self-absorption in the radio. Thus, it is unlikely that rc1 is an SNR. The source rc1 could be a background quasar or a radio supernova. A young radio supernova whose spectrum is inverted because the emission at lower frequencies is still absorbed (Weiler & Sramek 1988) is an example of a radio source with synchroton self-absorption that would be consistent with the observed properties of rc1.

The X-ray, radio, and ultraviolet emission from X1/rc1 may arise from parts of the source that differ in column density. For example, Bregman & Pildis (1992) find that the X-ray and radio observations of the very luminous SN 1986J in NGC 891 are consistent with a model in which the radio emission is from the high-temperature outward-moving shock whereas the soft X-ray emission is from the cooled material associated with the reverse shock.

NGC 2207/IC 2163, the spiral galaxy NGC 2276, and the Taffy pairs UGC 12914/15 and UGC 813/6 are unusual in having a λ20 cm radio continuum flux density a factor of 2.5–3 times greater than expected from the IRAS far-infrared flux. The λ20 cm radio continuum luminosities of these four systems range from 3.6 × 10²² W Hz⁻¹ for UGC 813/6 to 5.8 × 10²² W Hz−1 for NGC 2207/IC 2163, and they have strongly non-thermal values for the radio spectral index (Condon 1983; Condon et al. 1993, 2002).

In the Taffy pairs, the excess radio emission is visible as a bridge between the two spiral galaxies; the bridge is also prominent in H i and ¹²CO J = 1 → 0. Condon et al. (1993, 2002) interpret the Taffy pairs as resulting from a face-on collision between two spiral galaxies with pericenter less than the radius of the molecular disk. In the model by Lisenfeld & Volk (2010) for the radio synchrotron emission from the Taffy bridges, 10%–30% of the collisional kinetic energy of the two gas disks colliding with a relative velocity of 600 km s⁻¹ is converted into the energy of relativistic particles.

There are two reasons why the amount of kinetic energy available for accelerating particles to relativistic energies as a result of the galaxy encounter is at least an order of magnitude less in NGC 2207/IC 2163 than in the Taffy pairs. In the model by Struck et al. (2005), a warm gas stream (i.e., not H i) from IC 2163 impinges on NGC 2207 with a relative velocity of ∼200 km s⁻¹ as compared to 600 km s⁻¹ for the Taffy pairs. That alone gives a factor of nine down in v². In the Taffy pairs, it is a face-on collision of the two gas disks and thus the whole disks shock at the same time. The amount of gas involved in the gas stream in NGC 2207/IC 2163 is much less than in the Taffy pairs. Figure 2 shows that the H i tidal bridge arm of IC 2163 crosses NGC 2207 in projection at massive H i cloud N6 and then heads southwest, whereas the bright λ6 cm radio continuum emission in the "interarm" along the NE radio ridge is northwest of cloud N6.

Like NGC 2207, the spiral galaxy NGC 2276 has a long ridge of enhanced radio emission from the outer part of the disk on one side of the galaxy. In NGC 2207, we call this feature the NE radio ridge. The NGC 2276 radio continuum ridge is prominent in the images displayed in Hummel & Beck (1995) and Condon (1983).

NGC 2276 is a member of a small group of galaxies embedded in a large diffuse, intragroup, X-ray cloud. Unlike the NE radio ridge in NGC 2207, the large-scale bow-shock-like radio continuum ridge along the western edge of NGC 2276 is a site of active star formation and bright X-ray emission, visible in the Chandra observations by Rasmussen et al. (2006) and the ROSAT High-Resolution Image by Davis et al. (1997). The stellar and gaseous disks in NGC 2276 truncate just beyond this shock-like feature, whereas the NE radio ridge in NGC 2207 is not the outermost spiral arm on the affected side of NGC 2207. The lopsided appearance of NGC 2276 has been attributed either to a tidal interaction with the elliptical galaxy NGC 2300 (Hummel & Beck 1995; Davis et al. 1997) or to ram pressure from the hot intragroup gas (Rasmussen et al. 2006). From Chandra X-ray observations, Rasmussen et al. (2006) find that NGC 2276 is moving supersonically at 850 km s⁻¹ through the hot intragroup gas and that the ram pressure, which may have been acting for several × 10⁸ yr, could explain the observed compression of the H i gas and magnetic fields along the western edge of NGC 2276, leading to active star formation in this region. They argue that the X-ray emission from the western edge of NGC 2276 is dominated by hot plasma resulting from the vigorous star formation. This may be the reason why the NE radio ridge in NGC 2207 is not bright in X-rays, since it is not a site of extended vigorous star formation. If the scraping between NGC 2207 and IC 2163 and the bright radio continuum emission from the NE radio ridge are high off the disk of NGC 2207, this would explain why the observed H i column density contours are not unusually compressed on the NE radio ridge and why there is no widespread active star formation there.

Feature i is the most luminous radio continuum, 8 μm, 24 μm, and Hα source in the galaxy pair. Figure 12 displays the contours of emission from feature i and environs in various wavebands overlaid on the HST B-band image from observations made in 1996. As shown in this figure, feature i is bright in λ6 cm radio continuum and 8 μm emission but highly absorbed in the UVM2 band. Comparison of the UVM2 and 8 μm images of this region provides an extinction map. Elmegreen et al. (2000) point out the opaque (optically thick even in the I band) conically shaped dust cloud (labeled here) with a bright compact cluster at its apex. The core of the radio source and the brightest 8 μm and 24 μm emission are centered on this cluster. The UVM2 emission from the radio core and the conical dust cloud is highly absorbed and no X-ray emission is detected from either (see Figure 12).

Zoom In Zoom Out Reset image size

In the radio continuum, feature i is a core plus envelope source. Aside from a northern plume in the radio, it looks very similar in the λ6 cm and 8 μm images: the extended emission fills a triangular region with E–W base 8'' and N–S height 96 (= 1.4 kpc × 1.7 kpc), which includes the cluster arcs of Elmegreen et al. (2000) and two super-star clusters identified by Elmegreen et al. (2001). Just north of the filled triangular region, the λ6 cm emission forms a plume with position angle P.A. = 5°, whereas the 8 μm emission is a little west of north, i.e., along the arm at P.A. = −15°. We take as the definition of feature i the filled triangular region plus the radio plume. It has S_ν(6 cm) = 4.67 mJy.

We fit a simplified model consisting of the sum of two two-dimensional Gaussians plus a flat baseline to the λ6 cm emission from feature i to represent the core plus envelope plus general arm emission. Table 9 lists the results obtained by using our λ6 cm image with the circular 25 synthesized beam and the results obtained by using our original λ6 cm image which has higher resolution in the E–W direction (synthesized beam = 248 × 130 and BPA = 8°). These Gaussian models give for the core plus envelope S_ν(6 cm) = 4.6 ± 0.07 mJy, attribute roughly 60% of the emission to the core, and find the core is slightly elongated along the same line as the opaque dust cloud (position angles 40° for the core and about 40° + 180° for the dust cloud). The opaque dust cloud has a projected length of 2''–3'' (about twice the diameter of the radio core).

Table 9. Gaussian Fits to λ6 cm Emission from Feature i

Beam	Core			Envelope
HPBW	Sν(6 cm)	FWHM, P.A.		Sν(6 cm)	FWHM, P.A.
	(mJy)			(mJy)
250 × 250	2.90 ± 0.03	14 × 10, 39°		1.69 ± 0.05	56 × 19, 142°
248 × 130	2.66 ± 0.02	12 × 08, 43°		2.00 ± 0.07	55 × 26, 145°

Download table as:  ASCIITypeset image

From VLA B configuration snapshot observations in 1986, Vila et al. (1990) measured S_ν(6 cm) = 1.4 mJy for the 1'' core and 3.4 mJy for feature i as a whole, whereas our combined VLA λ6 cm data from B configuration (high-resolution) observations in 2001 and D configuration (low-resolution) observations in 1995 give 2.66 mJy for the 1'' core and 4.66 mJy for feature i as a whole on our highest resolution image (see Table 9), i.e., 1.3 mJy greater than the Vila et al. value for the core and 1.3 mJy greater than the Vila et al. value for feature i as a whole. Differences in flux density can arise from the missing short spacings and higher noise in the Vila et al. data and different ways of measuring the source. However, since the difference in flux density between 1986 and 2001 is the same for the core and for feature i as a whole and the λ6 cm flux density of the core in 2001 is nearly twice the Vila et al. value, we conclude that the λ6 cm flux density of the core increased by 1.3 mJy between 1986 and 2001. This corresponds to an increase in L_ν(6 cm) of 1.9 × 10²⁰ W Hz⁻¹ if isotropic. As this lies in the luminosity range of radio supernovae, it seems likely that in 2001 a radio supernova was present in the core. From these data, we cannot determine the year when outburst in the core of feature i occurred; Weiler & Sramek (1988) point out examples of radio supernovae that remained bright in the radio for a number of years. A supernova in the core of feature i may have been hidden from view optically by the high extinction. If our interpretation is correct, then NGC 2207 is remarkable in having had two optical supernovae (SN 1999ec and SN 2003H, both Type Ib) plus one radio supernova in recent years (i.e., between 1986 and 2003) and all with high-mass progenitors.

The simulation models by Elmegreen et al. (1995a) and Struck et al. (2005) for the NGC 2207/IC 2163 encounter estimate the inclination i of the main disk of NGC 2207 as 25°–35° (relative to face-on) with the minor axis of the projection at P.A. = 50°–70°. The near side (relative to us) of NGC 2207 is the northeastern side. The opaque dust cloud is aligned nearly parallel to the minor axis of the projection of the main disk of NGC 2207 into the sky plane. Elmegreen et al. (2000) point out a red V-shaped structure with apex at the radio core and opening to the north (see Figure 6 of that paper). The left fork of the V appears to be a continuation of the opaque dust cloud to the opposite side of the core, whereas the right fork of the V is aligned with the inside edge of the optical arm farther north, but straighter than most spiral-arm dust lanes, possibly as a result of extra compression of the original dust lane by the energetic events in feature i. The radio plume is midway between the two forks of the V. Given the orientation of the main disk of NGC 2207, the opaque dust cloud plus the left fork of the V could be outflow perpendicular to the plane of NGC 2207, generated at the radio core, i.e., the dark dust cone could be gas approaching us on the near side (relative to us) of the midplane and the left fork of the V could be gas receding from us on the far side and thus less prominent as an absorption feature because it is not obscuring light on the near side of the midplane.

Elmegreen et al. (2006) measure a 24 μm flux density S_ν(24 μm) = 248 mJy for the 24'' × 24'' field displayed in Figure 12. In the HiRES 24 μm image in Velusamy et al. (2008, which has a resolution of 19, a little better than that of the 8 μm and λ6 cm images shown here), this flux density comes from a 75 × 75 region centered on the radio core of feature i. This gives a 24 μm to λ6 cm flux density ratio for feature i of 248 mJy/4.67 mJy = 53, which is somewhat greater than the mean value S_ν(24 μm)/S_ν(6 cm) = 36 ± 9 for the M81 H ii regions in Table 4 but similar to that of the most luminous giant radio H ii region (K181) in M81. In Section 4 (see also Figure 5), we found that feature i is underluminous at 8 μm relative to its λ6 cm radio continuum emission when compared with the mean value for the M81 H ii regions and with the mean value for clumps in NGC 2207 containing OB associations. For feature i, S_ν(24 μm)/S_ν(8 μm) = 248 mJy/35 mJy = 7.1, which is high compared to the mean value of 1.9 for the M81 H ii regions in Table 4 and compared to the mean value for the SINGS sample (see Table 5), but similar to that of the dwarf (Im) starburst galaxy Mrk 33, which has S_ν(24 μm)/S_ν(8 μm) = 6.6 (Dale et al. 2007). Emission at 24 μm is from warm very small grains, whereas emission at 8 μm is a combination of PAH emission bands and continuum emission from warm very small grains. The high value of the 24–8 μm flux density ratio of feature i and the low value of the 8 μm to λ6 cm flux density ratio suggest some PAH destruction by the radiation field of feature i has depressed its 8 μm emission.

In the HiRES 24 μm image in Velusamy et al. (2008), feature i appears as a filled elliptically shaped region with minor axis/major axis ratio = 0.8 and minor axis at P.A. = 50°. Given the orientation of the disk of NGC 2207, it is probably a filled circular region in the disk of NGC 2207.

A long-slit optical spectrum taken by P. Martin (2000, private communication) with the Canada–France–Hawaii Telescope MOS prior to 2000 cuts E–W through the core of feature i and exhibits a normal H ii region spectrum with no unusual line ratios. The long-slit optical spectrum of SN 1999ec (8'' south-southeast of the radio core) taken by Matheson et al. (2001) has a P.A. of −17°, a resolution of 6.3 Å (380 km s⁻¹) FWHM at 5000 Å, includes cluster arcs in feature i, and crosses the east fork of the red V. North of SN 1999ec, this spectrum shows normal H ii region emission but the line profiles have a little asymmetry with a slightly more extended red wing, i.e., for the [O iii] λ5007 line, the center of a Gaussian fit is shifted systematically to the red relative to the intensity maximum with the shift increasing from 0.2 ± 0.3 Å at the southern cluster arcs to 0.6 ± 0.3 Å at the east fork of the red V. This may be an instrumental effect associated with non-centered sources projecting onto different positions on the chip. The optical emission-line velocities are generally consistent with the H i velocities given that the H i data from Elmegreen et al. (1995b) has low spatial resolution and a velocity dispersion of 56 km s⁻¹ at feature i. Neither of these optical spectra is suitable for looking for low-velocity outflows from feature i. At the feature i, positions sampled by these two long slits, photoionization dominates, and these optical spectra show no components in feature i at supernova or jet velocities.

Since photoionization dominates the optical spectra whereas the radio emission is non-thermal, we interpret feature i as a mini-starburst and compare it with the central starburst in M82. The central 50'' × 15'' starburst in M82 has S_ν(6 cm) equal to 3.4 ± 0.2 Jy (Hargrave 1974), a radio spectral index α of 0.5, and outflow over a wide range of solid angle (Seaquist & Odegard 1991). If the M82 central starburst were at the 35 Mpc distance of NGC 2207 instead of 3.6 Mpc, it would have a major axis of 51 (a little smaller than feature i) and S_ν(6 cm) = 36 mJy, which is 7.7 times the λ6 cm flux density of feature i as a whole. Except in the 1'' core of feature i, the λ6 cm radiation field in feature i is significantly less intense than the average value for the M82 starburst. We estimate the value of S_ν(24 μm)/S_ν(6 cm) of the central starburst in M82 for comparison with feature i. With a 25'' aperture centered on the M82 starburst, Kleinman & Low (1970) measured S_ν(22 μm) = 120 Jy. Extrapolating to 24 μm by taking I_ν∝ν⁻¹ gives S_ν(24 μm) = 130 Jy for a 25'' aperture. For the 50'' extent of the radio emission from the M82 starburst, S_ν(24 μm) should be somewhat greater than 130 Jy, and thus S_ν(24 μm)/S_ν(6 cm) should be somewhat greater than 130 Jy/3.4 Jy = 38. It appears that the ratio S_ν(24 μm)/S_ν(6 cm) = 53 for feature i is not unusual for a dusty starburst region.

The upper left panel in Figure 12 displays contours of the X-ray emission from source X2. The core of feature i is not detected as an X-ray source. The absence of soft X-rays could be the result of the high absorption seen from comparison of the UVM2 and 8 μm images. The X-ray emission is generally south and south-southeast of feature i, and the brightest X-ray knot is centered 24 north of SN 1999ec. The upper right panel in this figure shows that the UVM2 peak emission coincides with the supernova. As previously reported by Pooley (2007) from an archival search of Type Ib,c supernovae which used our XMM-Newton observations of NGC 2207, it seems likely that most of the X-ray emission from X2 is associated with the Type Ib supernova SN 1999ec. It is unusual for a Type Ib supernova to be seen as an X-ray source, and, interestingly, SN 1999ec is bright in X-rays 6 years after the optical SN was discovered. The X-ray emission may be from the supernova shock plowing into a circumstellar envelope. This makes it more plausible that the bright UVM2 emission coinciding with SN 1999ec is due to shock excitation of circumstellar gas by the supernova. The X-ray spectral index of X2 is steeper than typical of an X-ray binary.

We presented the X-ray and UV data observed by XMM-Newton and new λ6 cm radio continuum observations of the interacting galaxies NGC 2207/IC 2163. When combined with our previous observations in Hα, HST B band, Spitzer infrared, H i, and SEST¹²CO J = 1 → 0 and with Swift/XRT observations, these data allow us to see the effects of the grazing encounter in producing large-scale shocks, to study star complexes, supernovae, and the galactic nuclei in this pair, and to identify possible ULX candidates.

In X-rays we detect the nucleus of NGC 2207, nine possible ULX candidates, and extended X-ray emission, mainly from NGC 2207. One of the discrete X-ray sources corresponds to SN 1999ec and another has brightened in the radio in recent years and could be a radio supernova or a background quasar. The bright UVM2 and X-ray emission from SN 1999ec may be from shock excitation of circumstellar gas by the supernova. The strongest source in our XMM-Newton X-ray observations of NGC 2207/IC 2163 is the nucleus of NGC 2207. It is the only hard X-ray source in the field. The preferred model for its X-ray spectrum is a power law with a partial-covering absorber. Most likely this is a strongly absorbed, low-luminosity, Seyfert 2 AGN. However, optical spectroscopy is needed to confirm this assumption.

We measured values of the ratio of 8 μm to λ6 cm radio continuum flux density for the prominent, kiloparsec-size, star-forming clumps in the galaxy pair and compared them with those for the M81 H ii regions whose radio continuum is dominated by optically thin free–free emission. For the bright clumps in NGC 2207, the mean value of this ratio equals 18, which is the same as for giant radio H ii regions in M81, within the uncertainties. For the bright clumps on the rim (the eyelids) of the eye-shaped oval in IC 2163, the mean value is nearly a factor of two greater.

There are two types of large-scale fronts in this galaxy pair: the eyelids of the eye-shaped oval in IC 2163 and the NE radio ridge on the companion side of NGC 2207. Simulations suggest that the eyelid shock front is produced by both inflow and outflow, resulting in a convergence of orbits with vigorous star formation in a dusty environment. In the eyelids, the ratio of the 8 μm to λ6 cm surface brightness is two times greater than in the NGC 2207 and M81 giant H ii regions. This excess 8 μm emission could be the result of heating of PAHs by the current generation of OB stars in the H ii regions, in addition to an equal amount of heating by B stars in the same region from the previous ∼30 Myr of star formation. Unlike the flow of older stars through a steady spiral density wave, an ocular front should accumulate mass as it grows stronger. The excess 8 μm emission could also result from shock-heated H₂ emission or from fragmentation of dust grains down to PAH sizes by collisions in the shock region. The eyelids are located behind the outer part of NGC 2207; this may explain why they are not bright in extended soft X-ray emission.

The NE radio ridge in NGC 2207 is particularly bright in the radio continuum but not in any of the tracers of recent star formation. Values of the ratios of 8 μm to λ6 cm surface brightness and 24 μm to λ6 cm surface brightness are low on the NE radio ridge compared to those of the M81 H ii regions. Unlike the bright radio ridge in the outer part of NGC 2276, the NE radio ridge is not enhanced in extended X-ray emission relative to the rest of galaxy; this is probably because it is not a site of active star formation. The NE radio ridge, which previous models attributed to disk or halo scraping, is simply due to compression of the magnetic field and may be mainly in the halo on the back side of NGC 2207 (between the two galaxies). The X-ray flux of the NE radio ridge provides an upper limit of 0.036–0.081 cm⁻³ to the density of halo gas there if the line-of-sight path length through the shocked gas in the halo is 1–5 kpc. Having the bright radio continuum emission from the NE radio ridge originate high off the disk of NGC 2207 in a region with little neutral gas explains the lack of widespread vigorous star formation and the lack of unusually compressed H i column density contours on the NE radio ridge.

For NGC 2207/IC 2163, the global values of the ratios of infrared-to-radio continuum flux density in the Spitzer 8 μm, 24 μm, and 70 μm bands, and the IRAS FIR are significantly below the medians/means for large samples of galaxies. This is the result of excess radio continuum emission from large portions of NGC 2207, not just the NE radio ridge.

We find evidence that a radio supernova was present in the core of feature i in 2001. If so, then NGC 2207 had two optical supernovae and one radio supernova in recent years. The λ6 cm radio continuum luminosity of feature i on an outer arm of NGC 2207 is 13% of the central starburst in M82. In linear size, feature i is a little larger than the M82 starburst. Like the M82 starburst, feature i is a dusty starburst region with radio continuum emission that is mainly non-thermal and with a 24 μm to λ6 cm flux density ratio somewhat greater than the mean value for the M81 giant H ii regions. Whereas the M82 starburst has outflow perpendicular to the disk which is bright in X-rays, our only indication of outflow perpendicular to the disk in feature i is the peculiar morphology of the dust structures.

This work is based on observations obtained with XMM-Newton, an ESA science mission with instruments and contributions directly funded by ESA Member States and NASA. We gratefully acknowledge support from NASA Goddard Grant NNG05GR10G to M.K., D.G., and D.M.E. We acknowledge partial support of page charges by National Radio Astronomy Observatory. This research has made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, Caltech, under contract with the National Aeronautics and Space Administration. We also used a radio continuum image from the NRAO VLA public archives. We thank Tom Matheson and Perry Berlind for sending us FITS images of their long-slit spectrum of SN 1999ec. We thank Thangasamy Velusamy for sending us the HiRes FITS image of the Spitzer 24 μm emission. We thank Pierre Martin for providing figures displaying his long-slit spectrum of the core of feature i and his measurements of line ratios. This research has made use of the XRT Data Analysis Software (XRTDAS) developed under the responsibility of the ASI Science Data Center (ASDC), Italy. Swift at PSU is supported by NASA Contract NAS5-00136.