Radial I and V Band Photometry of NGC 4038/39: Exploring the Effects of galaxy-galaxy Interactions on Surface Brightness Profiles

The interacting galaxy pair NGC 4038/4039 was observed in both V and I, down to limiting surface brightnesses of 24.0 and 23.5 in those filters, respectively. These observations were used to obtain mean radial surface brightness profiles for both of the individual galaxies, in addition to a profile for the pair taken as one object. Both NGC 4038 and 4039 are found to have exponential profiles in their outer parts, but only NGC 4039 can be matched using parameter values which are similar to those of normal spiral galaxies. An particularly interesting finding is that the outer region of the interacting pair follows an law over a 100 arcsecond range in radius, even though the two galaxies are not yet fully merged. Implications of this are discussed.

1. Introduction

The "Antennae" galaxy pair (collectively Arp 244, individually NGC 4038 and 4039), is an ideal candidate for investigating the radial surface brightness profiles of interacting disk galaxies. The morphology of the pair is well reproduced by disk-disk encounter models (Toomre & Toomre 1972, Barnes 1988), so it is a safe bet that this system really is undergoing a merger. Also, the two galaxies overlap in a small region only, thus ensuring that the two disks can be analyzed separately.

Kuchinski et al. (2000) published the only radial surface brightness profile for this galaxy pair. Their surface brightness plot treated the galaxy pair as one object, rather than as a separate galaxy pair. The focus of their paper was galaxy morphology at various wavelengths, not the exploration of morphology changes due to interactions.

The observations and analysis reported here deal with finding azimuth-averaged radial surface brightness profiles from V and I band photometry of Arp 244, both in terms of the pair as a whole and the two galaxies treated independently. The use of the I band is particularly important, since it is less affected by dust obscuration than V images. The use of both bandpasses provides a check on features in the profiles that are strongly affected by dust (if any).

Arp 244 is at an approximate redshift of z=0.005477, implying a distance of 23 Mpc if Hubble's constant is taken to be 70 km/s/Mpc. At this distance, 1 arcsecond on the sky is equivalent to a linear distance of 0.11 kpc.

2. Observations

Calibration frames included Zeros, Darks, and Twilight and Dawn Flats. The number of exposures and exposure time is as follows: Zeros, 11 exposures for CCD temperatures of -5 and -15 degrees Celsius, Darks, 3 x 600 s, 3 x 300 s, 3 x 120 s, 6 x 15 s, Dawn Flats, 3 x 60 s in I, 3 x 30 s in V, Twilight Flats, 4 x 15 s in I, and 1 x 90 s, 1 x 120 s, and 1 x 180 s in V.

3. Reductions

1. The zeros were combined into one Zero
2. The darks of the same exposure time were combined
3. The combined darks were zero corrected and combined into one DarkZ
4. The flats were Zero and DarkZ corrected and combined by filter (FlatI and FlatV)
5. The Antennae and standard star images (both I and V) were then DarkZ and Flat(I,V) corrected
6. All images were cosmic ray cleaned using the IRAF task "Xzap." For Xzap, the parameter "box" was set to 6 to ensure that no stars were accidentally eaten, while ensuring that most of the cosmic rays were cleaned. The parameter "sigma" was also set to 6 so that the sky level remained intact.
7. Using 6 stars for reference in each NGC 4038/39 image, the I and V frames were registered and co-added (by filter). The co-added I and V images were then registered again.
8. The standard stars were then used to obtain the transformations from instrumental V and I to standard values, using Landolt's catalog (1983). To do this, the headers of the standard images had to be updated to include right ascension (RA), declination (DEC), sidereal time (ST), and the epoch (EPOCH). This header information was used to find the effective airmass for each standard star. The IRAF task "Phot" was used to measure the magnitudes of the stars. The following transformation equations were then solved: a) mV=V+v1+v2*XV b) mI=(V-VI)+i1+i2*XI (see the IRAF task "mkconfig" for an explanation of the meaning of these parameters).
9. The galaxy images were then normalized to counts per square arcsecond for a 1-second exposure. This was done by dividing each galaxy image by its total exposure time and the area of 1 pixel. The area of 1 pixel was measured to be roughly 0.52 square arcseconds per pixel. Also, the airmasses of the co-added I and V images were set to the arithmetic mean of the individual exposure times as an estimate of the effective mid-exposure airmass.
10. To do the photometry on the galaxy images, a sky value was subtracted from each of the co-added images. The V-band image had a near-constant value of 19.58 , which was subtracted from the image. The I-band image was found to have a sky gradient from the bottom of the image to the top (and slightly to the right/west) of 2.6% of the mean sky value of 18.43 . This was corrected by subtracting a gradient obtained from the average of a number of top/bottom areas in the image.
11. The IRAF task "ellipse" was then used to measure the mean radial surface brightness profiles of the sky-subtracted images. Profiles were found for each of the two galaxies individually (NGC 4038 and NGC 4039), in addition to profiles for which the galaxy pair is treated as one object (note: Arp244 will be used to designate the galaxy pair). The centers of the isophotes were set to be the centers of the bulge components of the galaxies. Profiles for all three objects (NGC 4038, NGC 4039, and Arp 244) were found using circular annuli. Additional profiles were found for NGC 4038 and NGC 4039 using image-measured ellipticities and position angles for the annuli. Treating the individual galaxies at ellipses in the co-added I band image, the major and minor axes were measured, giving approximate ellipticities of 0.32 and 0.61 for NGC 4038 and NGC 4039, respectively. The position angles were determined to be 87.6 and 53.2 for those two objects, assuming North is up in the image.
12. The IRAF task "invertfit" was used to calibrate the surface brightness profiles against the standard star solutions.

4. Results

Figure 1. False-color image of the NGC 4038/39, along with an I band image showing the tidal features (antennae) to the East.

Figure 2 shows the mean I (upper curve) and V (lower curve) radial surface brightness profiles of NGC 4038 using annuli with an ellipticity of 0.32, plotted against semimajor axis to the one-fourth power. For reference, a de Vaucouleurs profile (dash-dotted line) and an exponential profile (dashed line) are shown; the parameters values used are: =0.66,=15.2, =2.2, and =16.4. Notice that a de Vaucouleurs profile shows up as a straight line in this plot, whereas an exponential profile has some downward concavity. The overall profile of NGC 4038 shows three somewhat distinct regions: an inner point-spread-function (PSF)-flattened bulge, a flattened middle region showing the increased brightness from the star-forming regions in the galaxy, and an exponentially decreasing outer region. Both a de Vaucouleurs and an exponential profile matches the outer portion (of both the V and I data) fairly well. However, the curvature seems to suggest that the profile is closer to an exponential rather than a de Vaucouleurs profile. Also, the effective radius for the law is far too low; it would take a very large elliptical galaxy to have such an effective radius.

At first glance Figure 2 suggests that the outer disk has survived the encounter, since the scale height of 2.2 is within the range of 1 to 10 kpc for most normal spirals. This is probably not the case, however, since the central brightness is far too bright. The central value of 16.4 (for the I band) is at least one-and-a-half magnitudes brighter in I than any normal spiral galaxy with a comparable scale height (see Reshetnikov et al. 1993, Beijersbergen et al. 1999, assuming (R-I) is close to 1 and (B-I) is close to 2).

Figure 2. I- (upper) and V-band (lower) mean radial surface brightness profiles plotted as a function of the semi-major axis to the 1/4 power. The profile was obtained by averaging over annuli with an ellipticity of 0.32. The dashed line shows an example of an exponential disk profile with scale height 2.2 kpc and a central surface brightness of 16.4 . The dash-dotted line shows a de Vaucouleurs profile with an effective radius of 0.66 kpc and an effective brightness of 15.2 .

Figure 3 shows the mean radial surface brightness profile of NGC 4039, when averaged over annuli with an ellipticity of 0.61. At first glance, this galaxy appears to be following an law (from the straight line through the lower curve in the plot) over the entire range plotted. The dash-dotted curve has parameters =99.0 kpc,=25.4. On closer examination, however, an exponential matches the data better past 45 arcseconds. The model shown by the dashed line (=8.25 kpc, and =19.2 ) also has parameters that are within the normal range for spirals (scale length between 1 and 10 kpc and central brightness between 18 and 22 in I). This is an interesting result: it may be saying that the original outer disk of NGC 4039 has survived to a good extent. Unfortunately, without knowing the original characteristics of the galaxy definitive conclusions cannot be reached.

Figure 3. I and V band mean radial surface brightness profiles for NGC 4039 averaged over annuli with an ellipticity of 0.61. The dashed line shows an exponential disk profile with scale height of 8.25 kpc and a central surface brightness of 19.2 . The dash-dotted line shows a de Vaucouleurs profile with an effective radius of 99.0 kpc and an effective brightness of 25.4 .

Figure 4 shows the mean radial surface brightness profile for the two galaxies taken as one object, using circular annuli for the averaging. The striking feature about this plot is that an law (dash-dotted line, with =1.65 kpc,=15.9 ) fits the faint outer region from a=45 to 150 arcseconds. An exponential profile (=3.41 kpc, and =18.5 ) matches the data over about half of that range. The exponential law's deviation from the data occurs mostly in the region where the errors in the photometry are large (or becoming large), but there still seems to be a significant enough deviation to say that the de Vaucouleurs profile matches the data better. Also, the data seems to be becoming more concave up, rather than down, contrary to what is expected for an exponential profile. Possible interpretations of this result will be given in the discussion section.

Figure 5 shows a guassian smoothed and logarithmically scaled I band image. The scaling was done to show bring out the extended, low surface brightness component. As a reference to the scale here, the distance between the centers of the two bulges (bright white spots in the image) is about 67 arcseconds. This figure shows that the extended material really is there, and that the behavior is not just an artefact of improper sky subtraction. This figure does show that the non-uniformity of the sky was not fully removed when the sky gradient was applied (see the reductions section), but the area around the galaxy pair is not much affected by the remaining non-uniformity.

Figure 4. Mean radial surface brightness profiles in I (upper curve) and V (lower curve) for Arp244, taken as one object. The profile was generated by averaging over circular annuli. The dashed line shows an exponential disk with scale height 3.41 kpc and a central brightness of 18.5 . The dash-dotted line shows a de Vaucouleurs profile with an effective radius of 1.65 kpc and an effective brightness of 15.9 .

Figure 5. Gaussian smoothed and logarithmically scaled sky-subtracted I-band image.

Figure 6 is provided for comparison with the results of Kuchinski et al. (2000). This figure should be compared to Figure 6 below. The crosses in Figure 6 are data points taken from Figure 7. Notice that the data from the observations reported here are brighter by as much as one magnitude. There are two factors which may explain the discrepancies: first, the brightness of the data inside about 50 arcseconds depends on the center chosen for the annuli, and second, the data point at 100 arcseconds from Kuchinski et al. may be overly low because they may have overestimated the sky background. The difficulty with the center position arises from the fact that those authors did not clearly indicate the exact location of the zero point for the annuli when finding the surface brightness profile. Analysis showed that the surface brightness level can change by up to 1 magnitude for center position changes of less then 10 arcsecond or so. As for the data point at 100 arcseconds (and all of their data beyond that point), that data may be underestimated because the galaxy pair fills the frame in their images. Their sky level may not be correct.

Figure 6. The same as in Figure 4, but with surface brightness plotted against semimajor axis (radius). The crosses are data points taken from Kuchinski, et al. (2000).

Figure 7. Part of Figure 4 from Kuchinski et al. (2000) showing the mean surface brightness profile as a function of circular radius. This should be compared with Figure 6 of this paper.

5. Discussion

It must be noted that the two single-galaxy profiles may not be robust, since the elliptical annuli used in obtaining the radial profiles was based on the overall appearance of the galaxy as a whole in the I band, rather than on a true isophotal fit to the galaxy. It may be the case that the annuli used for averaging introduced an artificial smoothing to the profiles. Of course, this had to be done, since the true isophotes in the image aren't elliptical. The IRAF task "ellipse" was unable to fit ellipses to the images.

Of special interest is the overall radial profile of the galaxy pair, as shown in Figure 4 above. The profile is well represented by an law over roughly 100 arcseconds (11 kpc), even though the galaxy pair has not yet completely merged. This is surprising, since it is generally thought that laws form only in the late stages of the merging history, when the galaxy has (nearly) fully relaxed.

To get an idea of what might be happening here, note that there are two dominant processes that determine the configuration into which a gravitating system settles: phase mixing and violent relaxation (Binney & Tremaine 1987). Phase mixing is essentially the process of whereby compact groups of phase points spread into a larger region, resulting in a lower average phase space density. Violent relaxation, on the other hand, represents the changing energies of the stars due to the time-dependent potential of the interacting system. Both of these processes are at work to relax the system, possibly to an law for the complete merger of two disk galaxies. If the seen in the data presented here is correct, it would seem to suggest that the material in the outer envelope of the combined galaxy pair has mixed/relaxed well before the inner material. One would think it would be the other way around, since the dynamical timescales would be shorter in the inner regions. It would appear that, if the surface brightness profile is correct, mixing and relaxation may work in unusual ways in galaxy mergers. Like for the two individual galaxies, the circular aperture used for the measurements may not reflect the true behavior of the isophotes; no attempt was made to check this. It may be more appropriate to use annuli that match the faint (green and blue) halo in Figure 5.

It is worth considering whether the parameters to the law are similar to those of a large elliptical galaxy; it may be the case that an elliptical is being grown "from the outside in." From Figure 4 above, the model which reproduces the radial profile well has parameters =1.65 kpc, =15.9 . The effective radius is just about right for a large elliptical galaxy, but the effective brightness is at least 4 magnitudes too bright in I band for that effective radius (considering the data of D'Onofrio et al. 1994, assuming (B-I)=2). The current law is probably temporary; the parameters of the law must change.

The result of the law behavior of the radial surface brightness profile for the low surface brightness material is very interesting, but further study would be needed to test the reality of the measurement. To begin with, better photometry would be helpful. If the limiting magnitudes scale by collecting area of the telescope, then a 1 meter telescope would allow for I band measurements down to roughly 26.5 . This could be used to determine the true radial extent of the law that matches the data. Obtaining spectra to faint levels would allow the stellar population to be modeled. The underlying stellar populations must be known to properly determine dust absorption corrections.

Another avenue for further study of the unusual behavior would be to examine many N-body simulations of disk-disk mergers, to see if similar radial profiles are found for similar stages of evolution. This doesn't seem to have been addressed in the literature.

Acknowledgements

References