Results of the digital co-addition of thirteen Schmidt films of the Virgo cluster of galaxies

Astronomy and Astrophysics Supplement Series, Jul 2018

We have digitally co-added APM scans of 13 Kodak TechPan films of the SE region of the Virgo cluster of galaxies. The field of the R-band films combined with the resolution of arcsec pixel-1 results in data file sizes of about 222 MBytes. The 13 scanned films have been aligned, co-added, corrected for vignetting effects and cleaned of stellar features. To illustrate the astrophysical uses of this technique, we present high contrast images of a sample of Virgo cluster and background galaxies. Several very faint, but very clearly seen features, such as the interaction between IC 3481, IC 3481A and IC 3483 and filaments outside of the common envelope of NGC 4438 & NGC 4435, can be seen for the first time. We present an image of the halo of M 87 showing its great extent, and an image of the central regions of the cluster which shows no real evidence for interactions between the galaxies in this region. We also present high-quality images of the previously-identified "jet

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

Results of the digital co-addition of thirteen Schmidt films of the Virgo cluster of galaxies

Astron. Astrophys. Suppl. Ser. Results of the digital co-addition of thirteen Schmidt lms of the Virgo cluster of galaxies A.C. Katsiyannis 1 S.N. Kemp 0 D.S. Berry 1 J. Meaburn 1 0 Instituto de Astrof sica de Canarias , C/ Via Lactea, E-38200 La Laguna, Tenerife , Spain 1 Department of Physics and Astronomy, University of Manchester , Oxford Rd., Manchester M13 9PL , UK We have digitally co-added APM scans of 13 Kodak TechPan lms of the SE region of the Virgo cluster of galaxies. The 6:2 6:2 eld of the R-band lms combined with the resolution of 2 arcsec pixel−1 results in data le sizes of about 222 MBytes. The 13 scanned lms have been aligned, co-added, corrected for vignetting e ects and cleaned of stellar features. To illustrate the astrophysical uses of this technique, we present high contrast images of a sample of Virgo cluster and background galaxies. Several very faint, but very clearly seen features, such as the interaction between IC 3481, IC 3481A and IC 3483 and laments outside of the common envelope of NGC 4438 & NGC 4435, can be seen for the rst time. We present an image of the halo of M 87 showing its great extent, and an image of the central regions of the cluster which shows no real evidence for interactions between the galaxies in this region. We also present high-quality images of the previously-identi ed \jet" and shell features around M 89. We also present an image of the whole eld which appears to show large variations in brightness of the intra-cluster medium across the region, with the brightest regions in the northern part in the central regions of the cluster, though we caution against this interpretation until we have investigated the large variation in emulsion sensitivity across individual lms more thoroughly. techniques; image processing | galaxies; structure | galaxies; haloes | galaxies; interactions | galaxies; clusters; Virgo cluster 1. Introduction The extensive 6:2 6:2 observing eld of the United Kingdom 1.2-m Schmidt Telescope (UKST) of the AngloAustralian Observatory (AAO), allows the whole (or signi cant fractions) of nearby clusters of galaxies to be imaged simultaneously. The most sensitive photographic emulsions are required for such observations as most regions of extragalactic objects are of very low surface brightness, and techniques such as photographic amplication or co-addition (Malin 1979) increased the depth to which such faint features of galaxies could be identi ed on Schmidt plates. At the end of the 1980s, increased computing capabil ity and disk storage facilities allowed for the possibility of digitally co-adding microdensitometer scans of whole Schmidt plates in order to increase the signal-to-noise ratio (S/N) and enable the detection of fainter extended features of galaxies, and produce a resultant co-added data array from which fully quantitative results could be obtained. As a preliminary project, Kemp & Meaburn (1991a) co-added Automatic Plate Measuring (APM) machine scans of 8 IIIaJ sky-limited plates of a eld containing the IC 4296 cluster of galaxies, for the purpose of detecting faint haloes around galaxies and other extensive low surface brightness features; the scanned area was about 4:5 4:5 . Discoveries were reported in Kemp & Meaburn (1991a,b, 1993, 1994, 1995) and Kemp (1994) and included: a giant halo of dimensions 600 200 kpc around a cD galaxy; tidal tails, detached laments and distortions associated with a pair of interacting galaxies; a warped disk of an edge-on spiral galaxy; a set of ve spectacular trails \emanating" from an apparently normallooking lenticular galaxy; a possible \shell" feature associated with a late-type spiral galaxy; faint optical emission spatially associated with the radio lobes of IC 4296; and a halo surrounding an irregular SMC-like galaxy which is almost perfectly circular in projection. The faintest features were estimated to be at approximately 27 mag arcsec−2 in the BJ band, while surface brightness pro les could be followed to about 1 mag arcsec−2 below this. A full photometric calibration of this data is now being carried out using Anglo-Australian Telescope (AAT) and South African Astronomical Observatory charge-coupled-device (CCD) data. Recently, the availability of hyper-sensitised Kodak Technical Pan (TechPan) lms (Kodak 1981; Russell et al. 1992) at the UKST has increased the possibilities for projects studying faint features of galaxies, at least for the red end of the spectrum. Phillipps & Parker (1993) carried out tests on two lms during the initial trials of TechPan lm at the UKST. These used the same OG590 lter and plateholder combination employed for normal \OR" Rband IIIaF exposures. The di erence in quality between the lms and the IIIaF plates was easily noticeable visually, the galaxy images on the TechPan lms having much greater clarity, due to the Tech-Pan emulsion having extremely ne grain size, low rms di use granularity and high resolving power (Phillipps & Parker 1993 and references within). In good seeing, sky-limited exposures with TechPan lms are reported to go a magnitude or more deeper than those with the equivalent IIIaF plates. They found a limiting isophote for galaxy surface photometry of 27 mag arcsec−2 (0.2% of the sky background) for these lms. The RCA CCD combined with the Anglo-Australian Telescope (3.9 m aperture) would take only 2 minutes to achieve the same S/N (i.e. pre unit area for low surface brightness sources) as obtained in the individual lms. However, the greater sensitivity of the CCD, for any project which involves observing extended objects or elds, is more than o set by the cumulative exposure involved in mosaicing a large number of elds. Modern CCDs, on 4 m-class telescopes, typically cover elds of the order of 0.01 square degrees compared with 40 square degrees for UKST plates. In addition, the mosaicing of CCD elds creates problems with matching discontinuities between adjacent elds. Of course, the two types of detector are complementary, in that objects identi ed as interesting in the co-added array can be targeted for multicolour CCD follow-up observations, which can give information on the stellar populations contained in them etc. and we will be reporting such work in the future. After the success of the initial digital plate co-addition project of Kemp & Meaburn (1991a), the Virgo cluster became the natural follow-up target, as it is the nearest rich cluster of galaxies and has a high galactic latitude which minimises foreground galactic contamination. Thirteen sky-limited TechPan R-band lms of the Virgo cluster were used. The 6:2 6:2 area of each lm was digitally scanned and the resultant arrays co-added as described below, to produce a digitally co-added array of unprecedented area. New software had to be developed to cope with the alignment and co-addition of such large data les, and for making a non-symmetric vignetting correction to the stacked image. With the co-addition of 13 lms galaxy surface photometry is expected to be possible to below 28 R mag arcsec−2 (0.1% of sky) over large angular scales. Initial priorities for investigation included: study of warping in the 20 or so edge-on disk galaxies contained in the area; giant haloes around elliptical galaxies, including M 87; comparison of the low surface brightness intra-cluster medium with maps of the cluster at other wavelengths e.g. X-ray and radio; and a general study of the amount of baryonic dark matter which is visible at very low surface brightnesses. In this paper we assume all Virgo cluster members are at a distance of 17 Mpc (Mould et al. 1995), resulting in a spatial scale of 5 kpc arcmin−1. This paper reports the software developed and utilised to produce the co-added array of unprecedented spatial area. Some preliminary results are then displayed and described for a number of galaxies in the eld. In some cases e.g. M 87 and M 89 well-known galaxies and previously discovered features are seen with the new clarity of TechPan lm and new depth produced by the co-addition. In other cases e.g. NGC 4435/4438 and IC 3481/3481A new low surface brightness features indicating interactions between galaxies or disturbances in the outer parts of galaxies have been discovered. Further papers will report these and follow-up CCD observations in more detail. 2. Observations and data reduction The thirteen lms of the South-East area of the Virgo cluster of galaxies were taken with the UKST between 1991 and 1994 using the Kodak TechPan emulsion with the OG590 lter. The lter has a lower wavelength cuto around 590 nm (10% transmittance at 580 nm and 90% at 600 nm), while the emulsion has an upper wavelength cut-o around 690 nm. This particular combination of lter and emulsion has a very small colour term compared with standard R band (Phillipps & Parker 1993). The eld of the plates covers an area of 6:2 6:2 , affected by vignetting. The lms used, dates and exposure times, are given in Table 1. The rst eight lms were from the UKSTU archive, while the remaining ve were taken speci cally for this project. A sixth lm (OR16137) was also taken for this project, but has double images and so was not used. The 6:2 6:2 area of the lms were scanned using the APM machine at the Royal Greenwich Observatory, Cambridge, at a pixel size of 32 m ( 200). The size of the digitized les was approximately 10500 10500 pixels, which occupied about 222 MBytes. The thirteen raw frames were combined into a single co-added image using applications from the Starlink KAPPA and CCDPACK packages (see Starlink User Notes 95 and 139, available on the World-WideWeb through the Starlink home page at Some of these applications were modi ed to handle the very large image size by breaking the input image up into \chunks" of 512 512 pixels for processing. The facilities of the NDF subroutine library were used to do this (see Starlink User Note 33). The following steps were involved in creating the nal co-added image: { Conversion of the tape FITS les to Starlink NDF for mat. { Alignment of the thirteen frames. For each frame, the centres of a common set of twelve stars located towards the corners of the images were found using a centroiding technique. Transformations were then found which mapped the star centres from each frame onto the corresponding star centres in the rst frame. These transformations were of the form: xout = C1 + C2:xin + C3:yin yout = C4 + C5:xin + C6:yin where the coe cients C1 to C6 were chosen to minimise the sum of the squared residuals between the transformed star positions (xout, yout) and the corresponding star positions in the rst frame. The typical values found for the coe cient values and the values corresponding to already aligned frames can be found in Table 2. A simple linear transformation was used even though the images cover a large area of the sky. This was possible because the original images were mis-aligned by a very small amount. The maximum shift obtained from these values does not exceed fteen pixels for a frame of 10500 10500 pixels. The accuracy of the alignment is less than a quarter of a pixel. { Photometric normalisation. The aligned frames were then compressed by replacing each box of 21 21 pixels by the mean of the pixel values within the box. An exposure factor and zero point o set were then found for each of the compressed images using the makemos application from the CCDPACK package. The full sized images were then modi ed using the zero point o set and exposure factor for the corresponding compressed image by subtraction of the zero point o set, and division by the exposure factor. This results in the images having a common mean photometric calibration. { Co-addition. The thirteen frames were then combined into a single stacked image by nding the median of the thirteen corresponding input pixel values at each incident with extended background features such as output pixel. The median was chosen rather than the galaxies. mean in order to eliminate, as far as possible, artifacts { Gaussian ltering. The data area containing the Virgo such as satellite tracks and scratches which do not oc- galaxy M 89 was smoothed by a two-dimensional cur at the same places in all thirteen images. RA and Gaussian function with a full-width at half-maximum Dec. information were also stored in the nal co-added (FWHM) box of 40 40 pixels. The resulting image array. was subtracted from the original data, so any \sharp" { Correction for the e ects of vignetting. A procedure features (like shells, jets, laments, etc) can now be was developed to reduce the e ects of the vignetting clearly seen. ifttttrrecagbmawhhaeeaatnatcsgieetusecoaneute.ikstgoolvicdilteoSngmtenreoidicmrng.rsrniaoaoym,rgbTgogulefa,yaaopecnhblina2lsgttmotdean,dih1heioctnaadsehklnstncgittegenuh.taoavedrierecnAnenoalfpoukltoeausrlreaastsltceendeehdcrihecafnxddteoeeseositsripartntmlrvwitejeoaadtocuiuveaogpnoniedrerngnonseeelgesteteldtstewpaitsttahmdwettthseieiedieenineeraamgddtdtienpstnniooaemetrigfturherttbeaitrpohwgcaeptsegamineloetohoelntdfbuenrhtcticreantmhsltoetoyyhm.chaeomnakeeTcobotgvgepafeavhreuaiurrlctigiecusocengnnthdn;uhussehvvse,seldontiiateotagagsnedhttdftnnnrnbeieealodeteydetbccehnttvhkytosttearveaheetenssorliidnddaeest----f iifttsoddCKKMsnaiaetaumCAacnihutacnpeptaDelePtbitehtri.nbiodPoesopwgcyr)uPnAtlaibanotegrwmtoesmehhaRaTlfiletoenlies7enlmr-twtd0blahedrIeb.iaCsiisneCswcemsncea.oCdipdciur3tnpSoyahD4sCteante8aPaasdrTCd1(nenAiJeDtastdfdakKhCiuonaonterldeTKrrnttdxdoshscc)tpn,eoapIhafooiC-lLraexpinasrobcaupdermkCr31srldePaai4ue9Cscegtsa8u9ltadiDtel1nh5ltoamstAgiiafniosDonwartnantrg.egcehsaccTaartyei.oelflmhhiar-2b6ceioaeuxan2o0rmdsig1at{-eedeaot.t2sd0eeidtt36o(radhwd,am0mfnleoeleuoeerdsrasSJexxneirltoatppngUdehafctoogKeraowcsspslkbuotuiuSirrnhetunrrereThnkeees---mate of the vignetted background level. The ratio of these two images was stored as a correction factor im- 3. Results age and expanded to the full image size using bi-linear interpolation. The vignetted areas of the stacked image We present ve areas of this data array, showing large were then multiplied by this correction factor image, faint extensions of several galaxies, revealing interactions, leaving the unvignetted areas unchanged. common envelopes and shells. { Removal of stellar features. When the stacked image is compressed, the stars in each 21 21 pixel bin are combined together and result in an extended surface 3.1. The giant elliptical galaxy M 87 and its vast brightness in the compressed image which depends on extended halo the number of stars in each bin. To remove this arti cial surface brightness, a procedure was developed which automatically removed all stellar-sized features from the full-sized stacked image, prior to compression. This procedure band-pass lters the image to attenuate heavily all features except those of stellar size. Bright features in this image were identi ed and the corresponding pixels in the original image were removed by setting them to a value which indicates that the the pixels should be ignored. This basic procedure was re ned to retain foreground stars which are co M 87 (NGC 4486) is the largest, most massive galaxy in the Virgo cluster. Its radial velocity is 1282 9 km s−1 (de Vaucouleurs et al. 1991) and it is likely to be the central object of the cluster. Arp & Bertola (1969) suggested a diameter of almost 1 at the 27 B mag arcsec−2 isophote level, also an image of its faint envelope has been presented by Kormendy & Bahcall (1974). Weil et al. (1997) presented a B-band image of M 87 by photographically co-adding ve IIIa-J UK Schmidt plates. An asymmetry was seen in the form of a di use \fan" of stellar material extending out to about 100 kpc to the SE along the projected major axis. They show that accretion of a small are several almost edge-on disk galaxies (NGC 4388, spheroidal galaxy into a larger potential can account for NGC 4402 and NGC 4425) and dwarf elliptical galaxies this structure. (NGC 4387 and NGC 4413). This eld is in the vignetted A normal contrast image of M 87 from the co-added region of the lms and there is a noticeable variation in data array is shown in Fig. 1a, in which four neighbour- the background across this eld at high contrast. Linear ing galaxies are indicated and enlarged at higher contrast. marks, which can be scratches on an individual lm or A high contrast image of the same eld is also presented satellite tracks, can be seen in the eld, and there are sevin Fig. 1b. The approximate dimensions of the halo in eral blemishes (\white marks") in the SE region. Each of this eld are 490 250, ( 250 125 kpc2), to a surface these defects only occurs in one of the set of 13 lms, but brightness of approximately 28 R mag arcsec−2, which is the variation in backgrounds over the set of 13 lms (inlarge, but not as large as some cD haloes (e.g. A3571, cluding the relative variations across lms caused by the Kemp & Meaburn 1991a, which is of the order of 600 variations in emulsion sensitivity, see Sect. 3.6) results in 200 kpc2), although it shows the same increase of ellip- median- ltering not being able to remove such features ticity with radius as most cD haloes. The asymmetry of completely, and this is especially noticeable at high conM 87 noticed by Weil et al. is independently con rmed by trast in the vignetted areas. our image, although it is not as prominent as it is on the M 86 (NGC 4406) is a giant elliptical galaxy (SO1(3) blue plates. E3) with a high negative redshift, −227 km s−1 (Binggeli We note that the X-ray pro le of this galaxy is much et al. 1985) compared to the average velocity of the clusmore extensive, Fabricant & Gorenstein (1983) show it ter ( 1150 km s−1; Faber et al. 1989) and its velocity reaching a radius of at least 1000 (500 kpc). The mass dispersion (762 km s−1, Rangarajan et al. 1995). Forman of material contained within this radius is of the order et al. (1979) suggest that the galaxy is on a radial orof 1013 M (Fabricant & Gorenstein 1983; Nulsen & bit, passing through the centre of the cluster about every B¨ohringer 1995). Whether or not there is a direct rela- 5 109 years. It is also a luminous X-ray object (LX 1:7 tionship between the material detected optically and that 1042 erg s−1) with most of the X-rays radiated by therobserved at X-ray wavelengths is unclear, as is whether mal bremsstrahlung from the interstellar medium, and the optical material also continues outwards to radii of has a well-studied plume of stripped material to the NW 500 kpc or greater at lower surface brightnesses. Note that (Rangarajan et al. 1995), reaching 120 (60 kpc) from the the curved \ laments" on either side of M 87 in the high centre. The dimensions reached by the halo of M 86 in contrast images are circular \defects" (caused by reflec- Fig. 2 are 240 160 (120 80 kpc2) which is comparable to tion of the plate holder, S. Tritton, private communica- the dimensions of the X-ray halo. Rangarajan et al. (1995) tion) on one of the original lms, not quite removed by and others have suggested that the intra-cluster medium the median- ltering addition. of Virgo is ram-pressure stripping the ISM of M 86, pro The dwarf ellipticals NGC 4476, 4478, 4486a and b are ducing the plume and other features visible at X-ray waveall optically close to M 87 in projection (see Fig. 1a). In the lengths. Nulsen & Carter (1987) reproduce a deep image present data, there is no obvious evidence for interaction of M 86 obtained by Malin, showing distortions in its outer between M 87 and any of these galaxies, as there are no isophotes due to regions of excess optical emission associfeatures indicating such interaction at the outer isophote ated with features in the X-ray emission, and they suggest levels of these galaxies (distortions, elongations, laments that stars are forming from the cooling hot gas in these etc.). The redshifts of NGC 4476, 4478 and 4486b all in- regions, though Rangarajan et al. (1995) suggest that the dicate that they are likely members of the Virgo cluster, excess optical emission is due to scattering of starlight though NGC 4476 would have a high relative velocity of from a high dust concentration. +500 km s−1 compared with M 87. NGC 4486a appears M 84 (NGC 4374) is a giant elliptical galaxy (E1) with to be a foreground object at +270 km s−1. NGC 4486b an extended halo of 210 180 (105 90 kpc2). It has an total has been classi ed as a prototype compact elliptical (Rood X-ray luminosity of 1:4 1041 erg s−1 detected by Fabbiano 1965) though the results of Prugniel et al. (1987) suggest et al. (1992) within 1.50 (7.5 kpc) of its centre (i.e. the that it could equally be classi ed as a normal elliptical, optical emission is considerably more extended than the but they nd NGC 4478 to have the truncated pro le of X-ray emission). The galaxy appears to be a weak radio a compact elliptical. source (3C 272.1; Laing et al. 1983). An organised pattern of Faraday rotation is detected by Laing & Bridle (1987) at 1.4 and 4.6 GHz, implying the existence of a magneto3.2. The centre of the cluster ionic medium in front of the radio-emitting plasma { Laing & Bridle also suggest that this medium may be responsible We present normal and high contrast images of the cen- for the di use component of M 84's X-ray emission. tral region of the Virgo cluster, covering a eld of 430 380 Despite their apparent proximity, our data o ers no (see Figs. 2a and b). The giant elliptical galaxies M 84 evidence for an interaction between M 84 and M 86. and M 86 are very prominent in this image, and there Although the outer isophotes as seen in Fig. 2b almost overlap, an isophote map shows no particular evidence of distortion or twisting of the isophotes of either galaxy caused by the other, other than that already referred to above for M 86 (which has been interpreted as due to star formation in the cooling hot gas or scattered starlight from dust, rather than due to interaction with M 84). Of course, the very high relative velocity and radial orbit of M 86 makes such an interaction unlikely. Other galaxies in the eld also show no evidence of interaction either with M 84 and M 86 or with each other, except for NGC 4388 (see below). Caon et al. (1990) considered NGC 4387, a dwarf E5 box galaxy (Nieto & Bender 1989), to be embedded in a common envelope of the overlapping haloes of M 84 and M 86. However, our data suggests that the haloes of M 84 and M 86 may not overlap physically, and trast image (Fig. 3a) shows an elliptical galaxy apparNGC 4387 shows no distortions or features that suggest ently without unusual features. Figure 3b was produced it is physically interacting with either of them. Also, al- by masking the original data (for more details see the last though NGC 4402 has an optical warped disk detected by paragraph of the data reduction section of this paper). A Warmels (1988), the deep image (Fig. 2b) appears more number of M 89's features ( rst identi ed by Malin (1979) normal. A deep, blue (IIIaJ) image including this area was in his blue plates) are shown in Figs. 3b and 3c. The jetproduced by Kormendy & Bahcall (1974), showing sugges- like feature, the innermost shell at the SSE (50 from the tions of an extended halo around these objects, although nucleus), a semicircular patch on the opposite side of the this may be due to non-uniformities in the emulsion. In galaxy to the jet (\condensation D" of Malin (1979), only our data there appears to be a brighter background in the seen on his deepest IIIaJ plates) are seen in Fig. 3b, with area surrounding M 84 and M 86, hinting at a common a much greater clarity. In the much deeper Fig. 3c one halo, but these galaxies lie in the vignetted area of our can see the di use area at the NW (feature C of Malin eld, near the NW corner (Fig. 6), and so we cannot be (1979) which appears as a circular patch on the R-band certain about the reality of this feature. images), the \jet", and an extended halo of 11:50 10:40 One of the most interesting galaxies of the central or 58 52 kpc. Note that the curved \ lament" seen Virgo area is the near edge-on Seyfert 2 Sab galaxy extending to the west of the galaxy is actually a circular NGC 4388 (Binggeli et al. 1985). It has a high systematic defect (caused by reflection of the plate holder) on one velocity with respect to the Virgo mean { its radial veloc- of the original lms, not quite removed by the medianity is 2523 km s−1 (Binggeli et al. 1993) so may be a non- ltering addition (see Sect. 3.1). member (Faber et al. 1989). However Binggeli et al. (1985) Clark et al. (1987) obtained CCD images of the jet-like classify it as a member, while faint extensions along the feature in R and V at the 1.3 m McGraw Hill telescope. major axis, principally to the west ( rst noted by Phillips They found the jet to be bluer than the rest of the galaxy & Malin (1982) and visible in Fig. 2b) may be evidence by V −R 0:15 and to have a total luminosity of 0.25% of of tidal interaction with other cluster members. Strong, that of the galaxy. They suggested that both the jet-like hard X-ray emission (Hanson et al. 1990) and extended, feature and shells were produced by a tidal encounter with soft X-ray emission (to a radius of 4.5 kpc, Matt et al. a smaller galaxy. Similar interactions are well simulated by 1994) has been detected, although its origin is still uncer- Hernquist & Quinn (1988) and references therein. tain (Antonelli et al. 1997). The dimensions of the halo in Fig. 2b are 80 30 (40 15 kpc2), so the optical halo is more extensive than the X-ray emission for this spiral 3.4. The interacting galaxies NGC 4438 and NGC 4435 galaxy. The apparently interacting pair NGC 4435 and NGC 4438 (radial velocities 775 15 km s−1 and 30 9 km s−1 3.3. The giant elliptical galaxy M 89 and its \jet" and respectively; Binggeli et al. 1985) are featured in Arp's shells catalogue of peculiar galaxies (Arp 1966) as number 120. Kotanyi et al. (1983) present Einstein X-ray observations M 89, shown in Fig. 3, is an E0 elliptical galaxy with a ra- of the area of NGC 4438 showing di use extended emisdial velocity of 321 12 km s−1 (Binggeli et al. 1985) and sion on the west of the galaxy. Kotanyi & Ekers (1983) is a member of the Virgo cluster. Malin (1979) reported an also saw such emission in the 1.4 GHz radio continuum. optical jet-like feature extending 100 from the core of the Both sets of authors proposed that the most likely explagalaxy, which he detected on sets of photographically am- nation for this emission was that the interstellar gas is pli ed and co-added UKSTU IIIaJ and IIIaF plates. He being swept out of NGC 4438 by a wind caused by the also saw three shell features concentric with the galaxy. dense surroundings of M 87 (the projected distance beAll the features were more easily seen on the blue IIIaJ tween NGC 4438 and M 87 is 250 kpc). Kotanyi & Ekers plates than on the red plates. Although the nucleus of (1983) did not believe that there was a tidal interaction the galaxy is a compact radio source, there was no evi- between the two galaxies because of the apparent lack of dence of a radio feature corresponding to the jet (Heeschen distortion of NGC 4435. However, Combes et al. (1988) 1970). However, as yet unpublished observations at 21 cm present a CO map of NGC 4438 showing a central comobtained with the VLA C array show an extended fea- ponent and some molecular emission from the NW of the ture attached to the nucleus of the galaxy and aligned galaxy. Molecular gas cannot be stripped from a galaxy with the optical jet. Sramek (1975 and references therein) by intracluster medium gas, arguing in favour of a tidal found the nucleus of M 89 to be variable at several radio interaction between the two galaxies. A numerical simulafrequencies, while Forman et al. (1985) found an unre- tion of the interaction of the two galaxies is also presented solved nuclear point source with an X-ray luminosity of 3 and they argue that the galaxies's tidal interaction caused 1040 erg s−1. the stripping of CO and star formation in the molecular Our images of the jet-like feature and shells from the clouds caused the NW X-ray and radio continuum feaco-added array are shown in Figs. 3a{c. The normal con- ture. Hummel & Saikia (1991) mapped the centre of the galaxy at 1.49 and 4.86 GHz and detected a morphologically unique (amongst the spiral galaxies) shell-like central radio source. The optical nucleus has been classi ed as a Liner (by Stau er 1982; Keel 1983 and Heckman et al. 1983), but may not be the real nucleus, mostly because it does not correspond in position with the radio nucleus. There are indications that star formation in the nucleus is the most likely explanation for its strange properties. Hummel & Saikia (1991) propose that around the nucleus, a bubble of plasma developed, expanded, and escaped by making a tunnel through the surrounding gas. Keel & Wehrle (1993) found two extended optical emission-line (H and [NII]) gas laments in the vicinity of NGC 4438. They propose di erent origins for the two laments and argue that only a variety of di erent mechanisms (interaction with the Virgo cluster's intergalactic medium, with NGC 4435 and nuclear energy release), can explain all of NGC 4438's peculiarities. Kenney et al. (1995) also detected these and other ionized laments in H and [NII] 1:5 − 3:5 kpc from the nucleus and 5− 10 kpc toward the west and southwest. They proposed that the laments are shock-excited and as their velocities are much less than the escape velocity for the galaxy, they suggest that these laments consist of gas disturbed by the collision and returning to the galaxy, and they delineate regions where hot gas is coming into contact with cold gas, creating layers of shocked gas at the interface. They further suggest that the features of the disturbed ISM of the system could be produced by a high-velocity ISM-ISM collision between the massive gas-rich galaxy NGC 4438 and the less massive, less gas-rich galaxy NGC 4435 (i.e. resulting from the likely tidal interaction between the two galaxies as simulated by Combes et al. 1988). NGC 4435 su ered a more severe e ect from the collision. According to Kenney et al. (1995), most of its gas is captured by the much bigger companion, and they expect deeper images to show a more disturbed halo morphology for this galaxy. Malin (1993) presented a deep image of the system of the two galaxies and detected a faint tidal tail of stars, north-northwest, pointing away from the centre (see also Kenney et al. 1995, although note that this feature is also clearly visible in the image presented by Phillips & Malin 1982). Figure 4 presents a normal and a high contrast image of the two galaxies; the tidal tail detected by Malin (1993) can be easily seen extended up to 100 ( 50 kpc) to the NNW of NGC s4435 in the very high contrast image, although there is no particular further evidence for the distortion of the halo of NGC 4435 (both galaxies however share a common halo in projection). A newlydiscovered lament of similar length can be seen extending to the SW of NGC 4438 (it is located within the curve marked in Fig. 4b). This new lament is an extension of the SW tidal tail seen in the normal contrast image. There are hints of other possible radial laments immediately to the east of this. NGC 4438 NGC 4435 Fig. 4. High a) and very high b) contrast images of the interacting galaxies NGC 4438 and NGC 4435. The dimensions of the frame are 200 330 The expectations of Kenney et al. (1995) that NGC 4435 has the more disturbed halo are not con rmed by our deep images, although the most prominent lament (to the NW) is associated with it. These galaxies lie in a lm area a ected by variations in the background and many lm defects not removed by the median-stacking process. Nevertheless the two faint tidal tails to NW and SW are clearly visible, there are hints of others, and it is possible that more such features could be visible in a deeper image and a more detailed study of the morphology of the halo regions could be carried out. 3.5. The interacting group IC 3481, IC 3481A & IC 3483 Arp (1966) was the rst to mention the apparent interac tion of the IC 3481, IC 3481A & IC 3483 galaxies. This interaction is displayed in Fig. 5, which contains three images at di erent contrast, which display a number of remarkable newly-detected faint features. The normal contrast image, Fig. 5a, is similar to the one presented by Arp (1966), although much more extended; the interaction between IC 3481 & IC 3481A is visible, but the tail to- Fig. 6. Very high contrast image of the whole ( 6:2 6:2 ) wards IC 3483 is very faint. On the higher contrast image, eld covered by the Schmidt lms. A large scale, marginal, Fig. 5b, the tail of the interaction and its properties are variation in photographic density is apparent. This may be clearly seen: a loop-like structure is formed with one apex a real detection of the luminous, inter-galactic, medium, or apparently just touching IC 3481A, and the other apex simply an artefact as consequence of the variation in sensitivity apparently close to IC 3483. On the very high contrast over this large eld image, Fig. 5c, one can see the extent of the interaction around the IC 3481A galaxy and the common halo surrounding IC 3481 and IC 3481A. The details of how the loop-like structure connects (or not) to IC3483 is hidden reveals that the morphology of this background variation by the halo of the bright foreground star. R magnitudes of changes considerably from lm to lm, suggesting that, for 12.9, 15.9 can be measured from the CCD data obtained individual lms at least, that it is dominated by variations from the JKT, for the galaxies IC 3481 and IC 3481A in emulsion sensitivity across the lm. respectively. Only a lower limit of 16.5 mag for the newlydetected faint features can be reported as the frame of the We further investigated this large scale variation usCCD data was not extensive enough. ing three IIIaF plates of the north-east area of the Virgo Although visual inspection is strongly suggestive that cluster also taken with the UKST. This second eld overall three galaxies are interacting and connected by newly laps with our Virgo SE eld. The OG590 lter (described discovered faint lamentary material, there is a strong above) was used for the IIIaF plates to give results close to possibility that IC 3483 is not involved in this interac- the standard R-band (UKSTU handbook). We co-added tion, indicated by its H I radial velocity: 128 10 km s−1, the scans of three IIIaF Virgo NE plates using exactly the (Binggeli et al. 1993), while IC 3481 & IC 3481A have ra- same procedure, as discussed in Sect. 2. dial velocities of 7086 56 km s−1 (Binggeli et al. 1985) & 7304 56 km s−1 (de Vaucouleurs et al. 1991) respectively. The morphology of the possible background emission IC 3481 and IC3481A are therefore objects in the back- in Fig. 6 is such that the brightest regions are seen in the ground to the Virgo cluster, and the interaction is likely central parts of the cluster towards the north and northto involve only these two galaxies. western parts of the eld. As this is much what we would have anticipated, it appears that we could be detecting genuine optical, intra-cluster, medium emission, in which 3.6. Variation of background photographic density case the co-addition with median- ltering of the thirteen lm scans would not be dominated by the variations in A large-scale variation in background photographic den- sensitivity across individual lms. However, the reality of sity is apparent in a high contrast image of the co-added this most extensive emission must remain uncertain until array (see Fig. 6). Inspection of scans of individual lms further observations are made. 4. Conclusion to obtain a preliminary photometric calibration and a sky noise level of 27.9 0.3 mag arcsec−2 was measured for This paper illustrates preliminary results obtained from the co-added array for the area around these galaxies. the digital co-addition of 13 APM scans of Kodak TechPan The problems associated with the processing of large UKST lms of a eld containing the SE part of the Virgo (222 Mb) data les have been overcome by a combinacluster. With this technique we can detect previously un- tion of modifying existing Starlink software, and writing seen low surface brightness features of galaxies over the new software. These software items have been used to prowhole of the 6:2 6:2 area scanned, producing data vide new ways of implementing techniques such as alignwith an increased dynamic range and which can be cali- ment, normalisation, co-addition, correction for vignetting brated photometrically using CCD observations of objects e ects, and removal of stellar features. Positional informawithin the eld. Hence we can make full quantitative use of tion was also stored in the nal co-added array. this data array, and this important calibration will be re- Some new astrophysical results are presented here. ported in a future paper (some preliminary calibration re- Filaments are seen associated with the apparently intersults are given in this paper). Also with such digital data, acting galaxies NGC 4435 and NGC 4438. Connecting laprocessing techniques such as removal of stellar images can ments are seen between the galaxies IC 3481 and IC 3481A be performed, while methods such as \unsharp masking" for the rst time, and a further \loop" of laments apparare greatly facilitated. Mosaicing of CCD frames is not ently connecting these two galaxies with IC 3483 (though yet competitive in terms of covering such large elds in we note the discrepant redshift of the latter). Also more equivalent observing and processing times, and has prob- high-quality images are presented of previously detected lems with matching backgrounds at the edges of frames. features such as the N lament associated with NGC 4435 Many of the galaxies reported here would require mosaic- and NGC 4438, and the shells and \jet" features of M 89. ing of large numbers of contiguous CCD elds to cover Also deep images are presented of the giant haloes surtheir apparent area. rounding the ellipticals M 87, M 84 and M 86, although, While processing the lm scans we noted that there perhaps surprisingly, no trace of distortions are found in were variations in emulsion sensitivity over the wide eld the halo shapes of galaxies apparently adjacent to them, of the lms, varying in position from lm to lm. In the some of which might have been thought to be interacting co-added image it appears that the faintest structures we with them. Co-addition of further lms will clarify varidetect are reliable up to scales of about 1 square degree ous uncertain features (e.g. whether or not there are any on average. In Fig. 6 we see a strong suggestion of very more shells in the vicinity of M 89 or further laments faint intra-cluster medium emission occupying the north- associated with NGC 4435/8). ern part of the eld (central area of the cluster), but the Complete CCD calibration and multicolour photomereality of these structures needs to be tested with accu- try will be important for the scienti c interpretation of rate calibration of each lm, further lms and independent these features. We will also estimate the mass of lumimeasurements. This is one of our priorities for future work, nous material contained within the haloes of galaxies in as is investigating the causes of the apparent variation in the eld. Comparison of the revealed halo sizes with the emulsion sensitivity across each lm. extent of X-ray emission detected by ROSAT and other Nevertheless the TechPan lms clearly provide good X-ray satellites show that there is no consistent correlation imaging quality, even with the slight blurring introduced between the R-band and X-ray haloes. The X-ray haloes by digital aligning and co-addition (errors in alignment are can be much larger than the optical haloes (as in the case only a small fraction of a pixel), while new faint features of of M 87) or much smaller, nevertheless the combination of galaxies have been discovered, even though many of these visible (\red-emitting") mass in the haloes detected here galaxies have been studied before (some previous studies and the X-ray-emitting mass can provide new estimates have been at blue wavebands, where the relatively faint of the total luminous baryonic masses of galaxies. sky level makes low surface brightness structures easier to detect). When the data array is satisfactorily calibrated, we will present the accurate surface brightnesses for the features detected here and the gain in magnitude obtained by the co-addition. Acknowledgements. We are particularly grateful to the sta of both the UKSTU and the ROE for the provision of the photo With a single lm features should be detected at sur- graphic material, their help on various occasions and the useful face brightnesses as low as 27 R mag arcsec−2 (Phillipps & discussions. We also thank the Royal Greenwich Observatory Parker 1993). As the co-addition of 13 lms should reduce (RGO) and specially Dr. M. Irwin for the scanning of the lms. the sky photon noise per pixel by a factor q 2 13 ( 2:9), Many thanks also to Dr. B. Kellett for his hospitality and the interesting discussions, and Dr. R.J.R. Williams, Dr. Q.A. Parker, Dr. S. Phillipps and J.-M. Schwartzenberg for further interesting and useful discussions. then features as low as 28.1 R mag arcsec−2 should be visible in the co-added array. The CCD data from the sky area around the galaxies IC 3481 & IC 3481A were used

This is a preview of a remote PDF:

A. C. Katsiyannis, S. N. Kemp, D. S. Berry, J. Meaburn. Results of the digital co-addition of thirteen Schmidt films of the Virgo cluster of galaxies, Astronomy and Astrophysics Supplement Series, 387-399, DOI: 10.1051/aas:1998302