NGC 4254: An Act of Harassment Uncovered by the Arecibo Legacy Fast ALFA Survey
aa r X i v : . [ a s t r o - ph ] J u l Draft version October 25, 2018
Preprint typeset using L A TEX style emulateapj v. 08/22/09
NGC 4254: AN ACT OF HARASSMENT UNCOVERED BY THE ARECIBO LEGACY FAST ALFA SURVEY
Martha P. Haynes ,Riccardo Giovanelli , and Brian R. Kent Draft version October 25, 2018
ABSTRACTWe present an HI map constructed from the Arecibo Legacy Fast ALFA survey of the surroundingsof the strongly asymmetric Virgo cluster Sc galaxy NGC 4254. Noted previously for its lopsidedappearance, rich interstellar medium, and extradisk HI emission, NGC 4254 is believed to be enteringthe Virgo environment for the first time and at high speed. The ALFALFA map clearly shows a longHI tail extending ∼
250 kpc northward from the galaxy. Embedded as one condensation within thisHI structure is the object previously identified as a “dark galaxy”: Virgo HI21 (Davies et al. et al. (1996a,b; 1998). The smoothlyvarying radial velocity field along the tail as it emerges from NGC 4254 can be used as a timing tool,if interpreted as resulting from the coupling of the rotation of the disk and the collective gravitationalforces associated with the harassment mechanism.
Subject headings: galaxies: intergalactic medium — galaxies: halos — individual: Virgo cluster —radio lines: HI — galaxies: clusters — galaxies: interactions INTRODUCTION
Because of its proximity and relative richness, theVirgo cluster provides an especially useful laboratory forwitnessing in detail the impact of environment on galaxyevolution. In Virgo, studies of the HI deficiency firstquantified the impact of the cluster environment (Davies& Lewis 1973; Chamaraux, Balkowski & Gerard 1980),while subsequent mapping of the gas distribution provedthat the HI disks of the highly gas-poor objects were sys-tematically smaller than those of galaxies of normal HIcontent (Giovanelli & Haynes 1983; Cayatte et al. et al. et al. ◦ tothe NW of M87. Binggeli et al. (1985) assigned it mem-bership in the main Virgo A cluster, so that its projectedseparation from M87 is ∼ et al. − implies a velocity of ∼ − with Center for Radiophysics and Space Research, SpaceSciences Building, Cornell University, Ithaca, NY 14853. e–mail: [email protected], [email protected],[email protected] National Astronomy and Ionosphere Center, Cornell Univer-sity, Space Sciences Building, Ithaca, NY 14853. The NationalAstronomy and Ionosphere Center is operated by Cornell Uni-versity under a cooperative agreement with the National ScienceFoundation. respect to the cluster overall. It has no close neighborsbut is noted for its strong m = 1 asymmetric spiral pat-tern and vigorous star formation, especially in its brightsouthern arm. NGC 4254 has been the subject of nu-merous detailed studies, most of which focus on the gen-eration of its unusual optical lopsidedness by interactionwith the cluster environment.Of relevance to the present discussion, Davies etal. (2004) reported the detection, later confirmed via cor-roborating Arecibo observations (Minchin et al. × M ⊙ of HI at a position some 120 kpc northof NGC 4254. HI emission was detected in a total of fiveadjacent Arecibo pointings, spaced at intervals of 2.7 ′ ,and centered on ( α, δ , J2000) 12 h m s , +14 ◦ ′ ′′ .Based on the Arecibo observations, especially the rela-tively broad HI line width of 220 km s − , and the lack ofdetectable HI in a VLA synthesis map, they argued thatVirgo HI21 is the disk of a large ( >
16 kpc in diameter),optically–dark galaxy with a much larger total mass thanthat of the HI itself (Minchin et al. ∼
250 kpc to the northof NGC 4254. In section 2 we describe the ALFALFAdata and followup Arecibo observations. In section 3,we discuss the body of evidence regarding NGC 4254in terms of the galaxy harassment scenario of Moore et al. (1996a,b; 1998). All coordinates are for epochJ2000.0. OBSERVATIONS AND ANALYSIS
The region around NGC 4254 and Virgo HI21 has beenmapped as part of the ALFALFA survey program. AL-FALFA is a two-pass survey which uses the the AreciboL-band Feed Array (ALFA) on the 305 m antenna.The ALFALFA observing strategy and data processingpipeline were outlined in Giovanelli et al. (2005). Onceboth passes of the survey are completed, 3-D spectralline data cubes are constructed and a signal extractionalgorithm, discussed in Saintonge (2007), is applied forpreliminary identification of HI sources. HI sources inthe northern part of Virgo were included in the first AL-FALFA catalog release (Giovanelli et al. ◦ < δ < +16 ◦ . That tabulation includes 17HI detections within 1 ◦ of Virgo HI21 ( α = 12 h m , δ = +14 ◦ ′ ) and at velocities cz < − . Of the17 HI sources, four have no optical counterparts; threeof those are extended (in comparison with the 3 . ′ × . ′ Arecibo telescope beam) and appear associated with theNGC 4254–Virgo HI21 system, while the remaining oneis probably a foreground Milky Way high velocity cloud.The left panel of Figure 1 shows the integrated HIflux density contours as detected in the ALFALFA dataset overlayed on the 2nd generation Digital Sky SurveyBlue image in the vicinity north of NGC 4254. The HIemission is integrated over the range cz = 1946 to 2259km s − for the stream and cz = 2259 to 2621 for thegalaxy on a grid with 1 ′ spacing. The velocity cut for thestream is outside the 2250-2510 km s − range reported inVLA maps of NGC 4254 (Vollmer et al. α = 12 h m s , δ = +14 ◦ ′ ′′ in-dicates the centroid of Virgo HI21 as reported by Minchin et al. (2005a). HI features connect in a stream that ap-pears to emanate from NGC 4254.In order to confirm the extended nature of the streamand, with higher signal–to–noise, explore the velocityfield along it, a set of HI spectra were obtained alongthe ridge of the HI stream with longer dwell times thanthose of the ALFALFA survey. Thirty-six on/off pairswere taken at 1.5 ′ steps in declination, centered on thepeak emission indicated by the ALFALFA data, using thesingle-beam L-band wide (LBW) receiver of the Arecibotelescope. Difference spectra were then constructed fromeach pair, with 2048 channels per polarization coveringa 25 MHz band centered at 1410 MHz with spectral res-olution of 12.2 kHz (2.6 km s − at cz = 2000 km s − ).The integration times varied from three to six minuteson source, depending on the expected signal strength,yielding an rms noise per channel of better than ∼ et al. (2007) give an HI mass for NGC 4254itself of 4 . × M ⊙ . The total HI line flux associatedwith gas in the extended stream, including the clumpsreported by Giovanelli et al. (2007) and lesser ones near the ALFALFA detection limit is 6.5 Jy-km s − whichcorresponds to M HI ∼ . × M ⊙ . Given the extendednature of the source, the ALFALFA observations missdetecting any diffuse emission at column densities belowthe ALFALFA limit. Assuming that diffuse emission ispresent at column densities equal to or less than 3 × HI atoms cm − over a solid angle of 200 (arcmin) , up toan additional 1 . × M ⊙ of HI may be present in thestream, bringing the total HI mass to 5 . ± . × M ⊙ .The HI tail can be traced over ∼ ′ , which at the Virgodistance corresponds to 242 kpc. This length is certainlydramatic, but not unique. Of comparable length is thelongest of the stellar streams in the extended halo ofM87 which was detected in very deep images of Virgo byMihos et al. (2005). DISCUSSION
By many measures, NGC 4254 is an exceptional Scgalaxy, one of the brightest spirals in the Virgo A clus-ter. Its rare dominant m=1 spiral mode is evident in itsconspicuous southwestern spiral arm and the lesser butstill prominent arm with several muted branches seen tothe north and east. There is no primary distance mea-surement but Solanes et al. (2002) give a mean distancemodulus of 31.04 ± − ,with a dispersion of 758 km s − . With a heliocentric ve-locity of 2404 km s − (Giovanelli et al. m = 1 spiral mode, particularly in the absence ofan appreciable bar or nearby companion. Various studiesof the gaseous components have confirmed the stronglyasymmetric optical structure. The H α (Phookun, Vogel& Mundy 1993; Chemin et al. CO(J=1-0)maps are consistent with the picture of an externally–forced spiral density wave. Explanations for its asym-metry include the superposition of spiral modes inducedby global gravitational instability (Iye et al. et al. et al. α extent relative to its older R-band stellar population (Koopmann, Haynes & Catinella2006), a relatively normal molecular content (Nakan-ishi et al. et al. ∼ × M ⊙ with the mass in “high velocityclouds” ∼
3% of that number. As seen in their VLAHI map (Figure 5 of Phookun, Vogel & Mundy 1993),the extra gas forms a loop on the NE side of the galaxyat high velocity, complemented by several lower velocityclouds which form an extended tail on the SW side. Thelatter connect to the much longer HI stream seen in theALFALFA data. This connection is also clearly evidentin the top panel of Figure 1 of Vollmer, Huchtmeier &van Driel (2005) who reprocessed the original VLA map.Details of the distortion of the HI layer in a number ofgalaxies in the Virgo cluster are seen in an accumulatingnumber of HI maps of Virgo galaxies. The most dramaticevidence of gas compression and stripping is seen in theHI deficient spirals found close to the cluster center, yetextra-disk gas is found in a variety of cases. Ooster-loo & van Gorkom (2005) propose that an HI cloud of3.4 × M ⊙ and extending some 100 by 25 kpc hasbeen stripped from the strongly deficient Virgo core spi-ral NGC 4388 by ram pressure due to the surroundingintracluster medium. Further out from M87, Chung etal. (2007) report several one–sided HI tails pointing awayfrom M87 and suggest they result from ram pressure ex-erted by the hot intracluster medium.NGC 4254 is moving at high velocity with respect tothe cluster but is located 1 Mpc from M87 where the in-tracluster medium density is quite low. While Sofue etal. (2003) conclude that the distribution and kinematicsof the molecular gas in the inner regions of the galaxy areconsistent with the effects of ram pressure as NGC 4254falls into the Virgo cluster, Cayatte et al. (1994) suggestthat the gravitational restoring force within NGC 4254probably exceeds the ram pressure force at its currentlocation on the outskirts of the X-ray emitting region.That ram pressure alone cannot explain the HI and opti-cal asymmetry is further emphasized by Vollmer, Hucht-meier & van Driel (2005).While ram pressure is often invoked to explain lopsid-edness, gas compression and HI deficiency, gravitationalrather than hydrodynamical forces may produce similareffects. Of particular relevance are the simulations ofgalaxy harassment reported by Moore et al. (1996a) andMoore, Lake & Katz (1998). Those authors exploredthe collective impact of both high speed encounters andthe tidal effects of a smooth cluster potential in a va-riety of cases, one of which was a spiral galaxy withcircular velocity ∼
160 km s − , nearly identical to thatof NGC 4254. The most outstanding characteristics ofthe HI tail reported here are consistent with the harass-ment scenario: (a) its clear association with an excep-tional galaxy NGC 4254, located 1 Mpc from the centerof Virgo and traveling at high speed; (b) its overall length(242 kpc); (c) its modest gas mass (6 × M ⊙ ), about10% of the total HI associated with NGC 4254; and (d)its roughly sinusoidal velocity field (right panel of Figure1). Unlike the strongly deficient case of NGC 4388 (Oost-erloo & van Gorkom 2005), NGC 4254 is not HI poor, sothat we can conclude it has not passed through the Virgocore in recent epochs. Its distance of 1 Mpc from M87 is not coincidental: the galaxy density there is already quitehigh, so that a close high speed encounter is not improb-able. A curious aspect of NGC 4254, cited in support ofthe “dark galaxy” hypothesis for Virgo HI21 (Minchin etal. et al. (1996a)further note that such harassment can result in disrup-tion of the inner disk (Moore et al. ∼
220 km s − )of the section of the NGC 4254 HI tail identified asVirgo HI21 by Minchin et al. (2005a,b) also led those au-thors to argue that the feature must be a massive rotating“dark” galaxy. As Bekki, Koribalski & Kilborn (2005)have shown, gravitational forces are more likely to pro-duce broad velocity widths and large velocity gradientsthan is ram pressure stripping. The full view of the tailprovided by ALFALFA shows the width of Virgo HI21to be not necessarily due to galaxy–scale rotation, butrather a part of a larger scale velocity field.In the harassment interpretation, the velocity fieldalong the stream suggests a timing argument. We hy-pothesize that the roughly full cycle, sinusoidal velocitysignature apparent in the right panel of Figure 1 reflectsthe coupling of the spin and tidal motions. Because gasis stripped more readily from that portion of the galaxywhose rotation aligns with the direction of the tidal force,we assume that the HI present in the stream was removedfrom the edge of the outer disk of NGC 4254. The HIdiameter derived from the VLA maps is ∼ . ′ (Cayatte et al. ∼
150 km s − (Guhathakurta et al. ∼ . × years, which is com-parable with the cluster crossing time. NGC 4254 is nota particularly massive system: the mass enclosed withinan 18.5 kpc radius is of order 10 M ⊙ ; the Virgo clustermass within a 6 ◦ radius is estimated to be 4 . ± . × M ⊙ (Hoffman, Olson & Salpeter 1980). Thus tidal forcesdue to high speed encounters and/or the collective clus-ter potential could easily strip the outer gas layers ofNGC 4254 in a prograde orbit that passes within 1 Mpcof the cluster center.Based on the discovery of relatively long stellar streamsvisible in very deep optical images of Virgo (Mihos etal. etal. (2000) and Rudick, Mihos & McBride (2006) haveshown that such structures may arise when a relativelyluminous spiral galaxy suffers tidal shocking on a pro-grade encounter deep within the cluster potential. TheNGC 4254 event appears to be of a milder nature thanthose illustrated in those works, presumably becauseNGC 4254 has not plunged very deep into the clustercore. A stellar stream may be present but would beharder to discern in this case, given the significantlysmaller optical light radius of the galaxy and the conse-quent rarefaction of the stellar density outside 18.5 kpc.We propose that the most likely interpretation of thegalaxy’s asymmetry, its very extended HI tail and theorigin of Virgo HI21 is an on-going process of galaxyharassment, resulting from the high speed gravitationalperturbations experienced by NGC 4254 as it enters theVirgo cluster. Note added in proof:
We have recently learned thatBournaud & Duc (2007, in preparation) have modeledthe formation of a tidal tail in NGC 4254, with the char-acteristics described in this work, resulting from a very high speed encounter with another cluster galaxy now farremoved from NGC 4254.RG and MPH acknowledge the partial support ofNAIC as Visiting Scientists during the period of thiswork. This work has been supported by NSF grantsAST–0307661, AST–0435697, AST–0347929, AST–0407011, AST–0302049; and by a Brinson Foundationgrant.
REFERENCESBekki, K., Koribalski, B.S. & Kilborn, V.A. 2005, MNRAS, 363,L21Binggeli, B., Tammann, G.A. & Sandage, A. 1987, AJ, 94, 251Bournard, F., Combes, F., Jog, C.J. & Puerari, I. 2005, A&A,438, 507Calc´aneo-Rold´an, C., Moore, B., Bland-Hawthorn, J., Malin, D.& Sadler, E.M. 2000, MNRAS, 314, 324Cayatte, V., van Gorkom, J.H., Balkowski, C., & Kotanyi, C.1990, AJ, 100, 604Cayatte, V., van Gorkom, J.H., Balkowski, C., & Kotanyi, C.1994, AJ, 107, 1003Chamaraux, P., Balkowski & Gerard, E. A&A, 83, 38Chemin, L., Balkowski, C., Cayatte, V., Carignan, C., Amram,P., Garrido, O., Hernandez, O., Marcelin, M., Adami, C.,Boselli, A. & Boulesteix, J. 2006, MNRAS, 366, 812Chung, A., van Gorkom, J.H., Kenney, J.D.P., & Vollmer, B.2007, ApJ, 659, 115Davies, R.D. & Lewis, B.M. 1973, MNRAS, 165, 231Davies, J., Minchin, R., Sabatini, S., van Driel, W., Baes, M. etal. et al. et al. et al. etal. etal. et al. etal.
Fig. 1.—
Left: HI flux contours extracted from the ALFALFA survey dataset, which mapped the full field represented in the image,superposed on a DSS2 Blue image. The 36 filled dots indicate the locations of beam centers for the successive LBW observations. Thecontours centered on NGC 4254 are at 10, 15, 20, 30, and 40 Jy beam − km s − , integrated from 2259 to 2621 km s − . The contoursfor the HI stream are at 0.35, 0.52, 0.70, 0.87, and 1.0 Jy beam − km s − , integrated from 1946 to 2259 km s − . Note the difference indynamic range of contours for the galaxy and the stream, selected for viewing ease. The 3 ′ circle in mid stream indicates the position ofVirgo HI21 reported by Minchin et al.et al.