arXiv:astro-ph/0208266v1 14 Aug 2002 2/2 r w bl lsesat clusters Abell two are 222/223 A Introduction 1. by cieterdcino h pcrsoi n photometric and spectroscopic the of reduction the scribe lines. well of three as number than known, a more for previously 223, widths redshifts equivalent Abell of as and number 222 the Abell times of field the in (hereafter clusters. clusters be- the connection tween the possible between a indicating region kinematical region”), clus- “intercluster the the first at PEL) in galaxies a 4 redshift hereafter did found also ter and (2000, PEL system. spectra al. this 53 of et study of Proust list a before published Newberry 1988) clus- 1976; al. being al. et them et (Sandage of known most were spectra region, members, 9 cluster ter clusters. al. the massive in et be galaxies David to of confirmed 1997; are Ulmer and 1958). & 1999) (Wang been (Abell have ROSAT 3 they by clusters, class observed selected richness optically are Abell clus- these having Both While sample. rich photometric are (1983) ters al. et Butcher the [email protected] edopitrqet to requests offprint Send pcrsoyo h egbrn asv lsesAel222 Abell clusters massive neighboring the of Spectroscopy ⋆ DI ilb netdb adlater) hand by inserted be will (DOI: Astrophysics & Astronomy h ae sognzda olw.I et ede- we 2 Sect. In follows. as organized is paper The galaxies 183 for redshifts independent 184 report We ae nosrain aea S aSla(Chile) Silla La – ESO at made observations on Based ∼ ohcutr ntecutrrsfaeaeaottesame: the about are restframe cluster the in clusters both edtrietelmnst fbt lsesaddrv mass derive and clusters both of subst cluste luminosity for red evidence the the From find determine detected. we we be While can respectively. substructure 223, kinematic A and 222 A flnsfr13glxe,13o hmblnigt h 2 an 222 A the to belonging them of be 153 to galaxies, redshifts 183 for lines of e words. Key (202 aaisat galaxies 2 1 Abstract. 09.08.2002 Accepted / 29.05.2002 Received 2 aais itne n esit aais lumino Galaxies: – redshifts and distances Galaxies: – 223 A aao otistepstos redshifts, positions, the contains catalog 14 bevtieMd–y´nes M57,1 Av. Midi–Pyr´en´ees,Observatoire 14 UMR5572, Germany Forschu Extraterrestrische und f¨ur Astrophysik Institut ′ ± ntesy or sky, the on 43) epeetasetocpcctlgo h egbrn massi neighboring the of catalog spectroscopic a present We h z 70 ∼ aais lses eea aais lses individ clusters: Galaxies: – general clusters: Galaxies: M 0 z ⊙ . 242. 0 = /L ⊙ . ∼ 2126 n ( and .P itih e-mail: Dietrich, P. J. : aucitn.h3727 no. manuscript 2600 ± M/L .P Dietrich P. J. 0 h . 70 − 08frA22and 222 A for 0008 1 ) z A223 p,blnigto belonging kpc, ≈ 0 (149 = R . 1separated 21 magnitudes, bl 223 Abell 1 ± .I Clowe I. D. , 33) dur ei,340Tuos,France Toulouse, 31400 Belin, Edouard ´ h z 70 0 = V M − g nvri¨tBn,AfdmHue 1 32,Bonn, 53121, H¨ugel 71, dem Auf Universit¨at Bonn, ng, t60n n iprino 16n/m ihthe With nm/mm. 11.6 580 of of 24 dispersion resolution of a a pixels CCD and has 2048x2048 used which nm instrument were 2, 600 The grism nights at cirrus. with These con- EMMI high 1999. three was occasional on December with NTT the clear in 222 at nights performed Abell was secutive clusters 223 two Abell the and of spectroscopy Multi-object Reduction Data and Data 2. cosmology. h airto rmswr ae ttebgnigo the of beginning the calibration. at the taken were in frames used calibration lines lamps, wave- The 20 Helium–Argon The typically using available. only provided night done was which second was seconds the calibration 2700 in length 222 of A exposure on cluster. of one each centered on exposures field 3 two fields, the 6 For exception for taken were one each seconds With 2700 nm/pixel. 0.28 of r umrzdi et .Truhu hspprw as- we paper results this Our Throughout Ω 4. an 5. sume Sect. Sect. in in by sequence summarized clusters are cluster the red of the ratio selecting mass–to-light determine and We an luminosity with substructure. the 3 possible Sect. finding in on examined of are emphasis the kinematics system the in cluster double and values the distribution previous spatial from The deviations literature. discuss and data ⊙ . 2079 σ R /L iyfnto,ms function mass function, sity euneietfidi color–magnitude–diagram a in identified sequence r 1 1014 = oo,a ela h qiaetwdh o number a for widths equivalent the as well as color, n .Soucail G. and , ⊙ t–ih aisi the in ratios –to–light ± diinlyw dniyagopo background of group a identify we Additionally . 2 ytm edtrieteheliocentric the determine We system. 223 A d utr ntesaildsrbto fA23 no 223, A of distribution spatial the in ructure ⋆ 0 . 08frA23 h eoiydsesosof dispersions velocity The 223. A for 0008 Λ a:A22–Glxe:cutr:individual: clusters: Galaxies: – 222 A ual: − +90 71 0 = ecutr bl 2 n bl 2.The 223. Abell and 222 Abell clusters ve ms km . 7 , Ω − m 1 2 and 0 = . σ R 3 H , µ bn f( of –band 1032 = hslast dispersion a to leads this m 0 70 = − +99 76 M/L h 70 ms km coe 5 2018 25, October ms km ) A222 − 1 − for 1 = and Mpc − 1 2 J. P. Dietrich et al.: Spectroscopy of the neighboring massive clusters Abell 222 and Abell 223 night for the masks used during that night, before the science exposures were made.
0.25 2.1. Reduction of spectroscopic data For the data reduction a semi-automated IRAF1 package was written by the authors that cuts out the single spectra of the CCD frames and then processes these spectra using standard IRAF routines for single slit spectroscopy. The 0.2 sky spectrum was removed from all spectra using a linear 0.22 fit with a 2σ rejection on measurements on each side of the galaxy spectrum where the position of the galaxy on 0.215 the slit permitted it. Measurements from only one side of the spectrum were used otherwise. A 2σ rejection was 0.21 used to remove cosmic rays and hot pixels. Remaining hot 0.15 pixels or cosmic rays in the sky spectrum introduced fake 0.205 absorption features, while hot pixels or cosmic rays in the 0.2 spectrum itself lead to fake emission features. These were 0.2 0.205 0.21 0.215 0.22 removed by hand. 0.15 0.2 0.25 Because the sky spectrum removal was done column z (this work) by column and no distortion correction was applied, some Fig. 1. Comparison of redshift measurements for objects residual sky lines remained in the final spectra, most no- observed by us and with redshifts listed in PEL. The large tably of the strong [O i] emission at 5577 A.˚ At the typical panel shows the full sample, the inset is a blow up of the redshift of the cluster members of z ≈ 0.21 this line does cluster region. The error bars are the internal errors re- not coincide with any important feature and thus does not ported by RVSAO. cause any problems in the subsequent analysis.
2.2. Redshift determination their and our cluster members, indicating that PEL ob- The radial velocity determination was carried out using served a sub–sample with a significantly different mean the cross-correlation method (Tonry & Davis 1979) im- value from the larger sample we describe here. plemented in the RVSAO package (Kurtz & Mink 1998). We ruled out the possibility that this discrepancy could Spectra of late type stars and elliptical galaxies with have been caused by taking all calibration frames before known radial velocities were used as templates. The red- the science exposures were taken. This could have intro- shift determination was verified by visual inspection of duced a shift of the zero point if the masks were not moved identified absorption and emission features. The average back to their original position for the science exposures. internal error reported by RVSAO is cz = 68 km s−1. We confirmed that this is not the case by determining Adding this in quadrature to the estimated error of the the radial velocity of the subtracted sky spectrum in the wavelength calibration of cz = 110 km s−1, we estimate wavelength calibrated frames. We found that, if a zero −1 the error in the redshift determination of single galaxies point shift occurred, it must be smaller than 30 km s , to be δz =0.0004. confirming the accuracy of our data. Fig. 1 shows a comparison of our redshift measure- ments and the redshifts listed by PEL. The obvious out- 2.3. Equivalent widths lier in the large panel is from the sample of Sandage et al. (1976). The inset shows a broad agreement between our We measured equivalent widths for the [O ii]λ3727, results and the values of PEL. Ignoring the obvious out- [O iii]λ5007 emission lines, and Hβ and Hα emission and lier, the average difference between the measurements is absorption lines. The integration ranges for the features −3 hz − zPELi = (0.2±1.1)×10 . Student’s t-test rejects the and the continuum were fixed by the values given in Table null hypothesis of different sample means with same vari- 1. ance at higher than the 99% level for the 23 cluster galax- [O ii] and Hα are important indicators of star forma- ies we have in common with PEL. However, Student’s tion rates (Kennicutt 1998). To accurately determine the t-test confirms the hypothesis of different sample means equivalent widths and in particular estimate their signif- with same variance at higher than the 99% level for all icance level we follow the definition of equivalent widths given by Czoske et al. (2001): 1 IRAF is distributed by the National Optical Astronomy
Observatories, which are operated by the Association of Nint Universities for Research in Astronomy, Inc., under cooper- fi Wλ = ∆λ − Nint∆λ, (1) ative agreement with the National Science Foundation. i=1 fc X J. P. Dietrich et al.: Spectroscopy of the neighboring massive clusters Abell 222 and Abell 223 3
Table 1. Restframe wavelength ranges for equivalent widths measurement. All wavelengths are given in A.˚ The last column gives the continuum range that was used for estimating the signal–to–noise ratio.
Feature λcent line blue cont. red cont. SNR [O ii] 3727 3713 - 3741 3653 - 3713 3741 - 3801 3560 - 3680 [O iii] 5007 4997 - 5017 4872 - 4932 5050 - 5120 4450 - 4750 Hβ 4861 4830 - 4890 4800 - 4830 4890 - 4920 4050 - 4250 Hα 6563 6556 - 6570 6400 - 6470 - 6300 - 6450
where fi is the flux in pixel i, Nint is the number of pixels in to allow for both the pointing offsets in the dithering pat- the integration range, fc is the continuum level estimated tern and any changes in the distortion pattern between as the mean of the continuum regions on either side of images. By comparing the positions of stars among the the line, and ∆λ is the dispersion in A/pixel.˚ Note that individual images and to the positions in the USNO cata- with this definition emission lines have positive equivalent log, both systems of equations for coordinate transforma- widths. tions were solved using χ2 minimization of the final stellar The significance of an equivalent width measurement positions. The rms dispersion of the centroids of the stars ′′ is given by (Czoske et al. 2001) used in the fitting were 0. 016 among the input images and 0′′. 54 between the input images and the USNO co- 2 σWλ = ordinates, with the average offset vector being consistent −2 2 with zero in all regions of the image. Further details of this S (W + Nint∆λ) (W + N ∆λ) ∆λ + λ , (2) technique along with justifications for the linear transla- N λ int N " c # tion between CCD and detector plane can be found in Clowe & Schneider (2001). The mapping of each input where the signal to noise ratio S/N = fc/σc, and fc is CCD was performed using a triangular method with lin- obtained by averaging over Nc pixel in the wavelength ear interpolation which preserves surface brightness even range given in Table 1. if the mapping changes the area of a pixel. The resulting All wavelengths are given in the restframe of the ob- images were then averaged using a 3σ clipping algorithm ject. All spectra were normalized to a continuum fit be- to remove cosmic rays and moving objects. The final R- fore equivalent widths were measured. The catalog lists band image can be found in Fig. 6. all [O ii] and [O iii] emission features and all Hβ and Hα emission and absorption features that were detected with a significance > 2σ. Objects were detected in the R-band image us- ing SExtractor (Bertin & Arnouts 1996), and the V - band magnitudes for the objects were measured using 2.4. Photometry SExtractor in two-image mode. The FWHM of bright but unsaturated stars in the coadded images are 0′′. 87 for R Wide-field imaging of the cluster pair was performed over and 1′′. 05 in V . Zeropoints were measured from Landolt two nights in December 1999 with the Wide Field Imager standard fields (Landolt 1992), but the V -band data is on the ESO/MPG 2.2m on La Silla. Eleven 900 second known to have been taken in non-photometric conditions. exposures in R-band and three 900 second exposures in From isolating the red cluster galaxy sequence in a color- V band were taken using a dithering pattern which filled magnitude plot (Fig. 2), corrected for the A =0.086 mag the gaps between the CCDs in the mosaic in the coad- B dust extinction (Schlegel et al. 1998) using the conversion ded image. The image reduction was carried out using a factors from Cardelli et al. (1989), and comparing to pre- combination of self–written routines and routines which dicted colors of cluster elliptical galaxies in a passive evo- are part of the IMCAT software package written by Nick lution model (Fukugita et al. 1995), a correction of −0.23 Kaiser (http://www.ifa.hawaii.edu/∼kaiser/imcat). mag has been applied to the V magnitudes to correct for The images were flattened with medianed night-sky flat- the additional atmospheric extinction. This correction also fields from all the R or V band long exposure images taken causes the stellar V −R colors to have the theoretically ex- over the two nights. The images were aligned using a pro- pected values (Gunn & Stryker 1983). All magnitudes are cess which assumes each CCD in the mosaic can be trans- isophotal magnitudes with the limiting isophote at 27.96 lated to a common detector–plane coordinate system using mag/arcsec2. We determine the completeness limit of the a linear transformation of coordinates (a shift in both axes photometric catalog to be at R = 24 mag from the point and rotation allowed) and that the detector-plane coordi- where the number counts of objects departs from a power nates can then be transformed into sky coordinates using law. a two dimensional polynomial, in this case a bi-cubic poly- nomial. The linear transformation from each CCD to the detector-plane is assumed to be constant for all the im- An excerpt from the big catalog of spectroscopically ages whereas the transformation from detector-plane to measured galaxies listing the 10 brightest galaxies in R– sky coordinates is determined separately for each image band in both cluster regions can be found in Table 2. The 4 J. P. Dietrich et al.: Spectroscopy of the neighboring massive clusters Abell 222 and Abell 223 0 1 eicnrcrdhf n nenlerrrpre yRVS by reported error internal and redshift heliocentric 11. & 10. hu:iuescn)3 elnto dge:iuescn)4. (degree:minute:second) declination 3. (hour:minute:second) al 2. Table (1988), BL 223 ABELL BL 222 ABELL betN.R 20)Dc(2000) Dec (2000) RA No. Object s adg ta.(1976), al. et Sandage : xep rmtefl aao itn h 0bihetglxe in galaxies brightest 10 the listing catalog full the from Excerpt 5 01:37:57.74 154 9 01:38:15.49 198 4 01:37:57.22 149 9 01:38:07.25 193 7 01:38:02.30 173 1 01:37:50.32 119 4 01:37:56.02 140 7 01:38:02.19 172 6 01:38:01.02 167 8 01:38:04.83 186 ...... 901:37:28.32 39 301:37:21.21 13 501:37:42.08 95 601:37:22.65 16 901:37:34.2 69 701:37:43.03 97 501:37:34.01 65 001:37:26.34 30 301:37:41.54 93 01:37:17.97 8 − − − − − − − − − − − − − − − − − − − − 24:581.108 - - 0.83 17.41 12:47:55.8 24:841.808 - - 0.80 17.48 12:43:38.4 25:571.706 15 - 0.66 17.77 12:55:55.7 25:711.006 36 0.68 17.40 12:50:17.1 30:921.408 - - 0.81 17.74 13:00:39.2 24:301.608 - - 0.87 17.36 12:48:13.0 24:951.808 4 0.83 16.58 12:45:19.5 25:551.607 1 - 0.75 17.96 12:55:35.5 24:541.607 - - 0.75 17.36 12:46:15.4 24:981.208 - - 0.84 17.12 12:49:09.8 24:111.409 - - 0.91 17.34 12:45:41.1 24:161.007 - - 0.79 17.00 12:46:51.6 30:131.007 4 - - 0.76 17.60 13:00:21.3 24:401.808 - - 0.83 16.88 12:47:34.0 25:271.410 - - 1.01 17.44 12:56:52.7 25:371.108 - - 0.84 17.41 12:57:43.7 25:861.008 - - 0.89 17.00 12:59:28.6 25:691.708 - - 0.83 16.97 12:59:56.9 25:081.108 - - 0.89 16.91 12:58:30.8 30:071.308 - - 0.83 17.03 13:01:20.7 em ailvlct eie rmeiso ietemplate. line emission from derived velocity radial : V R − R [O . . 6 9 R ii ± ± ]/ admgiue5. magnitude band 1 O1.Rvleo or ai 17)1.notes, 13. (1979) Davis & Tonry of R-value 12. AO 3 [O A ˚ . . 3 9 15 4 R . . . . 2 9 8 7 bn nbt lse ein.Teclm nre r .Ojc No. Object 1. are entries column The regions. cluster both in –band iii ± ± ± ± ]/ 2 0 0 0 . . H A ˚ . . 0 9 14 - 8 9 5 − − − − − − − − − − − − − − − − − 7 5 3 5 4 3 3 6 7 4 4 6 6 7 7 4 7 . . V 3 9 ...... 7 2 5 9 4 4 5 7 3 6 7 8 7 7 0 0 5 β ± ± − ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± / H A ˚ 1 1 R 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 . . 42 2 .16 .01 6.13 0.00019 0.21161 - 4 ...... 14 .02 4.82 0.00025 0.21045 - 7 3 2 .02 .02 .50 4.55 0.00025 0.20427 - 6 .26 .02 5.97 0.00020 0.22067 - 3 .05 .02 4.44 0.00028 0.20957 - 5 .11 .01 5.87 0.00019 0.21612 - 3 .03 .01 6.66 0.00018 0.20132 - 4 .15 .02 4.34 0.00028 0.21050 - 9 7 .11 .02 5.14 0.00024 0.21118 - 6 .07 .02 5.07 0.00025 0.20972 - 1 .00 .02 4.05 0.00028 0.20809 - 9 .19 .02 4.86 0.00026 0.21392 - 6 .12 .02 6.11 0.00021 0.21329 - 6 .14 .02 5.34 0.00023 0.21347 - 4 .11 .02 4.14 0.00024 0.21318 - 6 .13 .01 11.30 0.00011 0.21132 - 7 oo –.osre qiaetwdh f[O of widths equivalent observed 6–9. color − 0 . . . 9 . 0 6 8 α ± ± ± ± / A ˚ 0 2 0 0 . . . .16 .02 4.69 0.00023 0.21560 9 . .00 .01 .70 6.67 0.00019 0.20707 4 .54 .02 .50 4.35 0.00023 0.05147 7 .47 .02 5.68 0.00023 0.24074 3 .03 .00 p 0.00004 0.20839 .07 .04 n 0.00040 0.20971 .15 .03 p 0.00030 0.21253 .00 .06 s p p 0.00167 0.00031 0.20600 0.00013 0.20506 0.20525 .17 .01 n 0.00010 0.21371 .18 .01 p 0.00012 0.21387 .15 .00 p 0.00007 0.21356 .10 .02 p 0.00022 0.21409 z hel p rute l (2000), al. et Proust : σ z hel Notes R [O [O [N . . . ii 20694 05126 20459 ii iii iii ,[O ], ] ,H ], ,H ], n .rgtascension right 2. ebrye al. et Newberry : ± ± ± iii β β H , 0 0 0 ,H ], . . . 03 esrdo [O on measured 00038 01 esrdo [O on measured 00013 00 esrdon[O measured 00009 α β n H and , α iii ii ii ], ], ], J. P. Dietrich et al.: Spectroscopy of the neighboring massive clusters Abell 222 and Abell 223 5
10
5
Fig. 2. Color–magnitude plot with the already corrected V magnitudes. The red cluster sequence is centered around V − R =0.8. 0 0 0.1 0.2 0.3 0.4 0.5 z full catalog is available in electronic form at the at the Centre de Donn´ees astronomiques de Strasbourg (CDS)2. Fig. 3. Radial velocity distribution of all galaxies in the sample. The main peak corresponds to the two Abell clus- ters. 3. Spatial Distribution and Kinematics After removing some obvious background and foreground by galaxies (z > 0.3 or z < 0.1), an iterative 3σ clipping was 2 2 (n − 1)σ 2 (n − 1)σ used to decide upon cluster membership. We found 81 2 < σ < 2 , (4) χ − χ − − galaxies belonging to Abell 222 and 72 galaxies belong- n 1;α/2 n 1;1 α/2 ing to Abell 223 or the possible bridge connecting both for n galaxies in the sample. The redshift and velocity clusters. dispersion for A 223 do not change significantly if the 3 The mean redshift of the individual clusters are z = galaxies from the intercluster region in our sample are re- 0.2126 ± 0.0008 and z = 0.2079 ± 0.0008, for A 222 and moved. Figs. 3 to 5 show the corresponding radial velocity A 223, respectively. This differs significantly from the val- distributions of the individual samples. ues of z = 0.2143 and z = 0.2108 found by PEL for The values for the velocity dispersion are somewhat A 222 and A 223, respectively. The quoted errors in- higher than those derived from X–ray luminosities. David clude the statistical error of δz = 0.0005 added linearly et al. (1999) report bolometric luminosities of LX =7.65× 44 −1 44 −1 to the estimated error of the cluster member selection of 10 erg s and LX = 6.94 × 10 erg s from ROSAT δz = 0.0003. This systematic error was estimated from PSPC observations for A 222 and A 223, respectively, for −1 −1 varying the cut level of the recursive clipping procedure H0 = 50 km s Mpc . Using the LX–σ relationships −1 from 2σ to 3.5σ and calculating the means of these cuts. of Wu et al. (1999) we get σX = 845 − 887 km s for −1 The statistical errors were calculated from a bootstrap re- A 222 and σX = 828 − 871 km s for A 223. Compared sampling of the cluster members. The measured velocity to the velocity dispersions derived from the spectroscopic dispersions have to be transformed to the restframe of the observations, both cluster appear to be underluminous in cluster according to the transformation law X–rays. Together with the data of PEL we now have radial σ σcor = , (3) velocities for 6 galaxies in the possible bridge connecting 1+ z both clusters. With our new values for the radial velocity (Harrison 1974). The restframe velocity dispersions for the of A 223 the observation made by PEL, that most of the +90 −1 bridge galaxies are in the low–velocity tail, does not hold individual clusters are σcor = 1014 −71 km s and σcor = +99 −1 anymore. In fact they all appear to be close to the maxi- 1032 −76 km s , for A 222 and A 223, respectively. This is in good agreement with the values found by PEL of mum or higher of the velocity histogram shown in Fig. 5. σ = 1013 ± 150 km s−1 and σ = 1058 ± 160 km s−1 for Because A 222 is the cluster at higher redshift, this is the A 222 and A223, respectively. The errors of the velocity expected behavior should these galaxies indeed belong to dispersion are the (1 − α) = 68% confidence interval given a bridge connecting both clusters SExtrator also provides an algorithm to sepa- 2 http://cdsweb.u-strasbg.fr/ rate galaxies from stars. Each object is assigned a 6 J. P. Dietrich et al.: Spectroscopy of the neighboring massive clusters Abell 222 and Abell 223
Fig. 6. Deep R band image of A 222 and A 223. The galaxy density contours are generated from a color selected sample of 702 objects. See text for details. The small circles mark all spectroscopically identified cluster galaxies.
CLASS STAR value, which is 1 for stars, 0 for galax- (α=01:38:01.8, δ=−12:45:07.0). These peaks are also visi- ies, and lies in between for ambiguous objects. Fig. 6 ble in the density distribution of the 181 spectroscopically displays a projected galaxy number density map gener- identified cluster galaxies. Also visible is an overdensity of ated with the adaptive kernel density estimate method color selected objects in the intercluster region, hinting at described by Pisani (1996) for a color selected sample of a connection between both clusters. 702 objects with R < 21 and 0.7 < V − R < 0.9 and SExtractor CLASS STAR < 0.1 from the WFI images. We applied the Dressler & Shectman (1988, DS) test Overplotted are the positions of all spectroscopically iden- for local kinematic deviations in the projected galaxy dis- tified cluster members. Fig. 6 clearly exhibits two den- tribution. The DS test is based on computing deviations sity peaks in A 223. Both peaks are separated by 4.′8 of local mean velocity and velocity dispersion from the and are centered at (α=01:37:53.5, δ=−12:49:21.2) and global values. The local values are calculated for each of the N galaxies and its n−1 nearest neighbors. The statis- J. P. Dietrich et al.: Spectroscopy of the neighboring massive clusters Abell 222 and Abell 223 7
Fig. 4. Radial velocity distribution of the members of A 222. The continuous line is a Gaussian with mean and measured dispersion value as given in the text. Fig. 7. Declination (top) and right ascension (bottom) wedge redshift diagram. Cluster members are plotted as solid circles while background and foreground galaxies are displayed with open symbols. The rectangle highlights a small background group of galaxies. The opening angle is from −13◦2′24′′ to −12◦16′24′′ for the declination wedge and from 01:37:04.8 to 01:38:48.0 for the right ascension wedge.
1 N 2 n = (v − v )2 + (σ − σ )2 (5) σ2 glob loc,i glob loc,i i=1 ( glob ) X h i The ∆ statistics is calibrated by randomly shuffling the ra- dial velocities while keeping the observed galaxy positions fixed. The significance of the observed ∆obs is assessed by the fraction of simulations whose ∆sim is smaller than ∆obs. We find that, while the DS test is clearly able to sep- arate both clusters at better than the 99.9% confidence level, it does not find any substructure in the individual clusters for values of 8 N The wedge velocity diagrams in Fig. 7 clearly show the Abell system at z ∼ 0.21. A small group of five galaxies ∆ = δi i=1 can be seen behind A 223 at z =0.242. We derive a veloc- X 8 J. P. Dietrich et al.: Spectroscopy of the neighboring massive clusters Abell 222 and Abell 223 ity dispersion of σ = 330km s−1, confirming that it is not only close in the projected spatial distribution but also in redshift space. 4. Mass–to–light ratio To determine the luminosity of the clusters we applied the same cut on color and CLASS STAR as above in a circle −1 with 1.4 h70 Mpc radius around the bright cD galaxy of A 222 and the center of the line connecting both density peaks in A 223. We binned the selected objects in bins of 0.5 mag and fitted a Schechter (1976) luminosity function, ∗ ∗ ∗ ∗ n(L)dL = n (L/L )α exp(−L/L )d(L/L ) , (6) where L∗ is the characteristic luminosity, α is the faint– end slope, and n∗ is a normalization constant. Written in terms of absolute magnitude Eq. (6) becomes ∗ ∗ N(M)dM = kn exp {[−k(α + 1)(M − M )] ∗ − exp [−k(M − M )]} dM, (7) Fig. 8. Differential R–band luminosity function for A 222. where M ∗ is the absolute magnitude corresponding to L∗ The points represent the objects selected by the criteria and k = ln 10/2.5 (Kashikawa et al. 1995). Using k cor- detailed in the text, while the dashed line is the best–fit rection and a passive evolution correction on the synthetic Schechter function. elliptical galaxy spectra of Bruzual & Charlot (1993), for dl the Abell system MR = mR −5 log Mpc −24.91, dl being the luminosity distance. The fit is performed by minimizing the quantity 2 2 [N(Mi) − Nf (Mi)] χ = 2 , (8) σi X with N(Mi) and Nf (Mi) being the observed and fitted number of galaxies in the ith magnitude bin. The variance of galaxies in each magnitude bin was assumed to be that of a Poissonian distribution. The best–fit Schechter function for A 222 has M ∗ = −22.1 ± 0.2, α = −1.04 ± 0.12, and n∗ = 31 ± 11. The χ2 value for these parameters is 12.7 with 7 degrees of freedom. The best–fit parameters for A 223 are M ∗ = −23.1 ± 0.2, α = −1.20 ± 0.06, and n∗ = 15 ± 5 with a minimum χ2 value of 7.0, also for 7 degrees of freedom. From Figs. 8 and 9 we see, that the Schechter function is a good representation of the faint end, while it slightly underpredicts the number of bright galaxies. Although the ∗ values of M differ by one magnitude they are compatible Fig. 9. Same as Fig. 8 for A 223. with values from the literature. E.g. Tr`evese et al. (1996) find M ∗ = −22.66±0.52 from a study of 36 Abell clusters ∗ ∗ for a fixed α = −1.25. For this value of α the M ∗ value for and L is the luminosity corresponding to the fitted M A 222 increases to −22.4, but the fit then has a reduced (Tustin et al. 2001). It follows from the chosen magnitude χ2 of 2.2. In the following we only use the lower value of cut and the size of the bins, that in our case the limit- ∗ M = −22.1 for A 222. ing magnitude MR = −19.0. This implies that we observe The total R–band luminosity of the red cluster se- 91.4% and 81.7% of the total light in A 222 and A 223, quence is determined by extrapolating the luminosity respectively. The total R–band luminosity of the red clus- 12 −2 function. The observed fraction of the total luminosity is ter sequence then is LR = (3.1 ± 0.3) × 10 h70 L⊙ for ∗ 12 −2 given by Γ(α+2,Llim/L )/Γ(α+2), where Γ(·, ·) is the in- A 222 and LR = (4.4 ± 0.4) × 10 h70 L⊙ for A 223, complete Gamma function, Llim is the completeness limit, where the errors reflect the uncertainty in the parameters which in this case is given by the selection parameters, of the luminosity function. J. P. Dietrich et al.: Spectroscopy of the neighboring massive clusters Abell 222 and Abell 223 9 We compute mass–to–light ratios assuming an isother- cluster we derived (M/L) ratios in R–band, which are mal sphere model for both clusters with the velocity dis- comparable for both clusters. The computed values are persion determined in Sect. 3. The mass of an isothermal (M/L)R = (202±43) h70 M⊙/L⊙ and (M/L)R = (149± sphere inside a radius r is given by 33) h70 M⊙/L⊙ for A 222 and A 223, respectively. This is within the range of values reported by other groups for πσ2 M(< r)= cor r . (9) other cluster. G Dressler (1978) gave a range of 140 − 420 h70 M⊙/L⊙ We get the following mass–to–light ratios inside in a study of 12 rich clusters. Typical values for −1 1.4 h70 Mpc: virial mass–to–light ratio are at values of M/L ∼ 210 h70 M⊙/L⊙ (Carlberg et al. 1996). Typical values A222: (M/L)R = (343 ± 60) h70 M⊙/L⊙ , derived from X–ray masses tend to be somewhat lower A223: (M/L)R = (253 ± 47) h70 M⊙/L⊙ . than those from virial masses. Hradecky et al. (2000) find a median value of (M/L) ∼ 140 h M /L in a study By selecting only the red cluster sequence we miss a sig- V 70 ⊙ ⊙ of eight nearby clusters and groups. nificant part of the cluster luminosity. Folkes et al. (1999) We cannot exclude the possibility that the values we found a ratio between the luminosity of their “type 1” report here are biased towards higher values by using an galaxies (E/S0) and the total luminosity of 0.59 ± 0.07 in isothermal sphere model. Both cluster geometries clearly the fields of the 2dF survey. Using this correction factor deviate from circular symmetric profiles. More robust we arrive at the final result: mass estimates may thus lead to lower (M/L) ratios. A222: (M/L)R = (202 ± 43) h70 M⊙/L⊙ , A detailled discussion whether the galaxies between both clusters indeed belong to a structure connecting the A223: (M/L)R = (149 ± 33) h70 M⊙/L⊙ . cluster pair will be part of a forthcoming weak lensing These values slightly overestimate the true mass–to–light study of this system. ratio because the use of isophotal magnitudes cuts off part Acknowledgements. We thank Joan–Marc Miralles and of the galaxy luminosity and the intracluster light. Lindsay King for help and useful discussions. This work was If we use the velocity dispersion derived from X–ray supported by the TMR Network ”Gravitational Lensing: measurement instead of the spectroscopically determined New Constraints on Cosmology and the Distribution of Dark velocity dispersion we arrive at mass–to–light ratios that Matter” of the EC under contract No. ERBFMRX-CT97-0172. are lower by ∼ 20% but still agree within the 1σ error with the values quoted above. References 5. Conclusions Abell, G. O. 1958, ApJS, 3, 211 Bertin, E. & Arnouts, S. 1996, A&AS, 117, 393 We have reported 184 independent redshifts measure- Bruzual, A. G. & Charlot, S. 1993, ApJ, 405, 538 ments for 183 galaxies in the field of Abell 222 and Butcher, H., Wells, D. C., & Oemler, A. 1983, ApJS, 52, Abell 223, as well as equivalent widths for [O ii], [O iii], 183 Hβ, and Hα, R magnitudes, and V − R color. Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, From a sample of 153 galaxies which we identified as 345, 245 cluster members, we derived a mean redshift and rest- Carlberg, R. G., Yee, H. K. C., Ellingson, E., et al. 1996, frame velocity dispersion of z = 0.2126 ± 0.0008, σcor = ApJ, 462, 32 +90 −1 1014 −71 km s for A 222 and z = 0.2078 ± 0.0008, Clowe, D. & Schneider, P. 2001, A&A, 379, 384 +99 −1 σcor = 1032 −76 km s for A 223. The values of the Czoske, O., Kneib, J.-P., Soucail, G., et al. 2001, A&A, redshifts are clearly outside the error margins of the val- 372, 391 ues previously reported by PEL. By comparing our wave- David, L. P., Forman, W., & Jones, C. 1999, ApJ, 519, length calibration to the sky spectrum, which provides an 533 independent wavelength standard, we were able to confirm Dressler, A. 1978, ApJ, 226, 55 the accuracy of our data and rule out the possibility of a Dressler, A. & Shectman, S. A. 1988, AJ, 95, 985 zero point shift of more than 30 km s−1 for each mask. R Folkes, S., Ronen, S., Price, I., et al. 1999, MNRAS, 308, and V band photometry was taken from WFI data. 459 Although the projected density maps of all spectro- Fukugita, M., Shimasaku, K., & Ichikawa, T. 1995, PASP, scopically identified galaxies and of a color selected sam- 107, 945 ple with 702 members clearly show spatial substructure in Gunn, J. E. & Stryker, L. L. 1983, ApJS, 52, 121 A 223, neither the DS test nor the DIP statistics were able Harrison, E. R. 1974, ApJ, 191, L51 to find any kinematic substructure. Also no indications of Hartigan, J. A. & Hartigan, P. M. 1985, Annals of a non–Gaussian parent population could be found. Statistics, 13, 70 We fitted a Schechter luminosity function to objects in Hartigan, P. M. 1985, Applied Statistics, 34, 320 the red cluster sequence identified in a color–magnitude Hradecky, V., Jones, C., Donnelly, R. H., et al. 2000, ApJ, diagram. Assuming an isothermal sphere model for the 543, 521 10 J. P. Dietrich et al.: Spectroscopy of the neighboring massive clusters Abell 222 and Abell 223 Kashikawa, N., Shimasaku, K., Yagi, M., et al. 1995, ApJ, 452, L99 Kennicutt, R. C. 1998, ARA&A, 36, 189 Kurtz, M. J. & Mink, D. J. 1998, PASP, 110, 934 Landolt, A. U. 1992, AJ, 104, 340 Newberry, M. V., Kirshner, R. P., & Boroson, T. A. 1988, ApJ, 335, 629 Pisani, A. 1996, MNRAS, 278, 697 Proust, D., Cuevas, H., Capelato, H. V., et al. 2000, A&A, 355, 443 Sandage, A., Kristian, J., & Westphal, J. A. 1976, ApJ, 205, 688 Schechter, P. 1976, ApJ, 203, 297 Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525 Tonry, J. & Davis, M. 1979, AJ, 84, 1511 Tr`evese, D., Cirimele, G., & Appodia, B. 1996, A&A, 315, 365 Tustin, A. W., Geller, M. J., Kenyon, S. J., & Diaferio, A. 2001, AJ, 122, 1289 Wang, Q. D. & Ulmer, M. P. 1997, MNRAS, 292, 920 Wu, X., Xue, Y., & Fang, L. 1999, ApJ, 524, 22