As galaxies orbit in the cluster environment, they are subject to tidal stripping from a variety of sources - interactions with individual galaxies, with groups of galaxies, or with the global cluster potential itself (see, e.g., the discussion in Gnedin 2003). Over time, this stripped starlight builds up the diffuse intracluster light found in clusters of galaxies. The properties of this ICL - its luminosity, morphological structure, metallicity, and kinematics - and their correlation with cluster properties can help unravel the dynamical history of cluster collapse, accretion, and evolution. To date, theoretical work has largely focused on tidal stripping from individual galaxies orbiting in an evolved cluster potential (e.g., Merritt 1983; Richstone & Malamuth 1983; Moore et al. 1996; Calcáneo-Roldán et al. 2000) and ignored two important effects: preprocessing in groups, and heating by substructure (Gnedin 2003). Full cosmologically-motivated simulations are needed to study the phenomenon in detail (e.g., Dubinski et al. 2001; Napolitano et al. 2003; Mihos et al. 2004).
An example of these models is shown in Figure 5 (from Mihos et al. 2004). In this simulation, we have excised a cluster from a flat CDM cosmological simulation and traced it back to z = 2. At that point we identify dark matter halos more massive than 1011 M which destined to end up in the z = 0 cluster and replace them with composite (collisionless) disk/halo galaxy models. The simulation is then run forward to the present day to examine the formation of tidal debris and the ICL. In essence, this simulation follows the contribution to the ICL from luminous galaxies, rather than from the stripping of low mass dwarfs. In this simulation, we see significant kinematic and spatial substructure at early times; at late times much of this substructure has been well mixed into a diffuse intracluster light. However, at low surface brightnesses, significant substructure remains even at z = 0.
Figure 5. Morphological (top) and kinematic (bottom) structure of the intracluster light in a simulated galaxy cluster. Left panels show the cluster at z = 1, while the right panels show z = 0. From Mihos et al. (2004).
Detecting this ICL has proved difficult, as at its brightest, the ICL is only ~ 1% of the brightness of the night sky. Efforts to detect this ICL include deep surface photometry to look for the diffuse ICL (e.g., Uson et al. 1991; Bernstein et al. 1995; Gonzalez et al. 2000; Feldmeier et al. 2002), as well as imaging of individual stars and planetary nebulae in nearby clusters (Ferguson et al. 1998; Feldmeier et al. 1998; Arnaboldi et al. 2002). Recently, these surveys have begun to reveal interesting substructure in the ICL, often in the form of diffuse arcs or streaks of material from tidally stripped galaxies (Trentham & Mobasher 1998; Gregg & West 1998; Calcáneo-Roldán et al. 2000).
To quantify the prevalence and properties of ICL as a function of cluster properties, we have begun a deep imaging survey of clusters using the KPNO 2m (Feldmeier et al. 2002, 2004). We target a variety of clusters, from cD-dominated Bautz-Morgan Type I clusters to Type III clusters which are typified by a more irregular distribution of galaxies. Examples from this survey are shown in Figure 6. The massive cD cluster Abell 1413 is marked by regular distribution of diffuse light, well-fit by a r1/4 distribution over a large range of radius, with only a moderate excess at large radius and little substructure. In contrast, Abell 1914 shows a variety of features: a fan-like plume projecting from the eastern clump of galaxies, another diffuse plume extending from the galaxy group to the north of the cluster, and a narrow stream extending to the northwest from the cluster center. We see similar behavior in other Abell clusters we have surveyed.
Figure 6. Left: the cD cluster Abell 1413, after subtraction of a smooth r1/4 law (the extent of which is shown by the ellipse). Very little substructure is seen. Right: the Bautz-Morgan Type III cluster Abell 1914, showing a rich variety of substructure. North is to the left; east is down. (From Feldmeier et al. 2002, 2004).
Although the sample size is small, these results are consistent with the expectations that substructure in the ICL is correlated with the dynamical state of the cluster as a whole. As clusters are assembled, the ICL is built up though the significant tidal stripping that occurs during interactions within the accreting groups, and between galaxies and substructure within the cluster. Does the total amount of ICL also correlate with Bautz-Morgan cluster type? Examining ICL measurements from a variety of sources, Ciardullo et al. (this conference) find only a weak dependence - the ICL fraction rises as expected from Type III to Type II clusters, but Type I (cD-dominated) clusters show fractionally less ICL than do the Type II's. However, the drop in the Type I's is likely due to the difficulty in distinguishing the ICL from the diffuse envelope of the cD galaxy itself; indeed, such distinction may not even be well motivated, since the cD envelope itself likely is formed from tidally stripped material. Including the luminosity of the cD envelope in the ICL budget would raise the fractional amount of ICL in Type I clusters and bring the trend in line with expectations from the dynamical models for generating ICL in clusters.
My numerous collaborators have all made many contributions to this work. In particular, I thank Cameron McBride for his work generating and visualizing the cluster ICL simulations. This work has been supported in part by the NSF through a CAREER award AST-9876143 and by a Research Corporation Cottrell Scholarship.