GLOBULAR CLUSTERS IN THE OUTER GALACTIC HALO: NEW HUBBLE SPACE TELESCOPE/ADVANCED CAMERA FOR SURVEYS IMAGING OF SIX GLOBULAR CLUSTERS AND THE GALACTIC GLOBULAR CLUSTER AGE–METALLICITY RELATION^*

The Galactic globular clusters (GCs) have long been recognized as useful probes of the formation and evolution of the Galaxy. Their spatial coherence allows their locations in age, metallicity, and Galactocentric distance (R_GC) space to be established with some accuracy. Their numbers allow them to be used as probes of the formation timescale and chemical enrichment history of the Milky Way halo, thick disk, and central bulge.

One of the earliest papers to examine the age–metallicity relation (AMR) of the Milky Way GC system was Sarajedini & King (1989). They calculated ages for 32 GCs using the magnitude difference between the horizontal branch (HB) and the main-sequence turnoff (MSTO). Taking a cue from Gratton (1985), Sarajedini & King (1989) compared the AMR for GCs inside 15 kpc of the Galactic center with that of GCs outside 15 kpc. They found a statistically significant difference between the two: the AMR of GCs with Galactocentric distance (R_GC) < 15 kpc was shallower than that of GCs with R_GC > 15 kpc. Sarajedini & King (1989) argued, based on the work of Larson (1972) and Tinsley & Larson (1978), that the inner halo GCs formed in a region of higher gas and dust density, thereby accelerating GC formation and chemical enrichment. In contrast, the outer halo GCs formed in lower density sub-galactic fragments in a slower, more chaotic fashion akin to the fragmentation and accretion scenario advocated by Searle & Zinn (1978, see also Navarro et al. 1997) and featured in models of galaxy formation based on the cold dark matter paradigm with a cosmological constant (e.g., Zentner & Bullock 2003; Robertson et al. 2005; Font et al. 2006).

Subsequent work by Chaboyer et al. (1992), Sarajedini et al. (1995), and Chaboyer et al. (1996) reinforced the results of Gratton (1985) and Sarajedini & King (1989) using larger and more reliable data sets of GC ages as well as a variety of techniques for measuring them. The most recent studies to determine ages from a large sample of GCs, by Marín-Franch et al. (2009) and Dotter et al. (2010), rely upon the uniform photometric data set provided by the Hubble Space Telescope (HST) Advanced Camera for Surveys (ACS) Galactic GC Treasury project (Sarajedini et al. 2007). This data set provides deep photometry for 65 GCs of up to ∼7 mag below the MSTO (Anderson et al. 2008). Based on a careful analysis of the ages and metallicities of the Galactic GC Treasury project GCs, Marín-Franch et al. (2009) showed the clearest evidence thus far that the GC AMR splits into two distinct branches. This result was reinforced by Dotter et al. (2010), who added the six most distant (known) Galactic GCs to the GC Treasury data set and used the combined sample to study the correlations between HB morphology and a variety of GC parameters. The analysis presented by Forbes & Bridges (2010) using data assembled from the literature for 93 GCs largely reinforces these conclusions.

The HST/ACS Galactic GC Treasury project preferentially targeted nearby GCs in order to maximize the depth that could be achieved with one orbit per filter. One consequence of this strategy is that it left out much of the outer halo. The most distant GC observed as part of the Treasury project is NGC 4147 at R_GC ∼ 21 kpc. Yet there are many GCs beyond this distance that can provide more leverage on the formation and subsequent evolution of the outer Galactic halo.

A number of investigators have used HST Wide Field Planetary Camera 2 (WFPC2) photometry to derive the ages of GCs with R_GC > 50 kpc. Harris et al. (1997) presented deep photometry of NGC 2419 (R_GC ∼ 90 kpc; Harris 1996, 2010 revision) and concluded that it is coeval with the similarly metal-poor GC M 92 based on a differential comparison of their color–magnitude diagrams (CMDs). Stetson et al. (1999) obtained deep photometry of Palomar 3, Palomar 4, and Eridanus (all with R_GC > 90 kpc) and concluded that the three are ∼1.5–2 Gyr younger than the relatively nearby GCs M 3 and M 5. Their conclusion rests on the assumption that the outer halo GCs have similar chemical abundances ([Fe/H] and [α/Fe]) to M 3 and M 5; this assumption has since been borne out for Palomar 3 (Koch et al. 2009) and Palomar 4 (Koch & Côté 2010). Dotter et al. (2008) presented CMDs of Palomar 14 (R_GC ∼ 75 kpc) and AM-1 (R_GC ∼ 125 kpc) and concluded that both are 1.5–2 Gyr younger than M 3, in keeping with the results of Stetson et al. (1999). For the purposes of the present study, the six known GCs with R_GC > 50 kpc are sufficiently well studied to firmly mark their place in the GC AMR. Chemical abundance information remains rather sparse for the most distant GCs, but what is known is consistent with what has been inferred from the CMDs.

Thus far then, deep and fairly homogeneous HST CMDs in the V and I (or equivalent) filters exist for the majority of GCs with R_GC ≲ 15 kpc and all those (known) beyond 50 kpc, allowing precise and internally consistent ages to be measured. What remains is to target the GCs within 15 ≲ R_GC ≲ 50 kpc so that the complete radial extent of the Galactic GCs, and what it can tell us about the formation of the Galaxy, can be probed. The present study is the first major step toward fulfilling this goal, wherein we present deep, homogeneous HST ACS photometry and use the resulting CMDs to estimate the ages, of six outer halo GCs: Pyxis, Ruprecht 106, IC 4499, NGC 6426, NGC 7006, and Palomar 15. Of the six, only IC 4499 has previous HST photometry (Piotto et al. 2002) and it is not of sufficient depth to allow an accurate age determination. In addition, a new age determination for Palomar 5 (R_GC ∼ 18 kpc) using archival HST/WFPC2 V, I data from Grillmair & Smith (2001) has been performed and included in the subsequent discussion.

The remainder of this paper proceeds as follows: the observations and data reduction are described in Section 2; the CMDs are presented in Section 3; the available information on the chemical abundances, distances, and reddening values of the sample is reviewed in Section 4; new age determinations based on isochrone fitting are presented in Section 5; the Galactic GC AMR is updated and compared with theoretical work in the context of the formation of the Galaxy in Section 6; Section 7 provides some discussion of the results; and Section 8 summarizes the findings presented in the paper.

The observations were obtained with the HST/ACS Wide Field Channel (WFC) under program number GO-11586 (PI: Dotter) during Cycle 17. Table 1 shows the observing log. The ACS images were retrieved from the HST archive and calibrated using the pipeline bias and flat-field procedures. All clusters, except for Palomar 15, were imaged over two orbits with the first orbit devoted to F606W and the second to F814W. Palomar 15 is the most distant and reddened in the sample and it received a total of five orbits: two in F606W and three in F814W.

Table 1. GO-11586 Observations

Cluster	Data Set	Date	R.A.	Decl.	PA_V3	F606W	F814W
Pyxis	jb1601	2009 Oct 11	$\rm 09^h07^m57\mbox{$.\!\!^{\mathrm s}$}7$	−37:13:17.0	76.86	50 s, 4 × 517 s	55 s, 4 × 557 s
Ruprecht 106	jb1602	2010 Jul 4	$\rm 12^h38^m40\mbox{$.\!\!^{\mathrm s}$}1$	−51:09:00.9	287.34	55 s, 4 × 550 s	60 s, 4 × 585 s
IC 4499	jb1603	2010 Jul 1	$\rm 15^h00^m18\mbox{$.\!\!^{\mathrm s}$}5$	−82:12:49.5	235.97	60 s, 4 × 603 s	65 s, 4 × 636 s
NGC 6426	jb1604	2009 Aug 4	$\rm 17^h44^m54\mbox{$.\!\!^{\mathrm s}$}6$	+03:10:13.0	294.67	45 s, 4 × 500 s	50 s, 4 × 540 s
NGC 7006	jb1605	2009 Oct 5	$\rm 21^h01^m29\mbox{$.\!\!^{\mathrm s}$}4$	+16:11:14.4	275.03	45 s, 4 × 505 s	50 s, 4 × 545 s
Palomar 15	jb1606	2009 Oct 16	$\rm 16^h59^m50\mbox{$.\!\!^{\mathrm s}$}9$	−00:32:17.9	261.25	10 s, 4 × 550 s	10 s, 4 × 500 s
⋅⋅⋅						65 s, 4 × 550 s	25 s, 4 × 560 s
⋅⋅⋅							55 s, 4 × 525 s

Download table as:  ASCIITypeset image

Photometry and astrometry were extracted from the images using the programs developed for GO-10775, the ACS Survey of Galactic GCs, as described by Anderson et al. (2008). The observing strategy in the current program was designed to follow the identical pattern of exposure times and dithers as GO-10775. Since Anderson et al. (2008) give an extremely detailed description of the data reduction process, only those aspects that have changed in the interim will be mentioned here.

Before performing photometry on the individual images, they were corrected for charge transfer efficiency using the algorithm developed by Anderson & Bedin (2010). The instrumental photometry was then calibrated to the HST VEGAmag system following Bedin et al. (2005) and using the aperture corrections from Sirianni et al. (2005) as described by Sarajedini et al. (2007) and Anderson et al. (2008). The VEGAmag photometric zero points were obtained from Bohlin (2007). The photometric catalogs and supporting data files from the reduction process of the six GCs presented herein will be made publicly available through the same archive that will host the ACS GC Treasury database.

The F606W, F606W–F814W CMDs are presented in Figures 1–6. Typical photometric errors are demonstrated toward the left side of each figure. In cases where ground-based V, I data are available, we also make a direct comparison with the ACS data converted to V and I using the empirical transformations provided by Sirianni et al. (2005). The ground-based comparisons indicate, in all cases, that the HST photometric system maintains a high degree of homogeneity post-SM4 and that the transformations to ground-based V and I magnitudes perform well.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The CMDs reveal several striking features. In particular, it can be seen that Ruprecht 106 and IC 4499 have strong binary and blue straggler sequences. NGC 6426, Palomar 15, and Pyxis show signs of differential reddening and, indeed, these are the most reddened clusters in the present sample. Palomar 15 and Pyxis are sparsely populated clusters; the red giant branch (RGB) of Pyxis loses coherence for F606W ≲ 19.

This section provides a brief review of noteworthy studies targeting the GCs considered in the present study. A few basic parameters are collected in Table 2.

4.1. NGC 6426

4.2. IC 4499

4.3. Ruprecht 106

4.4. Palomar 15

4.5. NGC 7006

4.6. Pyxis

4.7. Palomar 5

Dotter et al. (2010) presented isochrone fits to the photometric catalog of the ACS Survey of Galactic GCs, excluding only those known at the time to possess multiple, distinct stellar populations. The fits were performed with isochrones from Dotter et al. (2008). The procedure employed in the isochrone fits presented in this section follows that of Dotter et al. (2010, Section 4.2). The distance, reddening, and chemical composition of the isochrones are initially set to the best available literature values (see Table 2) and these are adjusted only when necessary to achieve the best simultaneous agreement between the models and the CMD on the unevolved main sequence and the base of the RGB. Dotter et al. (2010, see their Figures 6 and 7) showed that the distances derived in this manner were consistent with the absolute magnitudes of HB stars derived from RR Lyraes with the 1σ uncertainties and that the derived metallicities were typically within 0.1–0.2 dex of spectroscopically determined values, particularly in the range −2 ⩽ [Fe/H] ⩽ −1.5. The values derived from isochrone fitting for the present sample are compared to the values obtained from the literature at the end of this section.

Once the isochrones were registered in this manner, the age is estimated by selecting the isochrone which most closely matches the shape and position of the subgiant branch. The isochrone fits are performed "by eye" and the quoted uncertainties reflect the width of the data in the CMD and any mismatch between the data and the slope of the models through this region. As such, these uncertainties are suggestive of goodness of fit but lack statistical significance. Chaboyer & Krauss (2002) report that uncertainties in stellar models lead to an error of ±3% in derived ages if a star's properties are known exactly. At an age of 13 Gyr, 3% corresponds to ∼0.4 Gyr. Adding this value in quadrature with a typical uncertainty of 1 Gyr due to the fit (see Table 3) and ∼0.5 Gyr for 0.2 dex uncertainty in metallicity (e.g., Dotter et al. 2008) leads to a realistic error budget of ∼1.2 Gyr.

As can be seen in the figures that follow, age effects are only relevant from ∼2 mag below to ∼1 mag above the MSTO: greater depth in the CMD provides a tighter constraint on the models. The CMDs of Ruprecht 106, IC 4499, and NGC 7006 provide excellent constraints; differential reddening is a limiting factor in NGC 6426, Palomar 15, and Pyxis. The sparseness of the Pyxis and Palomar 15 CMDs also reduces the precision of the isochrone fitting.

Stars brighter than the subgiant branch are saturated in the WFPC2 observations of Palomar 5. To enable a more robust age determination, the WFPC2 data have been supplemented with ground-based observations from Stetson (2000). For the ∼20 stars that appear in both catalogs, the average offsets are less than 0.01 mag in V and I; the WFPC2 data were adjusted to match the photometric zero points of the standard stars measured by Stetson (2000).

The isochrone fits are shown in Figures 7–9; the results of isochrone fitting are presented in Table 3. For comparison purposes, the reddening and distance values listed in Table 3 have been transformed to standard DM_V and E(B − V) using the extinction coefficients given by Sirianni et al. (2005). The CMDs in Figures 7 and 8 were cleaned to improve the clarity of the fits by selecting stars according to photometric errors and other diagnostic information as described by Anderson et al. (2008, Section 7). As check on the results of isochrone fitting, synthetic zero age HB model sequences for the color ranges appropriate for each GC are plotted in Figures 7–9 as dashed lines. Generally speaking, these comparisons lend confidence to the results of the isochrone fitting. However, note that the models appear to inadequately represent those GCs that harbor both red and blue HB stars and, further, that the scarcity of HB stars in Pyxis and Palomar 5 make these comparisons of relatively little use.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Comparison of the [Fe/H] values listed in Tables 2 and 3 indicate that the literature values and those adopted in the isochrone fits differ, at most, by 0.1 dex in [Fe/H]. The largest departures are in cases where no information regarding [α/Fe] is available. Comparison of the derived distance moduli shows that four of seven GCs considered here differ by less than 0.05 mag. The largest discrepancy in distance modulus, that of Ruprecht 106, also corresponds to the largest discrepancy in [Fe/H] between the Harris (1996) catalog and the adopted spectroscopic value. The Harris catalog distances are based on the relation between the absolute magnitude of the RR Lyrae stars and the mean [Fe/H] of the GC.

As described in Section 2, all aspects of the observations and data reduction of the six GCs observed with ACS in program GO-11586 were designed to be homogeneous with the ACS Survey of Galactic GCs. The observations and data reductions of Palomar 5 are consistent with the outer halo GCs studied by Stetson et al. (1999) and Dotter et al. (2008). The isochrone analysis of all seven GCs presented herein is consistent with the approach used in Dotter et al. (2008, 2010). Hence, the results in Table 3 can be incorporated into the AMR assembled by Dotter et al. (2010).

The upper-left panel of Figure 10 shows the AMR of Dotter et al. (2010) with filled circles indicating R_GC ⩽ 8 kpc and diamonds indicating R_GC > 8 kpc. The current samples are plotted as open diamonds to highlight where they fall on the AMR. In order to give a sense of the uncertainties in the diagram, a "typical" error bar representing ±0.2 dex in [Fe/H] and ±1 Gyr in age is plotted in every panel of Figure 10. The remaining panels compare the GC AMR to various model predictions: the upper-right panel shows the simple closed-box chemical evolution model for the Sagittarius dwarf spheroidal galaxy from Siegel et al. (2007); the lower-left panel shows the chemical evolution models of the Magellanic Clouds from Pagel & Tautvaišienė (1998); and the lower-right panel shows the mean trend from the GC formation model of Muratov & Gnedin (2010). These comparisons are discussed in the following paragraphs.

Zoom In Zoom Out Reset image size

The closed-box model portrayed in the upper-right panel of Figure 10 was taken from the study of the star formation history of M 54 and the central field of the Sagittarius (hereafter Sgr) dwarf spheroidal galaxy by Siegel et al. (2007). The star formation history was derived with the same stellar models used in this paper and photometry from the ACS Survey of Galactic GCs. It has already been demonstrated by Layden & Sarajedini (2000) that the GCs associated with Sgr follow the same trend as its field stars. Forbes & Bridges (2010) reiterate that the AMR of Sgr is essentially identical to that of the putative Canis Major (CMa) dwarf galaxy and argue that ∼20 GCs originated with one or the other. The upper-right panel of Figure 10 indicates that the outer halo GCs follow an AMR consistent with dwarf galaxy chemical evolution. However, it appears that the outer halo GCs formed in a variety of environments within which chemical enrichment proceeded at different rates; one dwarf galaxy similar to Sgr or CMa is insufficient to explain the observed AMR.

The GC AMR is compared to the AMRs of the Small Magellanic Cloud and Large Magellanic Cloud (SMC and LMC) chemical evolution models by Pagel & Tautvaišienė (1998) in the lower-left panel of Figure 10. The LMC AMR is not markedly different from that of M 54/Sgr, as pointed out by Forbes & Bridges (2010). The SMC AMR tracks the majority of GCs with ages less than 12 Gyr better than either the LMC or M 54/Sgr.

The comparisons with Sgr and the MCs indicate that a collection of dwarf galaxies with a range of AMRs are likely to have contributed to the Galactic GC population, consistent with the findings of Forbes & Bridges (2010). Muratov & Gnedin (2010) devised a semianalytical model of the formation of a massive host galaxy's GC population and trained their model on the Galactic GC metallicity distribution. The model is built on top of cold dark matter galaxy formation simulations and assigns a GC formation probability and chemical enrichment history to each sub-halo that is accreted by the host galaxy. Among their stated goals is the development of a bimodal color distribution in the GC population without requiring two distinct formation mechanisms. While the model achieves this goal by design, it also fits the age–metallicity distribution of the outer halo GCs remarkably well. The lower-right panel of Figure 10 compares the GC AMR with the mean trend (squares) and standard deviation (indicated by error bars) of [Fe/H] for 0.5 Gyr age bins of nearly 10,000 simulated GCs drawn from the Muratov & Gnedin (2010) model. While the outer halo GCs are well represented by the mean trend, the model is discrepant with the metal-rich GCs in the inner Galaxy, i.e., the bulge and thick disk GCs. Muratov & Gnedin attribute this discrepancy to the lack of a metallicity gradient within the individual merging halos. As a substantial fraction of the Galactic GC population is old and metal-rich, it behooves a model of the GC population formation to describe the origins of these GCs.

The AMRs constructed by Marín-Franch et al. (2009) and Dotter et al. (2010, further updated in this paper) show that Galactic GCs split into two distinct branches at [Fe/H] ≳ −1.5. Marín-Franch et al. (2009) interpreted this branching as representing two groups of GCs: an old group with uniformly old ages (a flat AMR) and a young group with a significant trend toward younger ages at higher metallicities. They concluded that these groups likely originated from two different phases of Galaxy formation: an initial, relatively brief collapse and a second, prolonged episode driven by accretion. Marín-Franch et al. (2009) found a small number of metal-poor ([M/H] ⩽ − 1.5) GCs with somewhat younger ages than the bulk of the metal-poor GCs. Dotter et al. (2010) and the present study find all of the metal-poor GCs to be uniformly old,⁴ but this difference does not substantially influence the conclusion with respect to the proposed Galaxy formation scenario.

A number of influential studies have used the HB-morphology–metallicity diagram to infer the ages of stellar populations for which metallicity information and good photometry to at least the level of the HB are available. For example, Catelan & de Freitas Pacheco (1993) and Lee et al. (1994) both used synthetic HB models to demonstrate that, under the assumption of constant mass loss, the Galactic GC HB-morphology–metallicity diagram showed evidence for an age spread of several (∼4) Gyr for a fixed He content and heavy-element distribution.

The use of the HB-morphology–metallicity diagram must, however, remain a crude age indicator at best for a variety of reasons. The lack of detailed abundance information, including the possible spread in He with a given GC, has important implications for HB morphology. The effect of a spread in the total He content on HB morphology among different GCs has been explored by many authors (see, e.g., Catelan & de Freitas Pacheco 1993; Lee et al. 1994) as well as a spread in He within single GCs (see, e.g., D'Antona & Caloi 2008). Another important factor in matching GC HB morphology to model predictions is statistical fluctuations. The ACS GC Treasury project photometric database contains GCs with anywhere from ∼20 to several hundred HB stars, leading to uncertainty in the median color of the HB of up to 30% in the sparsest GC (Dotter et al. 2010, Table 1).

Addressing the question of which parameters most strongly influence HB morphology using full CMDs for large samples of GCs, Dotter et al. (2010) and Gratton et al. (2010) reached the same conclusion that age is the second parameter influencing HB morphology. The results presented in this paper lend further support to this conclusion. It is important to note that the term "second parameter" used here applies specifically to the variation of HB morphology among different GCs, the context in which the term was originally used in the 1960s (e.g., van den Bergh 1967; Sandage & Wildey 1967). "Second parameter" is not intended to apply to the spread in HB stars within a single GC, for which the same term is often employed, nor does this paper seek to address that important issue. Dotter et al. (2010) and Gratton et al. (2010) both found evidence for at least a third parameter as well, suggesting that age and metallicity alone are not sufficient to fully characterize HB morphology.

Figure 11 shows the HB-morphology–metallicity diagram of the full sample of 68 GCs. The left panel considers HB morphology as the median color difference between the HB and RGB (Dotter et al. 2010) and the right panel instead uses the (B − R)/(B + V + R) metric of Lee (1989) as compiled by Mackey & van den Bergh (2005). Dotter et al. (2010) demonstrated that these two HB-morphology metrics are strongly correlated, with the main difference being that Δ(V − I) does not saturate as (B − R)/(B + V + R) does at ±1. In this context, it is clear that age is the second parameter influencing HB morphology as concluded by Dotter et al. (2010) and Gratton et al. (2010).

Zoom In Zoom Out Reset image size

While Figure 11 is not well sampled over the full range of metallicity, it can still be seen that the transition from a red HB to a blue one happens over ≲ 0.5 dex in [Fe/H] for a fixed age, or ≲ 2 Gyr at fixed [Fe/H]. It is interesting to consider the ensemble in this manner, despite the fact that age uncertainties remain at or near the 10% level. It is a fortunate coincidence that we observe these GCs at a time when the HB-morphology–metallicity diagram is rich with information. Turn the clock back ∼2 Gyr and the HB-morphology–metallicity diagram would have been populated only on the red side—as, for example, in the SMC (Glatt et al. 2008). At such a time, the study by Searle & Zinn (1978) would have revealed little variation in HB morphology among the halo GCs and, hence, little or no insight into the formation of the Galaxy. Only the truly peculiar GCs, such as NGC 2808 (Dalessandro et al. 2010), might have shown a sizable variation. On the other hand, turn the clock forward ∼2 Gyr and all but the most metal-rich Galactic GCs would have blue HBs.

HST/ACS photometry of six Galactic halo GCs, IC 4499, NGC 6426, NGC 7006, Palomar 15, Pyxis, and Ruprecht 106, were presented. The resulting CMDs were used to derive ages via isochrone fitting. The age of Palomar 5 was derived using the same set of models and the combined photometry of Grillmair & Smith (2001) and Stetson (2000). The ages and metallicities of these seven GCs were added to the AMR of Dotter et al. (2010). The total number of homogeneously studied GCs in the sample is 68, excluding those seven GCs imaged by the ACS Survey of Galactic GCs that are known to harbor multiple, distinct stellar populations. Divided at R_GC = 8 kpc, the inner Galaxy GCs exhibit a pattern of rapid chemical enrichment spanning two orders of magnitude in metallicity over a timescale of 1–2 Gyr. The outer Galaxy GCs exhibit a much slower chemical enrichment that is evocative of dwarf galaxy chemical evolution and is reasonably well matched by the semianalytic GC formation models of Muratov & Gnedin (2010).

Data presented in this paper were obtained from the Multimission Archive at the Space Telescope Science Institute (MAST). Support for this work (proposal GO-11586) was provided by NASA through a grant from the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555. This research has made use of NASA's Astrophysics Data System Bibliographic Service as well as the SIMBAD database, operated at CDS, Strasbourg, France.

Facility: HST (ACS) - Hubble Space Telescope satellite