3. OBSERVATIONS

3.1. Optical SBF

Since the first successful prescription and quantification of SBFs as a distance indicator by Tonry & Schneider (1988), a multiplicity of works have used this technique to determine distances to globular clusters and (groups of) galaxies in the optical wavelength regime out to 40 Mpc. After this theoretical interpretation, the method was endorsed observationally with the application to Virgo galaxies in the VRI bandpasses (TAL90). Tonry (1991) established the first empirical calibration of the absolute fluctuation magnitude _I with a zero-point estimate relying solely on the M31 and M32 galaxies. A new indicator for measuring extragalactic distances was born.

In the following years the SBF method was applied to globular clusters (Ajhar & Tonry 1994) and tested in the blue wavelength regime (Shopbell et al. 1993; Simard & Pritchet 1994; Sodemann & Thomsen 1996). While the former approach gave good agreement with independent Cepheid measurements, the latter application yielded a rather poor characterisation (_B − _V ≈ 2.45 ± 0.15), mainly due to the problematic different sensitivities of _B and _V to age and metallicity as well as additional observational limits (e.g., much fainter fluctuation magnitudes, lower S / N, stronger residual signal from isophotal twist and morphological distortions, dust or disk/ring-like features, etc).

Several other studies utilised the method for individual galaxies and concentrated in refining the SBF method in the red optical bandpasses (Lorenz et al. 1993; Sodemann & Thomsen 1995; Fritz 2000, 2002).

John Tonry and collaborators have conducted a large I-band SBF survey to measure the distance to ∼ 300 nearby galaxies (Tonry et al. 1997, 2000, 2001 [hereafter T01]). The main result of this programme was an accurate empirical calibration of the SBF magnitude using optical colours as (Tonry et al. 2000):

(9)

This calibration is based on 10 Cepheid and 44 SBF distances in 7 different galaxy groups and defined for galaxies in a colour range of 0.95 ≤ (V − I)₀ ≤ 1.30. Both the zero-point and the slope of the relation are well defined (see Section 4.1 for a detailed discussion of uncertainties in the calibration of the SBF method). From ground-based I-band SBF measurements in spiral bulges, the zero-point in the _I relation is tied directly to the Cepheid distance scale to an accuracy of 0.08 mag (excluding systematic uncertainties in the Large Magellanic Cloud (LMC) distance modulus with the P − L relationship of ± 0.03 mag (Freedman & Madore 2010) or other methods, ± 0.16 mag (Mould et al. 2000); see Section 4.1), corresponding to ∼ 4% in distance (Tonry et al. 2000). A comprehensive review of the I-band SBF distance survey can be found in Tonry et al. (1997, 2000) and Blakeslee et al. (1999).

Jensen et al. (2003) used the Hubble Space telescope (HST) Near Infrared Camera and Multi-Object Spectrometer (NICMOS) ² F160W distances to 65 galaxies (see further Section 3.3) and established a slightly steeper slope for the _I relation in the H-band than in the I-band. These measurements were tied to the Cepheid distances of the HST Key Project (KP) by Freedman et al. (2001), neglecting the −0.2 mag dex⁻¹ metallicity correction. The need for such a correction was ambiguous at that time and argued that there appeared to be a better agreement with the SBF zero-point predictions from simple stellar population (SSP) models when the metallicity correction was omitted. More recent works provide additional evidence of a Cepheid metallicity dependence (Sakai et al. 2004; Macri et al. 2006; Scowcroft et al. 2009) and suggest that metallicity effects cannot be neglected when deriving SBF distances. Further, there is a better understanding of the stellar population predictions for NIR SBFs (Raimondo 2009; González-Lópezlira et al. 2010; Lee et al. 2010), which is discussed in detail in Section 5.

With respect to the NIR calibration and the collection of high-quality HST Advanced Camera for Surveys (ACS) ³ data (see further Section 3.3), in recent years there has been some confusion as to whether the T01 distances derived from of the ground-based I-band SBF distance survey (Tonry et al. 1997) require some correction. To overcome this discrepancy, Blakeslee et al. (2010) suggest to apply a correction to distances determined using the relation in equation 9, including both a zero-point (resulting from a revised Cepheid distance calibration of the KP (Freedman et al. 2001), which is −0.06 mag for the T01 distance moduli) and a second-order bias correction (originating from a distance bias in low-quality data of T01). Depending on the quality of the T01 data, the correction in the final T01 distance moduli (m − M)_T01 ranges between +0.05 mag (poorest quality) to ∼ −0.12 mag (highest quality). In order to be consistent with the revised Cepheid scale, the corrected T01 calibration, and other works using a metallicity correction, distance moduli obtained with the Jensen et al. (2003) IR SBF calibration would need to be increased systematically by +0.10 mag.

Ground-based optical SBF observations are restricted to relatively low redshifts. Therefore, these measurements are affected by the local peculiar velocities and difficulties are experienced in deriving direct numbers for the Hubble constant H₀. In combination with other distance estimators (e.g., D_n − σ, Tully-Fisher relation, see Section 3.3.1 for details), or space-based measurements, the SBF method provides much tighter constraints on the Hubble flow, as more distant galaxies can be considered. Using Fundamental Plane distances to ∼ 170 elliptical galaxies, Blakeslee et al. (2002) derived H₀ = 72 ± 4 (random) ± 11 (systematic) km s⁻¹ Mpc⁻¹.

3.2. Near-Infrared SBF

There are several obvious advantages for the application of the SBF technique at longer wavelengths. The fluctuation signal is significantly stronger, ∼ 35 times brighter in the K than in the I-band, making the SBF signal detectable to greater distances. The seeing is typically much better in the IR than in the optical and this greatly improves the SBF amplitude, which is inversely proportional to the seeing FWHM. As the fluctuations are red compared to the globular cluster population, the resulting contrast between stellar and GC contributions to the SBF signal is higher and can be easier separated. A drawback is that the sky background is larger in K. However, because of the red fluctuations, higher S / N ratios can be reached in shorter exposure times than in the optical. Further, the extinction correction is a factor of 4 lower in K′ than in I and for old, metal-rich populations the stellar population models predict a weaker dependence of the K-band SBF on the integrated (V − I) colour than in the I-band, reducing uncertainties in the colour determination of the stellar system (Worthey 1993). For this reason, the effects of age and metallicity are largely not degenerate. This fact acts in favour for stellar population studies of galaxies, but is less beneficial for simple distance calibrations.

The first feasibility studies of the IR SBF method were performed by Luppino & Tonry (1993); Pahre & Mould (1994) and Jensen et al. (1996) on a small set of galaxies in the Local Group and the Virgo Cluster. These works reported similar _K calibrations of _K ∼ −5.65 ± 0.20 mag. Jensen et al. (1998) extended the IR calibration to nine galaxies in the Fornax and Eridanus clusters using a wider range of (V − I) colours, metallicities and additional Cepheid distance measurements. Using high S / N ground-based NIR observations, Jensen et al. (1999) pushed their ground-based K-band observations to measure distances to elliptical galaxies in the Hydra (cz = 4054 km s⁻¹) and the Coma (cz = 7186 km s⁻¹) clusters. The authors suggested that _K itself can be used as a standard candle and therefore give direct constraints on the Hubble constant H₀ (see Section 4 for details). Based on the Coma SBF distance, this work derives H₀ = 85 ± 11 km s⁻¹ Mpc⁻¹.

3.3. SBF from HST

A number of works have mined the HST archive to use galaxy observations of other programmes for SBF analyses. After the HST refurbishment in December 1993 (STS-61, HST-SM1), SBF observations have been collected from several allocated programmes through Cyles 5 to 7 and 15 using the Wide Field Planetary Camera 2 (WFPC2) ⁴, ACS and NICMOS as well as a WFPC2 Guaranteed Time Observation (GTO) programme (Pahre et al. 1999). Further, new Wide Field Camera 3 (WFC3/IR) ⁵ observations are executed during Cyle 17.

3.3.1. Optical

Recent work on SBF distance determinations has concentrated on early-type galaxies observed with the HST. Space-based observations offer a number of advantages over ground-based measurements: Higher spatial resolution (FWHM_WFPC2 = 0.18″, FWHM_ACS = 0.098″), easier identification of GCs and dust due to the higher spatial resolution (reaching ∼ 2 mag fainter systems), higher PSF stability and photometric calibration, lower sky background, no disturbing atmospheric absorption hence no extinction correction, and insensitivity to focus changes of the camera (as in the Fourier analysis only wave numbers outside the range are affected, see Section 2.2 for details).

Using the WFPC2 onboard the HST, Ajhar et al. (1997) targeted 16 galaxies in the F555W (V) and F814W (I) filters. They found a steeper calibration of _I,F814W = (−1.73 ± 0.07) + (6.5 ± 0.7)[(V − I)₀ − 1.15] compared to Tonry et al. (1997). However, the usage of the WFPC2 camera and the harder calibration of the broader F814W was impractical for measuring distances greater than the Coma cluster (104.7 Mpc) because the observing time increases with the square of the distance.

Based on WFPC2 observations from a dedicated IDT/GTO programme, Pahre et al. (1999) have investigated the distances to NGC 3379, NGC 4373, and NGC 4406. One important goal of this work was to demonstrate the capability of the WFPC2 camera (despite its sampling rate) to measure extragalactic distances using the SBF technique. The refurbishment was a success and the results for the three galaxies were in good agreement with previous measurements. For the bright group galaxy NGC 4373 in the Hydra-Centaurus supercluster, the authors derive a distance of d = 39.6 ± 2.2 Mpc and a peculiar velocity for the galaxy of 415 ± 300 km s⁻¹. This peculiar velocity is about half as large as the 838 km s⁻¹peculiar velocity prediction using the Great Attractor model from Lynden-Bell et al. (1988) based on D_n − σ observations.

Ajhar et al. (2001) measured the SBF signal to E+S0s that have Type Ia Supernovae (SNe Ia) and found no systematic error with distance. However, they detected a systematic offset of ∼ 0.25 mag because of different Cepheid calibrations.

The higher quantum efficiency (a factor of 2 to 10, depending on the wavelength, see Table 1) and larger field of the Advanced Camera for Surveys Wide Field Channel (ACS/WFC) ⁶ compared to the WFPC2 chip (202″ × 202″ vs. 134″ × 134″) and better sampling of the PSF (factor of ∼ 2) allows researchers to extend the SBF distance scale to distances greater than 10,000 km s⁻¹. In an effort to study an unbiased, almost complete sample of nearby galaxies, Tonry and collaborators have obtained WFPC2 and ACS observations to ∼ 300 galaxies (Tonry et al. 2000, T01). The I-band SBF survey provided accurate calibrations and measurements to 91 elliptical and lenticular (E+S0) galaxies, and 9 spiral bulges up to 2000 km s⁻¹, thereby building a link to other distance indicators that are sensitive to and extend well out into the Hubble flow, e.g., D_n − σ (Dressler et al. 1987; Lynden-Bell et al. 1988), Tully-Fisher relation (Aaronson et al. 1989; Giovanelli et al. 1997), or SNe Ia (Sandage & Tammann 1982; Riess et al. 2009).

More recently the SBF technique was applied to derive distances to dwarf galaxies in groups and clusters in the southern hemisphere using large ground-based telescope facilities (e.g., Jerjen et al. 2001, 2004; Mieske et al. 2006, 2007; Dunn & Jerjen 2006; Mei et al. 2007; Blakeslee et al. 2009; 2010) or using HST ACS/WFC observations to nearby early-type galaxies (Cantiello et al. 2005, 2007a, b; Barber DeGraaff et al. 2007; Biscardi et al. 2008).

The first optical SBF distance out to ∼ 100 Mpc using ACS/WFC F814W bandpass observations to shell galaxies was measured by Biscardi et al. (2008). Barber DeGraaff et al. (2007) demonstrated that SBFs are also a powerful tool to determine the distance to barred lenticular galaxies. For the nearby SB0 galaxy NGC 1533 in the Dorado group, the authors derive (m − M) = 31.44 ± 0.12 mag, corresponding to d = 19.4 ± 1.1 Mpc. More recently Blakeslee et al. (2010) established an empirical calibration of the SBF method for the ACS/F814W filter (similar to the Johnson I-band). Because of the much higher throughput, the F814W (I₈₁₄) bandpass is more efficient and therefore to be preferred over the ACS/F850LP (z₈₅₀) bandpass (see Table 1 for a comparison). However, the most extensive observing campaign has been performed in the ACS/WFC z₈₅₀ band as part of the ACS Virgo and Fornax Cluster Surveys (ACSVCS, ACSFCS). In total, distances to 135 nearby galaxies were derived, for 90 early-type galaxies in the Virgo cluster (Mei et al. 2005, 2007) and 43 galaxies in the Fornax cluster (Blakeslee et al. 2009); the latter also give a refinement of the calibration of the combined cluster samples. Moreover, the SBF distances to two dwarf galaxies in Virgo were presented in Blakeslee et al. (2010).

3.3.2. Near-Infrared

Using HST/WFPC2 observations to the galaxy NGC 4373 in the Hydra-Centaurus supercluster, Pahre et al. (1999) found a good agreement (difference in P₀ ∼ 2%) with the results by Tonry et al. (1990, 1997), the latter using a different reduction and analysis package.

Jensen and collaborators measured the IR SBF distances to 16 galaxies located in the Leo, Virgo and Fornax clusters with ≤ 10,000 km s⁻¹ using the F160W NICMOS camera onboard the HST (Jensen et al. 2001). They were not able to give tight constraints on the Hubble constant. For the Hubble constant they derived H₀ = 72 ± 2.3 − 76 ± 1.3(statistic) ± 6(systematic) km s⁻¹ Mpc⁻¹, depending on the distance of the galaxies under consideration. Because the Local Group is located in an underdense region of the universe, the measurement is biased and therefore results in uncertainties and in a larger value of H₀ = 76 km s⁻¹ Mpc⁻¹. Soon thereafter, the team extended their efforts and established HST/NICMOS F160W SBF measurements for 65 E+S0 galaxies located in different environments (Jensen et al. 2003). Interestingly, they found evidence for intermediate age stars in the stellar populations of the galaxies. The central bluer colours of the E+S0 galaxies suggested younger stellar populations showing signs of recent star formation that are more metal rich. The SBF technique is therefore an independent tool to give insights to the composite stellar populations of early-type galaxies.

Figure 4 shows the Hubble diagram for galaxies that have SBF distances. The redshift in the cosmic microwave background (CMB) frame is displayed as a function of distance for ground-based data of the I-band SBF survey, divided into different galaxy types (Tonry et al. 2001), complemented by additional high-S / N HST/NICMOS F160W SBF measurements (Jensen et al. 2003). All redshifts were corrected for the peculiar velocity of the Local Group of galaxies with respect to the CMB. The expansion of the universe is clearly apparent and a linear, unconstrained fit to the data gives a Hubble ratio of 75 ± 4(stat) ± 7(sys) km s⁻¹ Mpc⁻¹. Local peculiar velocities are the origin for scatter in the Hubble ratio at distances of 40 Mpc, but the majority of data at this distance are pretty well described by the derived Hubble value. Part of the scatter in the SBF Hubble diagram is due to contributions of external sources (e.g., differences in the stellar populations), but does not solely rely on uncertainties in the distances. In the next few years, the SBF Hubble diagram will be better defined at distances reaching out into the Hubble flow (d > 40 Mpc) and likely be extended up to 100 Mpc using high-quality measurements of the WFC3/IR instrument onboard the HST.

Figure 4. Hubble diagram for galaxies with SBF distances. Recession velocities in the CMB frame are plotted as a function of distance for ground-based data (Tonry et al. 2001, circles and squares) and HST/NICMOS measurements (Jensen et al. 2003, triangles). The line has a slope of 75 km s−1 Mpc−1. Typical 1σ uncertainties are indicated for the HST data. For clarity, error bars are omitted for the ground-based data. The very deviant points near 39 Mpc are galaxies with high peculiar velocities belonging to the Cen 45 cluster of galaxies.

3.4. Stellar Population Gradients

Early-type galaxies are well known to show radial colour gradients due to variations in their stellar population properties (e.g., de Vaucouleurs 1961; Peletier et al. 1990). The first SBF gradients were discovered in ground-based observations for individual galaxies (Tonry 1991; Sodemann & Thomsen 1995, 1996; Jensen et al. 1996; Fritz 2000, 2002).

Using ACS/WFC observations, Cantiello and collaborators have investigated the SBF gradients of a larger sample of nearby galaxies (Cantiello et al. 2005, 2007a). Examples of SBF colour gradients are shown in Figure 5. The size of the slope between the intrinsic internal galaxy SBFs (_I) and the galaxy colour ((B − I)₀) for multiple annuli within a galaxy allows one to disentangle whether the gradients are due to age or metallicity variations. For the majority of galaxies in Figure 5, the gradients are metallicity driven (e.g., NGC 3258), whereas for a few cases the detection of a shallower slope suggests an age gradient which is equally or more important (e.g., NGC 1344). Galaxies with age gradients indicate a bluer, younger stellar population at larger galaxy radii and some evidence for recent merging activity or accretion events. More studies on this topic involving larger samples would allow scientists to draw further conclusions on the origin of SBF gradients.

Figure 5. Apparent I-band SBF magnitudes in a series of concentric annuli plotted as a function of (B − I)0 galaxy colour (Cantiello et al. 2005). With increasing radial distance from the galaxy centre, the mean galaxy colour becomes bluer and I brighter. The vertical offsets are due to the different distances of the galaxies.

For the foreseeable future, the WFC3/IR camera onboard HST offers a powerful new possibility of investigating the internal SBF gradients in the NIR (see also Section 8). It would be extremely exciting to compare the optical gradients with their NIR counterparts. In the NIR, pure age or metallicity variations affect the SBF magnitude dependence on galaxy colour very differently. More comprehensive studies on SBF gradients, in particular in the NIR, might allow us to disentangle the age/metallicity degeneracy in elliptical galaxies and therefore provide new insights into their different evolutionary histories.

² HST/NICMOS: http://www.stsci.edu/hst/nicmos Back.

³ HST/ACS: http://www.stsci.edu/hst/acs Back.

⁴ HST/WFPC2: http://www.stsci.edu/hst/wfpc2 Back.

⁵ HST/WFC3: http://www.stsci.edu/hst/wfc3 Back.

⁶ ACS-WFC: http://www.stsci.edu/hst/acs/ Back.

⁷ HST-IHB: http://www.stsci.edu/hst/wfpc2/documents/handbook/IHB_17.html Back.