DEEP HST/STIS VISIBLE-LIGHT IMAGING OF DEBRIS SYSTEMS AROUND SOLAR ANALOG HOSTS

Introductory notes. HD 207129 is a G2V close-solar analog and is one of the two oldest stars (along with HD 202628; Section 7.2) for which a CDS has been imaged in scattered light. The presence of cold dust around HD 207129 was inferred by Walker & Wolsterncroft (1988) from 12.5 to 60 μm IRAS excesses, and confirmed by Jordin de Muizon et al. (1999) with 2.5–180 μm ISO photometry with far-IR excess in its spectral energy distribution (SED) at >20 μm. A spatially extended disk was detected in thermal emission in Spitzer/MIPS imaging at 70 μm by Krist et al. (2010). Correcting for the broadening effects of the instrumental beam, they deduced an intrinsic size of an apparently elliptical disk at 70 μm of 188 × 81 (300 × 130 au). The disk, observed as part of the Herschel/DUNES program, has also been resolved with PACS imaging in thermal emission from 70 to 160 μm (Marshall et al. 2011) as modeled by Lohne et al. (2012). Krist et al. (2010) additionally resolved the disk with HST/ACS visible-light coronagraphy at 0.6 μm. The ACS image, derived from two-roll "self-subtraction" of the underlying stellar PSF, revealed a ring-like disk inclined 60° from face-on with a major axis P.A. of 127°, in approximate geometrical agreement with the 70 μm MIPS and Herschel/PACS images.

Observations and PSF subtraction (details). HST/STIS 6R-PSFTSC observations of HD 207129, and its color-matched PSF-template star Tau¹ Gru, were conducted according to the GO 13786 observing plan with no variances or anomalies on 2015 May 26 (visits 15–18) and 2015 August 8 (visits 11–14). With visit-level PSF subtraction, a low SB scattered-light excess from the disk was detectable (at low S/N) in each of the individual 1508 s exposure WedgeA-1.0 images, co-rotating with visit-to-visit reorientation of the celestial field. Despite the very close color-index matching of its PSF-template star, a stellocentric, (first-order) azimuthally symmetric, and radially modulated residual pattern (seen also independently in the WedgeA-0.6 images) at r < 6'' was characteristically betrayed by a small chromatic error in PSF subtraction with stellocentric angle (undersubtraction at r < 1'', oversubtraction at 1'' ≤ r ≤ 35, undersubtraction at 35 ≤ r < 6''). We illustrate this (hard stretched for maximum visibility), uncorrected, after combining all images in a common celestial frame in Figure 1 panel (A). To largely mitigate this problem, we follow the empirical color-correction process detailed in Sch14 as applied therein by example for HD 92945 (with source-masking here appropriate for HD 207129); see Figure 1 panel (B). After correction, the WedgeA-1.0 deep images and WedgeA-0.6 shallow images, in combination, both well reveal the ring-like disk reported by Krist et al. (2010) and provide SB limits on any dust (or lack thereof) interior to the starlight-scattering debris ring itself. For better visualization of the ring-like disk, in Figure 1 panel (C) we convolve the color-corrected image with a 3 × 3 pixel boxcar smoothing kernel to reduce the pixel-to-pixel instrumental background noise.

Zoom In Zoom Out Reset image size

Principal results. The STIS 6R-PSFTSC scattered-light imaging of the HD 207129 CDS (Figure 1) reveals an angularly large, ring-like disk of low SB at an intermediate inclination angle with a morphology as generally described by Krist et al. (2010). The broad ring-like disk appears to be devoid of starlight-scattering material interior to its inner edge to a stellocentric angle limit of r ≈ 5'' where, at smaller stellocentric angles, residual starlight dominates the centrosymmetric background. The SB ring model (Section 6) that best fits the observed data (by minimizing the sum of the squares of the observed-minus-model image residuals) is illustrated in Figure 2, with the model parameters given in Table 4. Although the model also allows for a stellocentric offset, none of significance was evidenced in initial fits to the HD 207129 image, and so it was fixed to zero, i.e., constraining the apparent ring center to be coincident with the location of the host star. We did not treat g iteratively, but produced a grid of models incrementally increasing from g = 0.0 (the isotropic case) in steps of Δg = +0.05, for which a best global fit was found for g ∼ 0.25.

Zoom In Zoom Out Reset image size

Table 4.  Debris Ring Parameters for HD 207129

Inclination (i)	579 ± 2° from face-on
P.A. major axis	122° ± 1° E of N
Ring peak SB on major axis (a)	934 ± 003	149.4 ± 1.0 aua
Inner-edge radius of ring HWHM major axis	713 ± 004	115.7 ± 0.9 aua
Outer-edge radius of ring HWHM major axis	1175 ± 003	188.0 ± 1.2 aua
Ring width (FWHM; Δr)	72.3 ± 1.5 au
g (HG scattering asymmetry parameter)	0.25 ± 0.05
Total ring flux densitya	310 ± 29 counts s−1 pixel−1	141 ± 13 μJy
Disk (0.6 μm) scattering fractionb	8.2 × 10−6 ± ∼8 × 10−7

Notes.

^aIncludes uncertainty in stellar parallax from SIMBAD. ^bV_star = 5.58 → 17.1 Jy.

Download table as:  ASCIITypeset image

The model appears to give a very good fit overall, though may slightly oversubtract the disk flux asymmetrically with respect to the locations (radii) of the peak brightness at some stellocentric azimuth angles. This is marginally evidenced in the somewhat darker (black) distorted "ring" of low-amplitude correlated negative-going residuals visible in Figure 2 panel (C). This may be indicative of the limitations of the current model, which does not account for possible anisotropies in the dust density or non-HG scattering properties.

1. Brightness and scattering fraction. In Figure 3 we compare the observed and model radial SB profiles along the disk's major axis where, at the ring ansae, the image contrast is less challenging and more accurately measured than closer to the minor axis. It is in this region that the deficit of light-scattering dust interior to the debris ring itself is visually apparent (e.g., in Figure 2 panel (A)). Using the model parameters in Table 4 closely replicates the main features of the observed profile on both sides of the star across the ring ansae, though some local deviations are seen—in particular here at ∼5'' < r < 7'' where the model slightly overpredicts the observed brightness. This could be due to a deficiency in the model, or simply a bias in the signal as the stellocentric angle is approaching the noise-dominated regime in the image at r < 5''. The peak SB of the debris ring along its major axis, at its diametrically opposed ansae, appears marginally brighter on its northwestern side at ∼0.96 μJy arcsec⁻² (23.7 V_mag arcsec⁻²) and on its southeastern side 0.68 μJy arcsec⁻² (24.1 V_mag arcsec⁻²). The total 0.6 μm flux density of the disk, as informed by the best-fit model, is 141 ± 13 μJy. This gives rise to a total disk optical scattering fraction, F_disk/F_star, of ∼8.2 × 10⁻⁶ to its host star with V_mag = +5.58. This is in good agreement with the prior estimation by Krist et al. (2010) of ∼7.6 × 10⁻⁶, independently derived with a different data set with iterative roll subtraction. Flux-loss and/or larger photometric uncertainties from "self-subtraction" for some disk geometries with this method are possible. However, the good agreement (≈±5%) in the independently derived global scattering fraction for the HD 207129 CDS suggests that the original estimation by Krist et al. (2010) was photometrically well calibrated.

Zoom In Zoom Out Reset image size

2. Anisotropic ("front/back") scattering asymmetry. From an analysis of the ACS imaging of the HD 207129 CDS from the 2006 epoch, Krist et al. (2010) suggested "at 0.6 μm, the ring shows no significant brightness asymmetry, implying little or no forward scattering by its constituent dust." They asserted g < 0.1 with an interpretive consequence that the "nearly isotropic scattering in the ring conflicts with any assumption of spherical particles that scatter according to Mie theory" (separate from a low-albedo inconsistency for silicate grains). The STIS data, in contrast, inform directionally preferential anisotropic scattering that, with g ≈ 0.25 from the best-fit modeling, is visually apparent as a "front/back" asymmetry (see Figure 2 panel (A)). In validation, though noisier, this same "front/back" asymmetry was seen in the STIS data when data from the first and second epochs were reduced separately into two independent three-roll images originating at different field orientation angles. The reason for this discrepancy in g with the prior published ACS imaging is unclear. We speculate that it may be related to a lower S/N (with shorter total integration time, narrower bandwidth, and lesser coronagraphic throughput) in the ACS data, or to the differing methods of reduction. The ACS image was derived from "roll subtraction," using the host star as its own PSF template. This obviates chromatic effects in PSF subtraction but, as noted previously, is subject to possible localized loss of disk flux from self-subtraction, with the dependence on stellocentric distance varying with azimuth angle, i.e., differently in the ring at the minor and major axes of the disk. The disparity may be more related to image structure (i.e., fidelity) than photometric efficacy, because both the STIS (six-roll) and ACS (two-roll) reduced data sets have relatively good agreement in the 0.6 μm total disk flux density and average ansal SB.

3. Radial sensitivity to starlight-scattering material:

• a.  
  Interior (endo-ring) dust. We do not detect any light-scattering dust (either diffuse or in a secondary ring) comparable in brightness to the debris ring itself to an inner limiting stellocentric angle of ∼5''. To quantitatively estimate the limits on imaging sensitivity to any (undetected) CS dust in the presence of incompletely suppressed starlight at, and interior to, this region, we plot in Figure 4 the 360° median absolute deviation¹⁷ (MAD) in the color-corrected PSF subtraction residuals as a function of stellocentric angle after subtracting the debris ring model of Figure 2 panel (B). Very similar results are obtained with direct measurement of the stellocentric field, masking the regions along the disk's minor axis that are superimposed upon the centrosymmetric residual pattern.
• b.  
  Exterior (exo-ring) dust. We also do not detect any light-scattering material beyond the exo-ansal radial decay of SB as shown, e.g., along the disk's major axis in Figure 3. Beyond the ring seen in sky-plane projection, the backgrounds in the sectors toward and flanking its semiminor axes at r ≥ 9''–10'' are wholly unaffected by the presence of the interior disk and the exterior is radially invariant. We directly assess the limits on imaging sensitivity in these regions (representative of fully circum-azimuthal measures) that are not dominated by PSF-subtraction residuals via multi-position aperture photometry as ≈25.8 ± 0.3 V_mag arcsec⁻².
• c.  
  Intra-ring dust. r ≈ 9''–10'' is also the distance to the ring peaks along the major axis, and so we approximately assess an intra-ring ansal average S/N ≈ 2.4 arcsec⁻².

Zoom In Zoom Out Reset image size

Introductory notes. HD 202628 is one of 10 nearby (<40 pc) solar analog (K0–G5) stars that were surveyed by Krist et al. (2012) for the presence of starlight-scattering CS dust potentially detectable with HST visible-light coronagraphy. This sample was IR-selected with Spitzer-derived excess thermal emission L_ir/L_star > 10⁻⁴, indicative of cold dust at stellocentric distances potentially within the imaging sensitivity capabilities of STIS for systems with favorable viewing geometries, dust properties, and ring-like morphologies with inner radii beyond the inner working angle limits of the coronagraph configuration employed. HD 202628 is a nearby (24.4 pc) solar-type (G2V) star with a thermal infrared excess L_ir/L_star = 1.4 × 10⁻⁴ (Koerner et al. 2010) estimated from 70 μm Spitzer/MIPS photometry, but lacking an excess at 24 μm, from which Krist et al. (2012) postulated an inner radius for the emissive dust of 80 au (41). Of the 10 stars surveyed, HD 202628 was the only one for which a visible-light debris system was imaged. The discovery imaging revealed a low-SB, intermediate-inclination, ring-like disk. Krist et al. estimated a semimajor axis length to the inner radius of the debris ring of ∼158 au, with characteristic width ∼58 au averaged radially across both ansae, and noted a significant stellocentric offset.

Observations and PSF subtraction. On 2015 May 30 and September 17, HST/STIS coronagraphic images of the HD 202628 CDS were obtained to improve, and expand upon, the discovery-mode imaging of Krist et al. (2012). The discovery images were (a) obtained using STIS occulting Wedge-A at its 18 wide location, (b) acquired in two HST orbits differing in a single orientation differential of 28°, (c) with total integration time 4512 s, (d) without contemporaneously observed, nearby, color-matched PSF-template stars, and (e) reduced with two-roll "self-subtraction" of the underlying stellar PSF. The GO 13786 follow-up images were (a) obtained using STIS occulting wedge-A at a combination of its (narrower) 06 and 10 wide locations, (b) acquired in six HST orbits with five field orientation differentials spanning 262° (see Table 3), (c) with total integration time 12,670 s, (d) with target-specific PSF reference stars, and (e) reduced with six-roll PSF-template subtraction.

All observations were executed as planned at both observational epochs. However, the preplanned PSF-template star for the first observational epoch (visit 07), BX Mic, was found to be "polluted" by third-light from an a priori unrecognized close-proximity companion (or ill-placed background star) rendering it unsuitable as a PSF-subtraction template. For the observations in the second epoch, the PSF-template star was changed to an alternative candidate, HR 8042, and found to be an excellent match. Fortuitously, PSF-subtraction using the second-epoch HD 8042 imaging showed little (if any) degradation due to possible differential breathing when applied to the first-epoch (visit 05, 06, 08) observations of HD 202628, and so was used as a PSF template for both sets of HD 202628 visits. We applied the same method of image reduction, and two-wedge six-roll visit combinations of PSF-subtracted images as discussed for HD 207129 (Section 7.1), to produce the AQ data image (shown in Figure 5) discussed herein.

Zoom In Zoom Out Reset image size

Principal results. The STIS 6R-PSFTSC AQ image of the HD 202628 CDS (Figure 5), from visual inspection, is generally in good morphological and photometric agreement with the discovery image of Krist et al. (2012) (see their Figure 2 top panel), though we note and discuss below some differences in details. The 6R-PSFTSC AQ image directly informs of the spatial distribution of the CS dust at all azimuth angles at r > 35–40, of greater apparent complexity than HD 207129. Principal geometric, photometric, and astrometric results discussed are summarized in Table 5.

Table 5.  Debris Ring Parameters of HD 202628

Inclination (i) from face-on	542 ± ∼ 4° (axial ratio: 1.708:1)
P.A. of major axis of debris ring	124° ± ∼ 3° (E of celestial N)
a (elliptical annulus of peak SB)	719		175.3 au
50% EDGES and PEAK RADII	From  Star		From  Disk Center
NW inner edge (major axis)	6409	156.4 au	5940	144.9 au
Peak SB radius of NW major axis	7628	186.1 au	7159	174.7 au
NW outer edge (major axis)	8948	218.3 au	8479	206.9 au
SE inner edge (major axis)	5522	134.7 au	5991	146.2 au
Peak SB radius of SE major axis	6740	164.5 au	7209	179.5 au
SE outer edge (major axis)	8060	196.7 au	8529	209.9 au
SB SLOPES (Power-law Index)	SE (Disk-centered)		NW (Disk-centered)
142 au < R < 158 au (inner)	+12.6		+7.7
175 au < R < 200 au (outer)	−4.5		−12.9
200 au < R < 250 au (outer)	−3.9		−2.8
250 au < R < 300 au (outer)	−6.9		⋯
NW Ring width (FWHM; Δr)	2539; 60.0 au
g (HG scattering asymmetry)	0.1 ± 0.1
Stellocentric offset from ring center in sky-plane projectiona	+0469 (ΔX), −0277 (ΔY) = 0535 toward ring-azimuth 2394
Stellocentric offset from ring center in face-on deprojection	0683 16.7 au
SB @ NW ansa	0.40 μJy arcsec−2		24.6 Vmag arcsec−2
SB @ SE ansa	0.51 μJy arcsec−2		24.4 Vmag arcsec−2
Total CDS flux density (Fdisk)	44.6 ± ∼9 μJy
Disk (0.6 μm) scattering fraction (Fdisk/Fstar)	7.65 × 10−6

Note.

^aMeasured along the disk's major (X) and minor (Y) axes.

Download table as:  ASCIITypeset image

1. Morphology. The HD 202628 CDS appears to comprise, in morphological description, two main components. First, dominating the system, a stellocentrically offset debris ring, and second, a low-SB diffuse "cloud" of exo-ring scattering material seen both exterior and somewhat interior to only the SE ansa.

• a.  
  The dust-scattered starlight from the ring appears to be mostly confined (within 50%-peak SB inner and outer isophotes) in an elliptical annulus between ∼135 and 218 au from the host star as seen in sky-plane projection. Interior to the ring, there is a notable deficit of scattering material in the two diametrically opposed sectors flanking the major axis (P.A. = 124°) from 17° < P.A. < 158° and 271° < P.A. < 354° to an inner working angle limit of ∼4'' where residual starlight begins to dominate the background interior to the debris ring. The debris ring itself deviates from bilateral symmetry with its geometrical center offset from the location of its host star.
• b.  
  A notable low-SB excess of diffuse scattering material is seen externally to, and somewhat superimposed upon and internally to, the SE ansa (not apparent in the discovery imaging of Krist et al. 2012), but with none seen at or beyond the diametrically opposed NW ansa; see Figure 8 panel (B). This "cloud" of material has a comparatively very shallow outer slope, leaving the outer "edge" of the ring beyond the SE ansa not as sharp or well defined as for the NW ansa. It may also apparently "intrude" (in morphological superposition) a short distance interior to the brightest part of the ring itself. The eastern sector of the ring between the NE minor axis (at P.A. = 34°) counterclockwise to near the SE ansa is marginally brighter (and/or wider at comparable brightness) than the diametrically opposed sector.
• c.  
  Krist et al. (2012) found the SB of the ring's SE ansa itself about half as bright as the NW ansa. This is an anisotropy that is not reproduced in the follow-up imaging presented here (e.g., see the radial profile of the major axis in Figure 7). We speculate that the previously reported ansal SB asymmetry and dearth of detectable exo-ring material beyond the SE ansa in the discovery imaging may have been a consequence of two orientation roll-subtraction partially cancelling the diffuse dust signal at these conjoined locations. This conjecture may eventually be tested with additional multi-roll imaging from HST GO program 13455 (J. Krist 2015, private communication).
• d.  
  As is also discernable in the prior discovery imaging, the apparent geometrical center of the debris ring in the follow-up imaging is offset to the east (toward P.A. ∼ 60°) from the location of its stellar host. However, in the follow-up imaging no (expected) corresponding pericentric "glow" is seen. This suggests the possibility of an azimuthally nonuniform dust density distribution (with or without anisotropic scattering) that, elsewhere, may be also locally implied by the presence of the SE primarily exo-ansal diffuse cloud of material.

2. Geometry. Beginning with a first visual estimation of the debris-ring geometry from the AQ image, we then parametrically fitted the ring-like component of the CDS to the same model described in Section 6. Our best fits to the simple ring models, illustrated in Figure 6, inform a less steep intermediate inclination of ∼54° from face-on, as opposed to ∼64° estimated previously by Krist et al. (2012) from apparent major:minor axial length ratios. These authors noted that in the discovery imaging most of the disk in the direction of its minor axis was obscured (remediated in the follow-up six-roll imaging reported here). We conjecture that with this lack of additional information there may have been a bias in the original estimation of the inclination. We also note a small (∼6°) difference with the prior estimation of the celestial position angle of the disk's major axis, derived as a free parameter in visual ellipse fitting with an a priori adopted inclination that had been fitted only to the inner edge of the elliptical annulus in the discovery imaging. While no formal errors were given by Krist et al. (2012) for these, and other estimated parameters, they note that reliable least-squares fitting was precluded due to low S/N. Given the above we (somewhat subjectively) do not find these differences discrepant with significance.

Zoom In Zoom Out Reset image size

3. Stellocentric offset. The originating set of GO 13786 images at each of the six field orientation angles, before PSF subtraction, record the HST diffraction spikes centered on the occulted star. Following Sch14, using the "X marks the spot" diffraction-spike fitting method, we are readily able to determine the location of the occulted star in the 6R-PSFTSC AQ image with an uncertainty of ∼±4 mas rms. We directly compare this to the geometrical center of the best-fit ring models to ascertain the presence and amount of a stellocentric offset of the debris ring that may inform a forced eccentricity (e.g., by an unseen planet). Krist et al. (2012) identified such an offset in the discovery imaging estimated from ellipse fits to the inner edges of the debris ring (only). We confirm such an offset from the follow-up imaging wherein we find from our best-fit ring model the apparent geometrical center of the debris ring offset from the star by (ΔX, ΔY) = [−047, +028] (measured in a Cartesian coordinate system defined by the major and minor axes of the debris ring seen in sky-plane projection), i.e., an apparent offset of 0545 toward P.A. = 2394. With sky-plane deprojection, assuming an inclination of 542 from face on, this is a deprojected offset of 067 (16.4 au). This offset, determined in fitting the entire debris ring ellipse, while closely agreeing in celestial P.A., is smaller than estimated by Krist et al. (2012) using the inner edge only in the discovery imaging.

4. (Nearly) isotropic scattering. With an assumption of an azimuthally uniform dust density distribution, a pericentric brightening of the debris ring of ≈+13% is expected toward P.A. ≈ 240° (the direction from the center of the debris ring toward the host star), and is encoded in the isotropic (g = 0.0) scattering model illustrated in Figure 6 (middle panels). This, however, is not seen in the observed image (top panel), and indeed it is the diametrically opposed sector toward P.A. ≈ 60° that is marginally brighter. Such a brightening, independent of the r⁻² scaling expected with offset stellocentric distances, suggests the possibility of an enhancement in dust density to the NE (though we cannot rule out some contribution to the possibility of some directionally preferential scattering). This is also consistent with contiguously adjacent brightening by the low SB "cloud" of material seen beyond the eastern ansa and may (or may not) be a physically contiguous feature. The ring itself is relatively well fit, over the explorable range of scattering phase angles, assuming nearly isotropic Henyey–Greenstein scattering, g = 0.1 ± 0.1, with the forward scattering direction along the ring's minor axis to the NE with residuals shown for both flanking cases in Figure 6 (bottom panels). After subtracting the g = 0.0 model, the residuals along the ring are enhanced relative to g = 0.2 to the north of the star (toward scattering phase angle zero) and diminished to the south, and vice versa. (This is readily apparent when "blinking" between these two images.) Those residuals, however, are (globally) not fully eliminated with intermediate asymmetry parameter values also modeled. In any case, the residuals to the SE of the star, near and beyond the SE ring ansa are (not surprisingly) dominated by excess light from the previously discussed exo-ring "cloud" that is not included in the simple ring model.

5. "Clumpiness" along the ring? A careful examination of the spatial distribution of the residuals in the location of the ring, after subtracting the best-fit model, reveals some "clumpy" substructures, in particular along its southern side (see Figure 6 bottom panel) on multi-pixel spatial scales that one would not expect from an unperturbed (smooth) dust density distribution. These clumps may simply be imaging artifacts at the faint limit of image fidelity—though comparable posited artifacts are not seen elsewhere in the field—or possibly confusion from ill-placed background sources. In principle, with four years of stellar proper motion of ∼10, the discovery image of Krist et al. (2012) could be used to arbitrate the latter. However, the same substructure pattern is not apparent in a discovery-epoch, roll-subtraction, image (see their Figure 2 top panel). This is likely due (if "real") to a combination of poor roll-angle coverage (wedge obscuration) in the sector(s) flanking the disk's minor axis, and differences in imaging efficacy with roll-subtraction versus multi-roll PSF-template subtraction. At this time a possible intrinsic origin for the intra-ring substructures seen in the follow-up images cannot be confirmed, but is testable with comparable (or improved) third-epoch imaging in the future.

6. Radial SB profile of the major axis. The major axis radial SB profile of the CDS, presented in Figure 7 (black line), was measured along the disk's major axis in both directions extending ±15'' (∼±300 pixels) from the center of the debris ring. The rectangular measurement region with its long dimension parallel to the major axis of the debris ring is depicted in Figure 6 (top panel). Along that strip we measured the average flux density in bins one pixel wide and ±20 pixels in extent orthogonal to the disk's major axis, at stellocentric angles >5'' (beyond the central noise-dominated region). In Figure 7 we also overplotted the identically measured major-axis radial SB profiles of the best-fit ring models previously discussed with characterizing parameters given in Table 5. Along the major axis (and therein through the diametrically opposed ansae) the difference in the g = 0.0 and g = 0.2 models is indistinguishable at the resolution of Figure 7. Here, we constrained the models to reproduce (scale) the SB of the ring at the NW ansa. The ∼26% difference in the predicted diametrically opposing ansal flux densities arises from the stellocentric offset of the debris ring. The NW side of the ring along the major axis is well fit to the model (with a noted deviation, however, at 75 < r < 87). The model, in the presence of the exo-cloud clump on the SE side of the CDS, underpredicts the SB, though the inner edge location (defined by the 50% intensity point) and outer "edge" slope (though offset in brightness) are well matched to the observations.

Zoom In Zoom Out Reset image size

7. Ring width and edge slopes. Following Schneider et al. (1999) (in the case of HR 4796A), we characterize and measure the inner and outer radii of the debris ring along its major axis from the locations where the radial SB has declined to 50% of its peak. We additionally estimate the SB gradient across the inner edge of the ring, i.e., its slope, fit with a radial power-law dependence along the major axis of the debris ring. For these (steep) inner edge slopes, we fit radial regions from 142 to 158 au (∼58 to 65) on both sides of the star centered on the disk. Radially beyond the radii of peak brightness, we fit separately in three contiguous regions between 175 and 300 au (see Table 5). These metrics, in combination with the ring diameter, may be used to set constraints on the locations and masses of unseen planets that may be responsible for the ring architecture, e.g., see Rodigas et al. (2014b). The presence of the low-SB cloud of material producing additional scattered light at and beyond the SE ring ansa biases the measurements of the adjoining outer slope and thus width, and should be viewed with caution. Physical modeling, beyond the scope of this current paper, is necessary to disentangle the two.

8. Exo-ring extent. Figure 7 also shows that the radial SB profiles of the major axis across both ring ansae are comparatively asymmetric, in particular beyond the radius of peak brightness. The observations inform (different from Krist et al. 2012) similar ansal peak SBs on opposite sides of the disk. Except for a local deficit between ∼74 and 85, on the NW side the model reproduces the outer-edge slope and fall-off in brightness to the level of the sky noise at r < 11'' (∼270 au). On the SE side of the disk, in the presence of a local enhancement in exo-ring brightness visible in Figure 5, the model underpredicts the brightness beyond the ring ansa to ∼125 (∼305 au) on (and flanking) the major axis. This may be evidence of exo-ring scattering material seen preferentially only beyond the NE ansa. Speculatively, such material may arise from the launching of small grains after an episodic collision, or from density perturbations ("confinement" by resonances) by unseen planet(s). Higher (still) S/N observations to better constrain the PSF would be beneficial to disambiguate the possibilities.

9. Total disk flux density and scattering fraction. Following chromatic correction in PSF subtraction, a faint, but not insignificant, azimuthally and centrosymmetric negative residual, diminishing with stellocentric distance, remains present beyond the ∼35 radius where highly structured PSF-subtraction residuals otherwise dominate the background (see Figure 8, panel (A), dark area beyond the digital mask). The innermost r < 35 region (shown in Figure 5) is digitally masked in making direct photometric measurements of the CDS. However, the presence of this remaining radial bias in the stellocentric background exterior, though of only minor consequence at the locations of the ring ansae at r ∼ 72, correspondingly biases aperture photometry by its inclusion in the directions of and flanking the ring's semiminor axes of length r ∼ 42. We therefore use the best-fit parametric model of the ring to estimate its total brightness, exclusive of the SE exo-ansal extension that is not included in the ring model. We refit with the same geometrical parameters of the ring, but to separately match the brightness of the ring in only the diametrically opposed 90° sectors flanking the major axis, where the radial background gradient is a much smaller bias. In this manner we roughly estimate that the full-ring fit underestimates the total ring flux by ∼30% in total flux density. With this systematic correction we then estimate the total instrumental brightness of the debris ring as 88 counts s⁻¹ ± ∼20% or (4.0 ± 0.8) × 10⁻⁵ Jy. This does not include the additional light scattered by the SE exo-ring "cloud." To estimate that contribution to the total brightness we subtract the best-fit ring model from the observed image and directly measure the (somewhat subjectively bounded as shown in Figure 8) region contributing an additional ∼(4.6 ± 0.9) × 10⁻⁶ μJy, i.e., F_disk = 45 ± ∼9 μJy. We adopt a catalog V_mag for HD 202628 of 6.75 Jy (5.83 Jy in the STIS 50CCD band), from which the optical scattering fraction of the disk is F_disk/F_star ≈ 7.7 × 10⁻⁶, for spectrally flat gains.

Zoom In Zoom Out Reset image size

10. SB of the debris ring. We characterize the representative SB of the CDS, independent of asymmetries expected from a stellocentric offset of the debris ring and anisotropic scattering, by measuring and averaging the SB at both ansae. We do so in circular regions, 1 arcsec² in area, centered on the radial SB peaks at the ring ansae. At this stellocentric distance along the disk's major axis (∼72) the residual chromatic error in the background that is corrected for in the estimation of the total disk flux is small to negligible. We find (averaging the two ansal brightnesses, which differ by ±12% with the NW side brighter) a mean SB of ∼0.46 μJy arcsec⁻² (or ∼24.5 V_mag arcsec⁻²).

Introductory notes. HD 202917 (HIP 105388; G7V, 43.0 pc) is a member of the Tucana–Horologium stellar association with an estimated age of 10–40 Myr (Moór et al. 2006). The presence of a thermally emissive CDS associated with HD 202917 was first suggested by Silverstone (2000) from 60–100 μm ISOphot data, and later confirmed from its ∼3 × 10⁻⁴ infrared excess determined from Spitzer/MIPS observations (Beichman et al. 2005). An asymmetric scattered-light excess to the north (only) of the star was suggested by Krist (2007) in "wide" V-band observations obtained in 2005 with ACS coronagraphy (HST program 10695) sensitive to dust-scattered starlight at stellocentric angular distance ≳15; see Figure 9 panel (A) where we have independently reduced those raw data obtained from the MAST archive with template PSF subtraction. Earlier (1999 in HST GO program 7226), NICMOS 1.6 μm coronagraphic images of HD 202917, designed for point-source (giant planet companion) detection by two-roll "self" subtraction, were obtained without PSF-template observations. Later (2005 in HST GO program 10849; one of 21 targets), images of HD 202917 at 1.1 μm were acquired for disk detection, but without a contemporaneous or specifically programmed PSF-template star. In neither case was a scattered-light excess from a CDS detected. More recently, Soummer et al. (2014) re-reduced those prior data with principal component analysis (PCA) using a now-available extensive suite of PSF reference data from the MAST-hosted¹⁸ (NICMOS coronagraphic) Legacy Archive PSF Library (Schneider et al. 2011). The PCA reduction of Soummer et al. at 1.1 μm (which provides the best NICMOS spatial resolution) is reproduced in Figure 9 panel (B), wherein the northern "half" of the disk (well correlated in position with the ACS image) was detected. From this image morphology, Soummer et al. (2014) suggested that the "disk exhibits an asymmetric arc suggestive of a partial ring inclined ∼70° to the line of sight with a major axis P.A. ≈ 300° [=120°]'', also noting that "the northwest side is significantly brighter and more extended than the southeast side." Our new STIS PSFTSC imaging (Figure 9 panel (C) and Figure 10) confirms these findings while providing both a higher-fidelity view of the ring-like disk and deconflating now well-seen exo-ring material from light scattered by the debris ring itself.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Observations and PSF subtraction (details). WedgeA-1.0: (a) All coronagraphic observations of HD 202917 from both the 2015 August 1 epoch (visits 55–58) and October 13 epoch (visits 51–54) were executed nominally as planned with no absolute constraints imposed on orientation. The observations of the WedgeA-1.0 PSF template of HD 8893 obtained contemporaneously (as visit 57) with the August 1 CDS images, however, were too deeply exposed. As a result, those images were saturated to larger-than-planned stellocentric angles in the directions orthogonal to the mid-line of Wedge-A to r ∼ 11 (22 pixels), and so do not, with application, provide information on the CDS at distances closer than this. However, the serendipitous absolute orientations for the relative roll constraints in these visits (but not the October 13 visits) had the Wedge-A itself obscuring the disk itself (so would not have been visible in these data in any case). (b) Exposure times for the HD 8893 PSF template visit (#53) of the second epoch were correctly adjusted prior to execution to circumvent saturation at r > 05 (10 pixels). With that, in the PSF-subtracted images of the second epoch (only), the ring-like disk is seen protruding on opposite sides of the Wedge-A at all three image orientations with a revealed major axis ring radius of ∼15 (29.5 pixels).

WEDGEA-0.6: (c) The debris ring itself is also seen, but with inferior fidelity, in the less deeply exposed WedgeA-0.6 PSF-subtracted imaging for the more favorable wedge orientations of (second epoch) visits 51–54. With their PSF subtraction residuals, they contribute only marginally to improving the Inner Working Angle (IWA) over WedgeA-1.0 imaging from these visits. (d) Separately, the WedgeA-0.6 PSF-template star in visit 57 of the first epoch (unfavorable wedge orientations), while properly exposed, was (unfortunately) de-centered in acquisition with respect to contemporaneous target imaging. With PSF subtraction this causes wedge-flanking excesses (or deficits) in instrumental brightness, obscuring stellocentric visibilities interior to the WedgeA-1.0 limits achieved, and contributes only "noise" in these smallest stellocentric regions.

AQ data image. For reasons (a)–(d) above, the AQ data images for WedgeA-1.0 and WedgeA-0.6 for HD 207129 combine the PSF-subtracted data as follows:

• (1)  
  for WedgeA-1.0: (disk-favored orientation) visits 51–54 are used to an IWA limited by the extent of the wedge,
• (2)  
  for WedgeA-1.0: (disk-disfavored orientation) visits 55–58 are used to a larger IWA limited by the PSF saturation distance,
• (3)  
  for WedgeA-0.6: (disk-favored orientation) visits 51–54 are used interior to regions unseen with WedgeA-1.0,
• (4)  
  for WedgeA-0.6: (disk-disfavored orientation) visits 55–58 are not used due to their inferior image quality.

N.B. With respect to (4), the WedgeA-0.6 AQ images independently confirm the ring structure and morphology seen in the WedgeA-1.0. However, they additionally contribute little—and with noise do not actually improve—knowledge of the very small additional area unsampled with the six-roll WedgeA-1.0 images. Therefore, in this case for quantitative measurements, only the six-roll WedgeA-1.0 combined AQ image is used.

Roll coverage. The HD 202917 CDS is well revealed in the STIS GO/13786 fully reduced (six-roll combined) AQ image, as shown in Figure 10 with the morphological major axis of its ring-like disk on the image horizontal. Due to constraints on the scheduling of HST, the six field orientations that satisfied other imaging requirements to produce a high-fidelity AQ image could not also provide the complete roll/IWA coverage needed to otherwise fully reveal the debris ring at all CS azimuth angles close to the star. As a result, the debris ring at, and in its opposing sectors closely flanking, its apparent minor axis was not imaged (digitally masked as shown in Figure 10). The remaining unobscured majority of the CDS, however, was imaged with high fidelity.

Principal results.

1. Morphology. Unlike prior ACS and NICMOS reprocessed images (Figure 10, panels (A) and (B)), the STIS image clearly reveals a complex "two-component" CDS morphology with (1) a relatively narrow "half" ring-like disk that is apparently (at least partially) centrally cleared, and (2) a larger and fainter exo-ring with a "fan-like" starlight-scattering structure. The latter is not unlike those seen in some other more highly inclined CDSs, e.g., HD 61005 and HD 32297 (Sch14) wherein, its debris ring is also seen on only one side of the star. Both "components" exhibit significant azimuthal asymmetries.

2. Anisotropic ("front/back") scattering asymmetry. The debris ring itself clearly exhibits a very strong "front/back" asymmetry in scattering phase angle, as suggested from the prior ACS and NICMOS imaging. The SB along the ring is asymmetrically "mirrored" about the major axis of the debris ring. This implicates strongly directionally preferential scattering, with (as a result) the fainter (southern) "half" of the debris ring undetected. The ring itself is seen to "turn around" to the south of both ansae, but only marginally so at its western extremity. The debris ring is undetected from stellocentric celestial P.A. ≈ 120°–280°. The exo-ring "fan" of material is, locally, significantly brighter beyond the ring's western ansa than on the diametrically opposed eastern side (see Figure 10 panel (B)). The broad outer scattering "fan" seen only to the north of the star is morphologically skewed, by about 7° clockwise, with respect to the apparent celestial orientation of the axes of the debris ring (see Figure 10 panel (C)).

3. Model fitting. The best fits to the ring-only model, minimizing the residuals along the visible portion of the debris ring after model subtraction, were found with a Henyey–Greenstein asymmetry parameter g = 0.6 ± 0.1 with a correlated uncertainty in scaling the brightness of the model to the observed data by ±∼20%. In Figure 11 we show these best fits with the model scaled in brightness by +20% with g = +0.5 and −20% with g = +0.7 (with the forward scattering direction toward the northern, visible, major axis) in Figure 11. To be clear, we do not attempt to model the exo-ring "fan," but rather to disambiguate the two "components" of the CDS. Parametric results are summarized in Table 6.

Zoom In Zoom Out Reset image size

Table 6.  HD 202917 Debris Ring Parameters

Inclination (i)	686 ± 15 from face-on
P. A. major axis	108° ± 1° E of N
Ring peak semimajor axis (a)	146 ± 002	62.8 ± 2.8 aua
Inner-edge radius of ring HWHM major axis	129 ± 002	55.4 ± 2.5 aua
Outer-edge radius of ring HWHM major axis	160 ± 002	68.6 ± 3.0 aua
Ring width (FWHM; Δr)	013 ± 003	13.16 ± 1.3 aua
g (HG scattering asymmetry parameter)	0.6 ±  0.1
Flux density of debris ring	182 ± 36 counts s−1 pixel−1	83 ± 17 μJy
Total CDS (fan+ring) flux densityb	∼660 counts s−1 pixel−1	318 μJy
Total CDS (0.6 μm) scattering fractionb	∼3 × 10−4

Notes.

^aIncludes uncertainty in stellar parallax from SIMBAD. ^bInner unsampled region obscured by coronagraph wedge and estimated with best-fit model.

Download table as:  ASCIITypeset image

4. Geometry of debris ring. The observed ring geometry has only minor dependences upon g within its estimated uncertainty (see below). In Table 6 we give the best-fit parameters for g = 0.6 with systematic uncertainties included for solutions with g varied by ±0.1. Fitting the (visible portion of the) debris ring to the best-fit scattering models (Figure 11), we find a semiminor:semimajor axial ratio ≈0.182 with a corresponding inclination ≈686 from a face-on viewing geometry for an (assumed) intrinsically circularly symmetric debris ring, with a celestial P.A. of its major axis of 108° east of north. We find the semimajor axis of the debris ring seen in sky-plane projection ≈146 (=62.8 au) in length as the distance between the SB peaks along the major axis in the best-fit model. The characterizing width of the ring, based upon its full-width to half-intensity maximum from inner edge to outer edge, is 13.2 au.

5. Asymmetries and anisotropic scattering. We note two major scattering anisotropies in the HD 207129 CDS. First, the debris ring itself exhibits highly directionally preferential scattering (g ∼ 0.6) with its brighter side toward its projected minor axis (Figure 10, panel (A)). The STIS data, unfortunately, did not image the brightest part of the ring. However, the imaged regions along the northern ring-arc appear asymmetrically bright on opposite sides of the minor axis (eastern side brighter). This is perhaps better seen after subtracting the bilaterally symmetric ring models (best fit in the range 0.5 < g < 0.7), as shown in Figure 10, panel (C), where a bright feature superimposed at the location of the ring-arc remains on the eastern side of the unsampled area, but not so on the western side. This suggests that there may be some azimuthally differentiated substructure along the debris ring. Second, scattering by exo-ring material on the western side of the CDS flanking (to the north) of the ring's major axis is significantly brighter than on its otherwise mirror-symmetric eastern side (Figure 10, panel (B)). This suggests a local enhancement in the surface density of scattering particles (through a recent collision or other cause) beyond the bright ring toward celestial P.A. ∼ 300°. This was suggested previously by Soummer et al. (2014) from the PCA-reduced NICMOS image, though the direction of the major axis of the debris ring that we now see is 12° further clockwise.

6. CDS extent. Exo-ring material of lower surface brightness in the CS "fan" is seen over the same range of scattering phase angles as for the ring itself (suggesting, perhaps, some common physical properties of the material). This material is detected to a maximum stellocentric distance (+3σ SB above the sky background) of ∼5'' toward celestial P.A.s ≈ 90° and 295°, just to the north of the western ansa of the debris ring where the aforementioned exo-ring bright scattering region is locally seen. This is coincident with the region detected in ACS PSF-subtracted imaging from archival data processed from the MAST. This feature is also likely the origin of the suggestion by Soummer et al. (2014) that the northwest side is significantly brighter (than the SE side), though the significance in the ACS image is not so apparent.

7. Photometry. To estimate the total flux density of the HD 207129 debris ring only, superimposed on its embedding larger CDS structure, we use the best-fit scattering models of the ring with uncertainties in brightness scaling for the range 0.5 < g < 0.7. To estimate the total flux density of the full CDS (both the "skirt" and the embedded ring together) we perform area-aperture photometry fully enclosing the CDS and then use the superimposed portion of the ring model coincident with the interior region of the CDS that was unsampled in the STIS image. Both are given in Table 6, showing that about 2/3 of the total CDS brightness originates in scattering by exo-ring material. With HD 202917 at V_mag = 8.67, the visible light scattering fraction of the CDS, F_disk/F_star, is ∼3 × 10⁻⁴. This is ∼40× "brighter" (in F_disk/F_star) than both of the ∼2.3 Gyr G-star disks discussed in Sections 7.1 and 7.2, and is approximately equal to its IR excess (2.5 × 10⁻⁴).

HD 207129 (G2V) and HD 202628 (G5V), both ∼2.3 Gyr in age, are the "faintest" solar-analog CDSs yet observed in terms of optical scattering fraction, F_disk/F_star, with both ∼8 × 10⁻⁶. Despite their ansal SBs of only ∼24.0 and 24.5 V_mag arcsec⁻², respectively, they both are readily detectable and mapped with STIS 6R-PSFTSC in part due to the large angular radii of their major axes of ∼93 and ∼71. This is, in angular scale, beyond the interior region where PSF subtraction residuals dominate. The optical scattering fractions of these two oldest solar-analog CDSs are in stark contrast to those of the ∼23 Myr HD 181327 (F6V) and ∼90 Myr HD 61005 (G8V) with their ∼200–300 times brighter F_disk/F_star.

The well-determined (≲10% uncertainty) total disk optical scattering fractions of all seven intermediate-inclination FGK-star ring-like CDSs that are shown in Figure 12 are plotted in Figure 13 (left panel) as a function of the estimated ages of their host stars, along with that also of the nearly edge-on HD 61005 (Sch14 and Figure 16). Age estimations and their uncertainties are taken from the literature; see Table 1 of this paper for HD 202917, HD 207129, and HD 202628, Mamajek & Bell (2014) for HD 181327, and Table 1 of Sch14 for the other stars¹⁹ plotted in Figure 13.

Zoom In Zoom Out Reset image size

Both HD 181327 and HD 61005 have been shown to possess very large, and optically bright, exo-ring scattered-light halos well beyond the peak radii of their thermal IR excess emission (Sch14); for HD 181327 this is posited as arising from a recent massive collision in its birth ring (Stark et al. 2015), and for HD 61005 with small-particle mass loss through a ram-pressure interaction with the local ISM (e.g., Hines et al. 2006). Excluding these two systems with non-steady-state dust production events, the remaining more quiescent CDSs are seen as an ensemble to decay in visible light scattering fractions, F_disk/F_star, over time (t) approximated by a t^−0.8 power law from ∼23 Myr to 2.3 Gyr. The robustness and more general applicability of this trend, however, should be viewed with due caution given the small sample size and origin. The fit is suggestive of an evolutionary optical dimming timescale for the overall ensemble, but in general may be quite inhomogeneous for individual CDSs that are diverse in stellar host, disk, and environmental characteristics. Additionally, this follow-up sample was culled from sensitivity-limited discovery-mode survey imaging for CDSs with L_ir/L_star ≥ 10⁻⁴, and thus could be biased for systems intrinsically fainter in their excess IR emission. The efficacy of the suggested power-law timescale remains to be tested when a larger, and fainter, homogeneous sample of solar-analog debris systems may be available in the future and is necessary to reliably establish a posited evolutionary sequence.

A similar, but somewhat less steep, decay in 24 μm IR excess emission, tracing larger particles, was predicted by Gaspar et al. (2013) over the same time for a broader range of host-star spectral types with also less massive disks than our HST down-selected sample with L_ir/L_star ≥ 10⁻⁴. The modeled average disk mass in their population synthesis has not yet reached quasi-steady state and is therefore decaying at a slower rate. However, the more massive members of their population synthesis have, and show, IR decay slopes in agreement with the visible-light observations of the more massive HST-observed disks. Steeper slopes, more closely matching those seen in our visible-light solar-analog ensemble, are predicted by Gaspar et al. (2013, see their Figure 6) for initially more massive disks.

For each of these spatially well-resolved systems, as an ensemble in Figure 13 (right panel), we also compare the total CDS flux density to that of the flux density in a 100 au² (seen in projection) region of their debris rings at their ansal SB peaks, also ratioed to their host star's flux density (F_ansa/F_star). A somewhat steeper, ∼t^−1.3, ansal-only power law is seen. Though this is not as well correlated in age for the younger (≲ 2 × 10⁸ Myr) systems alone, a similar declining trend is seen with the inclusion of the (∼10×) oldest systems. The larger dispersion about the trend line of F_ansa/F_star may be due to local anisotropies in the dust density distribution (clumping, deficits, or enhancements from dust production events) in the small (100 au²) areas measured in any given system at any given epoch.

Both of the oldest close-solar analog (G-star) CDSs reported in this paper, HD 207129 and HD 202968, have similar IR excesses of ∼1 × 10⁻⁴ (see Table 1).²⁰ This is only about an order of magnitude less IR-bright than the ∼100× younger stars in this sample (HD 181327 at 2.5 × 10⁻³ at ∼23 Myr and HD 202917 at 1.2 × 10⁻³ at ∼30 Myr) despite their advanced ages. Although no color information is available within the broad STIS spectral passband to assess wavelength dependence (see Section 8.3), the fact that F_disk/F_star for both is an order of magnitude fainter than their IR excesses suggests grains of low optical albedo in these two oldest starlight-scattering CDSs.

Starlight scattered by exoplanetary debris at visible wavelengths traces the small (micron sized) particles in CS environments. When ring-like structures exist, their locations, widths, and inner/outer edge slopes may predictively be able to constrain the dearth/existence, locations, masses, semimajor axes, and eccentricities of co-orbiting planets (Rodigas et al. 2014b). These small particles also react easily to non-planetary perturbative forces intrinsic to the systems such as stellar corpuscular winds, radiation pressure "pushing" material outward, and Poynting–Robertson drag causing particle orbits to decay inward. CDSs should not, however, be considered as closed or isolated systems. There now exists strong and growing evidence for ISM interactions with the small grain materials at large stellocentric distances in some CDSs manifested with the visual appearance of bow/termination shocks as well as large "skirts" of exo-ring material, for example in HD 61005 (Hines et al. 2007) and HD 32297 (Sch14). The latter may also be the case for HD 202917 and HD 15745 (Figure 14).

Zoom In Zoom Out Reset image size

8.2.1. Comparing HD 202917 to HD 15745 and HD 61005

We compare the morphology of the HD 202917 (Figure 10) CDS with those of HD 15745 (Figure 15) and HD 61005 (Figure 16). All are in the likely middle-age range of ∼30–90 Myr, with HD 202917 the youngest of the three. All are hosted by FGK stars, with HD 61005 (G8V) a very similar close-solar analog to HD 202917 (G7V). All three bear some significant morphological resemblances with large "skirts" or "fans" of exo-ring scattering material.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In Sch14, the HD 15745 CDS was discussed and presented to best illustrate its extended "fan"-like outer scattering structure (as so originally described from its ACS discovery imaging by Kalas et al. 2007), without emphasis on the debris ring itself. In Figure 15 we show the HD 15745 CDS to better illustrate its intermediate-inclination (i ≈ 57°), r = 117 (74 au), debris ring within. The ring is a lower-contrast feature, relative to its exo-ring "fan" (or "skirt"), than HD 202917 and HD 61005, and is obscured in a linear representation (Figure 14 panel (A), similar to Figure 13 panel (C) of Sch14). With a logarithmic compression of the dynamic display range, as in panel (B), the debris ring is readily apparent. In panel (C) we scale the SB in the plane of the ring to compensate for the 1/r² dimming of stellocentric illumination, which further enhances the visibility of the ring and of the endo-ring region of the CDS that is relatively devoid of starlight-scattering material. While the region of the ring along its minor axis (in particular flanking its fainter semiminor axis) was unimaged due to HST observing constraints, it is clear that the CDS exhibits highly directionally preferential scattering. This is also the case (more extremely) with HD 202917 at a somewhat similar intermediate (68°) inclination angle.

While the HD 61005 CDS has a highly inclined (∼85°) debris ring (unlike those in the intermediate-inclination FGK star sample in Figure 12) it, like HD 202917's, has nonetheless been well detected and resolved in scattered light—at least along its brighter "half" (e.g., Buenzli et al. 2010; Sch14; J. Olofsson et al. 2016, in preparation). Like HD 202917, a highly asymmetric scattering phase function is suggested as being responsible for the non-detection of the fainter "half" of HD 61005's debris ring. Quite differently, however, at its likely intermediate age of 90 ± 40 Myr, the HD 61005 CDS sits well "above" the power-law fit of F_disk/F_star ∼ t^−0.8 to the intermediate-inclination systems plotted in Figure 13.

HD 61005's near edge-on inclination (like also HD 32297's with a more massive A6V²¹ host) contributes to the HST detectability of its very large and asymmetric outer scattering structure (see Figure 16, top panel) of small (micron size) starlight-scattering grains. Such asymmetric structures have been posited as arising from an ISM wind blowing back the small grains in the disk halo (Artymowicz & Clampin 1997). In our viewing geometry this is manifested as a large exo-ring "skirt" of scattering material seen in sky-plane projection. Its endo-ring inner clearing is very apparent when compensating in the ring plane for the 1/r² attenuation of the stellar illumination²² ; see Figure 16, lower panel. The stirring/launching/ejection of the starlight-scattering exo-ring grains by ISM interaction may contribute to this CDS's notable morphology and to its F_disk/F_star brightness for its age, rather than, e.g., invoking a recent catastrophic collision.

The HD 202917 CDS, though less highly inclined to our line of sight than the HD 61005 CDS, bears morphological resemblance (at lower S/N) in its SB image (compare Figure 10 panel (C) to Figure 16 top). In both of these systems only the brighter "half" of the debris ring with a very large front-to-back scattering asymmetry is detected. Both, along with HD 15745, also possess large skirts or fans of exo-ring scattering material seen on the brighter side only of the CDSs. HD 202917 differs, however, from both HD 61005 and HD 15745 in that it is significantly asymmetrically brighter on one side (the east) of the ring's minor axis at and beyond the ring itself. The morphology of such a structure may be difficult to reproduce through an ISM interaction alone, and may be indicative of a recent dust enhancement event through a catastrophic collision or other dust production event. Kalas et al. (2007) disfavored ISM blowback as a cause of the HD 202917 fan, noting the star's tangential motion and location well within the local bubble. However, a plausibly analogous bow shock seen in IR observations of δ Vel (Gaspar et al. 2008), also in the local bubble with similar relative motion to the ISM, may suggest otherwise and is indicative of the fact that cloud patches of higher density exist in the local bubble. Reproducing the morphology of the HD 202917 CDS is a challenge to models of the structural evolution of disks inclusive of forces both intrinsic and extrinsic to the system (beyond the scope of this current paper).

8.2.2. Intercomparing the Oldsters—HD 207129 and HD 202628

Both filter-band color diagnostics and directionally preferential scattering derived from spatially resolved imaging photometry can inform or constrain the physical properties of the starlight-scattering materials (e.g., Debes et al. 2008a, 2008b). The former, however, cannot be obtained with STIS coronagraphic observations alone. However, the flux density and SB measures in the STIS 50CCD spectral band that we report in Tables 4–6 may be combined with future data to be obtained with continuing improvements to ground-based high-contrast imaging systems in the near IR, and with the advent of James Webb Space Telescope/nirCAM coronagraphy. Separately, photometrically determined scattering efficiencies as a function of the phase angle of stellocentric illumination with spatially resolved imaging, even in a single spectral band, parameterized by g, can provide a useful constraint in ascertaining the physical properties of materials in some CDSs. Unlike face-on or edge-on CDSs, the intermediate-inclination CDSs we focus on in this paper simultaneously provide an explorable range of circum-azimuthal scattering phase angles (lacking in face-on disks) that (with sufficient resolution and image contrast) are spatially resolvable with significance and without self-obscuration of debris structures (in edge-on disks).

In fitting the STIS 50CCD band SB distributions of the ring-like components of the intermediate-inclination CDSs of HD 202917, HD 207129, and HD 202628, we assumed a simple, single characterizing HG asymmetry parameter, g; see Section 6. More completely descriptive scattering models, e.g., those considering optical properties dependent upon a diversity in material composition, sizes, and mixes, are beyond the intended scope of this paper. Also see, e.g., Stark et al. (2014) for a discussion of the necessity for alternative or more complex formulations for some CDSs as explored in part for HD 181327. For the three G-star CDS rings we have observed, over the range of explorable scattering phase angles²³ , we found a range from nearly isotropic scattering (HD 202628, g ≈ 0.1) to mildly asymmetric scattering (HD 207129, g ≈ 0.25) in the case of the two oldest CDSs, to strongly asymmetric scattering (HD 202916, g ≈ 0.6) such that only the brighter side of the debris ring (centered on φ = 0°) was detectable.