Angular Offset AOx (arcsec)

Figure 18.16 Interferometric images of 13C16 O (J = 1 ^ 0) emission around the stars AB Aur (left) and HD 163296 (right). Angular offsets are relative to the phase center of each field, which does not coincide precisely with the stellar position, shown here symbolically. Also shown in each case is the beam size and orientation.

shape is hard to discern, as the minor axis is only marginally resolved. The mean, i. e., intensity-weighted, radial velocity exhibits a smooth gradient along the major axis, with a total range of 2 km s-1. Whether or not this variation represents Keplerian motion is still unclear. In the right panel, we show gas around the star HD 163296, now mapped in the 12 C16O (J =2 ^ 1) line at 1.3 mm. Here, the full structure is better resolved, but it is also more irregular in shape. The mean radial velocity again changes along the major axis.

Both AB Aur and HD 163296 have spectral types of early A. It is significant that nearly all the objects with excess millimeter emission, whether in the continuum or lines, are either Ae or late-Be stars. Conversely, stars whose spectral type is earlier than about B5 rarely have significant flux in this regime. It is true that more massive stars tend to be farther away, but the higher millimeter luminosities already seen would be detectable. The implication is that the circumstellar gas and dust being sampled by these observations is absent in the higher-mass population.

We have encountered this situation before, in a slightly different context. After noting the paucity of evidence for disks around main-sequence O stars, we saw in Chapter 15 how their ultraviolet photons are able to destroy, through photoevaporation, any structures inherited from the formation epoch. However, this mechanism is inadequate for the objects of interest here. Equation (15.50) gives the theoretical rate of efflux from a disk outside its gravitational radius mg. Consider a Herbig star of spectral type B4, corresponding to Teff = 1.7 x 104 K. If we substitute into equation (15.50) M* =7 Mq and use the appropriate ionizing photon output of N* = 6 x 1042 s_1, we find a mass loss rate of 3 x 10~8 Mq yr-1. Thus, a disk with initial mass 10 percent of M* would need 2 x 107 yr to vaporize, far longer than the object's pre-main-sequence lifetime. Either the disk never formed, or other factors, such as wind erosion, predominate.

Even among objects that do have millimeter flux, the additional disk signatures found in T Tauri stars are curiously weak or absent. For example, we recall that optical forbidden lines, most notably [OI] 6300 A, tend to be blueshifted. The common interpretation of this fact is that the redshifted emission that would arise from receding gas in the stellar wind is being occulted by an opaque disk. Many Herbig Ae/Be stars also exhibit [O I] emission, presumably arising from shocks generated in their winds. Usually, however, the line profile is either symmetric or only modestly suppressed in the red wing. This finding may mean that the winds are relatively broad, as suggested also by the apparent lack of molecular outflows and jets (§ 18.3).

Permitted optical emission lines in Herbig Ae/Be stars, such as those of the Balmer series, do not commonly exhibit deep absorption dips in the red. Such inverse P Cygni profiles are the spectroscopic sign of infall for the T Tauri class. Of course, we do not know the true source of the infalling gas, even among these objects. But if even part of the material flows from the inner regions of disks, then the rarity of the effect in intermediate-mass stars is puzzling. We note in passing that the infall signature does appear more frequently in ultraviolet lines. Perhaps a third of Herbig Ae stars show the effect, which is unseen at earlier spectral types.

These negative or ambiguous results should not allow us to lose sight of the secure fact that matter just outside Herbig Ae/Be stars is distributed anisotropically. The difficulty has been in establishing the specific, disk-like geometry and kinematics. As further evidence of anisotropy, we need only recall that the optical emission is often linearly polarized. Photons are evidently being scattered by dust grains in the stellar environment. These grains cannot reside in a purely spherical envelope, which would yield no net polarization. The effect is strongest, and most variable, in the UX Ori stars discussed previously. Note that these sources all have spectral types of either A or late B. It appears, once again, that the more massive Herbig Be stars have efficiently cleared out their immediate surroundings.

18.4.2 Dust Grain Processing

Those stars that still retain dusty material often exhibit broad emission near 10 pm, as mentioned in § 18.3. The detailed profile of this spectral feature is sensitive to the grain structure. It is interesting, therefore, that the profile evolves with time. Figure 18.17 illustrates the point through two representative spectra. The top panel shows observations of the now-familiar AB Aur, an A0 star with an age of 2 x 106 yr. Below is HD 100546, an object similar in spectral type (B9) but closer to the main sequence, with an age of about 1 x 107 yr. The most obvious change in the profile is the shift in its peak, from 9.7 pm in AB Aur to 11.3 pm in the older object.

Dust radiating in this regime is being heated to a temperature of roughly 500 K. Laboratory studies, coupled with theoretical modeling, find that it is the smaller, and more amorphous, silicates whose emission peaks near 9.7 pm. (The same feature, often seen in absorption, is created by dust in the interstellar medium.) Conversely, larger and more crystalline material has a profile that peaks at a longer wavelength. The particular value of 11.3 pm seen in HD 100546 indicates the presence of the mineral forsterite (Mg2SiO4) and also coincides with a PAH emission feature.

The picture, then, is that the grains, while heated, are nonetheless in a quiescent environment that allows them both to agglomerate and to gradually assume a more ordered structure. This finding is in accord with the inference, based on optical colors, that grains near Herbig Ae/Be

Wavelength X (|xm)

Figure 18.17 Spectra of mid-infrared emission from AB Aur (top panel) and the older star HD 100546 (bottom panel). Note the shift with age of the peak flux.

Wavelength X (|xm)

Figure 18.17 Spectra of mid-infrared emission from AB Aur (top panel) and the older star HD 100546 (bottom panel). Note the shift with age of the peak flux.

stars are larger, on average, than those in the interstellar medium. Finally, we noted earlier that spherical envelope models need the dust opacity to have a relatively weak dependence on wavelength in the millimeter regime to match observed flux values. Enhanced grain size naturally gives this property.

The top panel of Figure 18.18 shows the full spectral energy distribution for HD 100546. It is clear that the object still emits copiously in the infrared. The excess, long-wavelength luminosity is 0.51 times the stellar value of 32 Lq. Indeed, the rise in flux from the near- to mid-infrared is reminiscent of Group II Herbig Ae/Be stars (recall Figure 18.12). In the present case, however, the visual extinction is too low for the star to be embedded within a massive, opaque envelope. Again, a flattened structure is more appropriate. Note that the spectral energy distribution exhibits a strong dip just shortward of 10 ^m. Such a depression, we saw in Chapter 17, suggests truncation of the disk at its inner edge.4

Absence of gas close to the star is not a surprise. Ultraviolet photons from this B9 object can dissociate ambient molecules and ionize the atomic component. The dip in the spectral energy distribution, however, attests to the lack of dust within roughly 10 AU from the stellar surface. There is copious material beyond that distance, as seen through the large infrared excess. It appears, then, that the system of orbiting grains has been somehow cleared, perhaps by infall onto the star. Infall motion is directly seen in the ultraviolet, where gaseous emission lines exhibit redshifted absorption. Curiously, the species displaying this effect are either neutral

4 All three spectral energy distributions in Figure 18.18 also have strong dips shortward of 0.4 |im. This feature is intrinsic to the photosphere and represents absorption by partially ionized hydrogen in the star's outer layers.

Wavelength log X. ((Jim)

Figure 18.18 Spectral energy distributions of intermediate-mass stars close to, or on the main sequence. From top to bottom, the three objects have increasing age and decreasing infrared excess, as measured relative to blackbodies at the appropriate effective temperatures (dashed curves). Note the strong dips in photospheric flux near log A = —0.4. The top curve has been shifted upward by A log (AF\) = +2.5, the bottom downward by A log (A F\) = —4.0

Wavelength log X. ((Jim)

Figure 18.18 Spectral energy distributions of intermediate-mass stars close to, or on the main sequence. From top to bottom, the three objects have increasing age and decreasing infrared excess, as measured relative to blackbodies at the appropriate effective temperatures (dashed curves). Note the strong dips in photospheric flux near log A = —0.4. The top curve has been shifted upward by A log (AF\) = +2.5, the bottom downward by A log (A F\) = —4.0

(e. g., CI) or lightly ionized (S II). Given our previous remarks, the presence of gas moving at high speed, and thus situated relatively close to the star, is unexpected. Perhaps this gas arises through evaporation of infalling, solid material.

The dust component of the disk around HD 100546 is directly observable through scattered optical and near-infrared light. Since the star is relatively luminous and close (100 pc), this structure has been probed in some detail. It can be traced out to 500 AU in radius and has a surface density that rises inward, at least to the (unresolved) truncation edge. Several dark lanes are seen that may be segments of larger, spiral structures. There is also a much lower-density envelope of grains above and below the disk plane, some 1000 AU in radius. This more diffuse material appears as a haze in scattered light.

Another well-studied disk is that associated with the A0 star HR 4796A. The star itself is the brightest member of the TW Hya Association, described earlier in § 17.5. The age of HR 4796A is thus known to be about 1 x 107 yr, close to that of HD 100546. Nevertheless, the excess, long-wavelength luminosity is a good deal less, only 5 x 10~3 of the stellar value. Spectral observations in CO have failed to detect any flux. It appears that most circumstellar gas has been driven off, leaving behind the collection of orbiting grains we have termed a debris disk.

The upper left panel in Figure 18.19 displays an image of this disk in scattered, near-infrared light. To reduce contrast, the star was masked with a coronagraph, as shown. We see, with remarkable clarity, that the circumstellar material in this case is actually confined to a narrow, highly inclined ring. The width of this ring is measured to be 17 AU. Interior to it, i. e., within a radius of some 60 AU outside the star, all grains have been cleared. We are reminded of the interior gap in the debris disk of e Eri, the K dwarf situated in the immediate solar vicinity (Figure 17.27). Another point of similarity is that the observed dust grains in both cases are larger than those generally found in the interstellar medium.

18.4.3 Main-Sequence Disks

Scattered-light images of disks can be striking in their detail, but are available only for a handful of relatively nearby systems. Detection of unresolved, infrared flux from a star has always provided the initial evidence for circumstellar matter. This longer-wavelength radiation represents emission from heated dust grains. Because of their higher luminosities, A stars, such as the zero-age main-sequence object HR 4796A and its older counterparts, have constituted the bulk of the sample. All told, a substantial (but quantitatively uncertain) fraction of field stars are Vega-like, i. e., display excess thermal emission, but have little sign of circumstellar gas.

The most intensively studied debris disk has been that around 3 Pictoris, a Southern hemisphere star lying at a distance of 19 pc. One may assess the age of this 1.8 Mq object by utilizing the fact that it is moving through space with several, lower-mass companions that are still on pre-main-sequence tracks. Through this means, the age appears to be 1.5 x 107 yr, not much older than HR 4796A. The spectral energy distribution in Figure 18.18 shows that there is slightly less infrared excess luminosity, about 3 x 10-3 times the L*-value of 8.5 Lq.

Coronagraphic optical imaging reveals a remarkably narrow structure (Figure 18.19; upper right panel). Apparently, we are viewing the debris disk almost exactly edge-on. Also striking is the manifest asymmetry. While the northeast portion of the disk extends out to 790 AU from the star, the opposite side can only be seen in this image to a radius of 650 AU. The gravitational influence of another body, either a planet or passing star, may have created the distortion.

A wide range of observing facilities, including both spaceborne and groundbased telescopes, have provided many more details. The general picture, in any case, follows the pattern we have already seen. From the thermal, infrared emission, we know that the orbiting grains are relatively large, typically 10 pm in size. There is a steep falloff in this material within some tens of AU from the central star, but smaller grains persist farther inside. A relatively small amount of gas exists, some of it lying in a ring within 1 AU of the star. All of this attention, we note, has been lavished on a structure whose total mass amounts to 10 percent that of the Earth. The interest lies mainly in the generation and erosive depletion of this material, and in the evidence from both processes for more massive, orbiting bodies.

As a star continues to age, its circumstellar dust eventually becomes undetectable through scattered light. However, thermal radiation at submillimeter wavelengths still yields images, as exemplified in the two lower panels of Figure 18.19. On the left is an 850 pm map of Fomalhaut (a Piscis Austrini), an A3 star 8 pc away. All main-sequence objects in the field, including

HR 4796A 0 Pic

HR 4796A 0 Pic

Figure 18.19 Debris disks around four main-sequence A stars. The top two images are in scattered starlight (optical and near-infrared), while the bottom two display submillimeter thermal emission from grains.

this one, have highly uncertain ages. Here, there happens to be a nearby K star with identical proper motion and radial velocity. Assuming the two objects were born together, we may use the fact that the lower-mass companion has a detectable lithium abundance to assign an age to both of 2 x 108 yr. The submillimeter image of Fomalhaut itself shows a ring-like debris disk seen edge-on, i. e., essentially in cross section. A hollow central cavity extends 30 AU in radius. Within the ring are relatively large (10 pm) grains but no observed gas.

The prototype of all objects with excess, long-wavelength emission but little accompanying gas is Vega (a Lyrae), shown in the lower panel of Figure 18.19. The spectral energy distribution, plotted in Figure 18.18, reveals that the actual luminosity contributed by dust has declined to 2 x 10~5 L*. Judging from its position in the HR diagram, the star has already left the zero-age main sequence and has an age of 4 x 108 yr. The 850 pm map in Figure 18.19 shows a disk viewed nearly face-on. In addition to a low-emission gap of radius 60 AU surrounding the star itself, there is also an interior, bright peak that is curiously off-center. Note, however, that this region is marginally resolved in the image.

Table 18.2 summarizes key properties of the debris disk systems we have discussed. Here we have included LIR/L*, the fraction of the stellar luminosity reemitted at long wavelengths, as well as estimates or upper bounds for min, the interior radius of the disk itself. While the age uncertainty is large, there is clearly a tendency for the total mass in orbiting dust to diminish with time. Erosion apparently proceeds faster on the inside, as evidenced by the central holes. What accounts for these trends?

Table 18.2 Nearby Stars with Debris Disks

Name Distance Spectral M* L* L\R/L* min Age

Table 18.2 Nearby Stars with Debris Disks

Name Distance Spectral M* L* L\R/L* min Age

Was this article helpful?

0 0

Post a comment