Ast 777: Star and Planet Formation Jonathan Williams, University of Hawaii
Protoplanetary disks
https://www.eso.org/public/images/eso2008a/ The main phases of star formation
https://www.americanscientist.org/sites/americanscientist.org/files/2005223144527_306.pdf Conventional / historical usage: Protoplanetary disks = Class II
But many planetesimals and perhaps planetary cores form in the earlier embedded stage
https://www.americanscientist.org/sites/americanscientist.org/files/2005223144527_306.pdf First resolved images of disks by HST
https://www.spacetelescope.org/images/opo0113b/ More often, a disk is inferred from its SED
http://www.spitzer.caltech.edu/images/1179-ssc2004-08c-Spectra-Show-Protoplanetary-Disk-Structures But now we can really image them!
https://public.nrao.edu/gallery/hl-tau-birth-of-planets-revealed-in-astonishing-detail-2/ Some theoretical considerations
✓ expect centrifugal radii ~ 100au ✓ viscous accretion disk profile (ok, maybe not so smooth…)
(problem set #3)
What about the temperature profile? Viscous heating
Williams ISM book Radiative heating
same r-3/4 dependence as viscous heating! Williams ISM book Modeling the SED
flux density
dust opacity
see problem set #5
Williams ISM book Modeling the SED… sort of
Disks are much brighter than expected in the MIR/FIR…
Williams ISM book T(r) ⇢ F(λ)
Terrestrial planets Gas and Ice Giants Icy planetesimals
0.1AU 1AU 5AU 100AU ~1000 K ~300 K ~100 K ~20 K NIR MIR FIR mm
Sean Andrews PhD Flared disks
To match the enhanced MIR/FIR fluxes, disks must intercept more starlight at large radii than a flat disk => they flare
Naturally explained by the decreasing gravitational force with radius => increased scale height Flared disks
hydrostatic equilibrium
isothermal eos
—> Flared disks
This fits the SED much better (but the math gets more complicated - see Chiang & Goldreich) Scattered NIR from the surface of IM Lup
https://www.eso.org/public/news/eso1811/ Thermal mm from dust in IM Lup
Same disk, different appearance…
https://www.eso.org/public/news/eso1811/ Huang et al 2018 How do protoplanetary disks compare to what we thought formed the Solar System?
https://www.eso.org/public/news/eso1811/ Ansdell et al 2017 The Minimum Mass Solar Nebula
Ruden 2000, adapted from Weidenschilling 1977 & Hayashi 1981 Measuring dust masses from submillimeter continuum flux
Disks have similar masses as the MMSN
Williams & Cieza 2011 Fitting SED and interferometer visibilities to infer surface density
Andrews et al. 2009 AA49CH03-Williams ARI 14 July 2011 19:21
4.3.1. Surface density. A resolved image of a disk at millimeter wavelengths provides not only a measure of its total mass and radius but also the distribution of mass or surface density. Until p recently this was characterized as a pure power law, ! R− , with values of p generally in the ∝ range 0–1 (Mundy et al. 1996; Lay, Carlstrom & Hills 1997; Wilner et al. 2000; Kitamura et al. 2002; Andrews & Williams 2007b). γ The exponential tapered fits of the form in Equation 4 approximate a power law, ! R− ∝ for R R , but the fitted values of R 30 200 AU correspond to 1 arcsec in all but the # c c $ − ! closest disks, and γ is actually determined largely from the steepness of the exponential taper. The Hughes et al. (2008) comparison of pure power law versus exponentially truncated power- law fits shows similar indices but with slightly steeper pure power-law fits due to the soft edge, p 1.2, γ 0.9forfourdisks. % & = % & = From larger samples, Andrews et al. (2009, 2010b) find a tight range consistent with all having the same value γ 0.9. Using a different modeling technique, however, Isella, Carpenter & % & = Sargent (2009) find a very wide range γ 0.8 to 0.8 with mean γ 0.1 in their data. Negative = − % & = values of γ correspond to decreasing surface densities for R < Rc , which may be an important signature of disk evolution (see Section 6), but these were also found in the smaller disks with
Rc < 100 AU, which are barely resolved and, thus, the hardest to characterize. In general, all results agree that young protoplanetary disks have flatter central density profiles than the canonical power law p 1.5 MMSN (Weidenschilling 1977). There is even more = uncertainty in the density profile of the MMSN than its mass, however, and an exponential tapered power-law fit by Davis (2005) has γ 0.5. The more relevant comparison is of absolute values = in the planet-forming zone. Andrews et al. (2009, 2010b) infer surface densities, ! 10 100 g 2 ≈ − cm− at 20 AU, in their sample of Ophiuchus disks that are in good agreement with the MMSN (Figure 2). Whereas the disk-to–star mass ratio may be very high during the initial stages of formation (Section 3), the Toomre Q parameter, Q(R) c #/πG!,wherec is the sound speed and # the = orbital angularFitting velocity SED (Toomre and 1964), isinterferometer generally much greater than visibilities unity for these Class II YSO disks, implying that theyto are infer gravitationally surface stable atdensity all radii (Isella, Carpenter & Sargent
3 Figure 2 MMSN densities AS 205 Saturn Radial surface Elias 24 density (gas dust) + Uranus GSS 39 profiles for Class II Neptune AS 209 2 YSO disks in DoAr 25
) Ophiuchus based on by INSTITUTE FOR ASTRONOMY on 10/17/11. For personal use only. WaOph 6 –2 VSSG 1 fitting an SR 4 exponentially 1
Annu. Rev. Astro. Astrophys. 2011.49:67-117. Downloaded from www.annualreviews.org SR 13 tapered power-law WSB 52 profile to 880-µm DoAr 33 visibilities and IR WL 18 spectral energy 0 distributions (gas + dust) (g cm ∑ Disks have (Andrews et al. 2009,
log 2010b). The dark Current similar surface gray rectangular –1 resolution densities as regions mark the limits the MMSN minimum mass solar nebula (MMSN) surface densities for –2 10 100 1,000 Saturn, Uranus, and Radius (AU) Neptune.
Andrews et al. 2010
www.annualreviews.org Protoplanetary Disks 77 • Infrared excess ➞ disk lifetime
Disks are common at early times, and almost all gone by ~6 Myr
A compilation of many Spitzer surveys Infrared excess ➞ disk lifetime
fIR ~ e-(t/2.8Myr)
⇒ disk half-life = 2 Myr
But... MIR observations only show the existence of warm dust, and not the amount
A compilation of many Spitzer surveys Disk dust masses decreases very rapidly (faster even than you might think from the infrared disk lifetime)
MJup
MMSN The mass is not going away (planets are common) — its just becoming harder to see because its surface area is decreasing
MJup
MMSN µm mm REPORTS range from 4567.32 T 0.42 My to 4564.71 T thermal history involving multiple heating and substantial radiogenic Pb by the time the last 0.30 My (Fig. 2, B and C, and Table 1). The melting events. The isochron for the oldest melting event occurred at 4564.71 T 0.30 My. oldest chondrule age overlaps with our estimate Allende chondrule (C30) projects back to an ini- Chondrules from the Allende and NWA of CAI formation and thus requires that ag- tial Pb isotopic composition that is less radio- 5697 chondrites define age ranges (Fig. 4) that gregation of the chondrule precursor material genic than the most primitive estimate of the indicate the presence of multiple generations of and its thermal processing occurred within the initial Pb isotopic composition of the solar sys- chondrules in individual chondrite groups. To uncertainty of its Pb-Pb age. Moreover, the age tem (Fig. 3), based on the Nantan iron meteorite explore the spatial significance of this age range, of the oldest chondrule indicates that it was not (14). The low m value of the precursor material we have measured the 54Cr/52Cr ratios of these heated to temperatures above the Pb closure tem- for chondrules in general and the antiquity of chondrules, because 54Cr/52Cr variations within perature after 4567.32 T 0.42 My and therefore this chondrule in particular indicate that the the inner solar system track genetic relationships has a formation and thermal history indistinguish- Pb isotopic composition of the Nantan iron me- between early-formed solids and their respective able from that of CAIs. These data demonstrate teorite does not represent the initial Pb isotopic reservoirs (25). The five chondrules analyzed that chondrule formation started contemporane- composition of the solar system, but instead an here show significant 54Cr variability (Table 1) ously with CAIs (within the uncertainty of our mea- evolved composition inherited after accretion that is not correlated with their ages. Moreover, surements) and continued for at least ~3 My. and differentiation of its parent body before most chondrules have 54Cr/52Cr ratios that are The majority of chondrules are believed to core formation. Similar to chondrule C30, three distinct from those of their host chondrites (26). have formed as dust aggregates of near-solar of the four younger chondrules we analyzed Collectively, these observations indicate that composition subsequently thermally processed define isochron relationships that project back chondrules from individual chondrite groups by a flash heating mechanism creating the ig- to Pb isotopic compositions that are more pri- formed from isotopically diverse precursor ma- neous textures we observe today (10). However, mitive than the current estimate of the solar terial in different regions of the protoplanetary the presence of relict grains, igneous rims, and system initial Pb composition. This implies that disk and were subsequently transported to the ac- compound chondrules suggests that some chon- the precursor material of these chondrules, es- cretion regions of their respective parent bodies. drules may have grown by collisions and re- pecially C3 with its high m value of ~183, were This is consistent with the proposal that radial melting (22, 23). Given the low solar 238U/204Pb not thermally processed until at or near the transport of material in the protoplanetary disk, ratio (m)of~0.15(24), the Pb isotopic com- derived Pb-Pb age. Thus, the range of ages we such as by radial diffusion (27)and/orstellar position of a chondrule precursor is not expected observe for individual chondrules reflects pri- outflows (3), was important during the epoch of to evolve measurably during the lifetime of the mary ages associated with a chondrule-forming CAI and chondrule formation (28). protoplanetary disk (~3 My) until its m value is event and not secondary disturbance of the Some models of chondrule formation such on November 19, 2012 increased by Pb devolatilization during ther- Pb-Pb chronometer. Only the youngest chon- as current sheets (29), colliding molten planet- mal processing. As such, internal isochron rela- drule, C2, yields an isochron that projects to a esimals (30), and recycling of fragmented dif- tionships of chondrules are expected to project more evolved Pb isotopic composition, requiring ferentiated planetesimals (31)arebasedonthe back to nonevolvedConnection initial Pb isotopic compo- that this inclusion wasto thermally cosmochemistry processed for presumed offset of 1 to 2 My between the for- sitions, unless an object experienced a complex the first time early enough to have accumulated mation of CAIs and chondrules and therefore are inconsistent with the contemporaneous for- Fig. 4. Time scales of sol- mation of CAIs and the oldest chondrules in- id formation and disk evo- ferred from our study. Moreover, differentiated lution. The brief formation planetesimals typically have enhanced U/Pb val- www.sciencemag.org interval of 160,000 years ues (32), which would result in chondrules with for the CAI-forming event radiogenically evolved initial Pb isotopic com- is similar to the median positions. However, the initial Pb isotopic com- lifetimes of class 0 proto- position of individual chondrules suggests that, stars of ~0.1 to 0.2 My in most cases, chondrule precursors retained the inferred from astronomical solar U/Pb value up to the chondrule-forming observation of star-forming event(s). regions (37). Therefore, Nebular shock waves are currently the fav- Downloaded from the thermal regime re- ored mechanism for chondrule formation. The pro- quired for CAI condensa- posed sources of shock waves include infalling tion may only have existed clumps of dust and gas (33), bow shocks gen- during the earliest stages erated by planetary embryos (34), spiral arms of disk evolution typified by high mass accretion rates and clumps in a gravitationally unstable proto- −5 –1 planetary disk (35), and x-ray flares (3). Similar (~10 M☉ year )tothe central star. to the colliding planetesimals model, the forma- tion of chondrules by bow shocks requires at least 1 My to allow for the growth of planetary embryos of adequate size and therefore cannot explain the existence of old chondrules. Accretion- driven shock models, including models based on agravitationallyunstabledisk,requirecopiousmass accretion rates to the central star on the order of −5Most of the solid–1 mass is locked ~10 solar mass (M☉)year to be plausible (36).up Astronomical in particles observations greater indicate that suchthan about a high accretionmillimeter rates are achieved in size only in within the deeply 1-2 Myr embedded class 0 phase of star formation (37), and such accretion rates can only last for ~0.1 My. Thus, chondrule formation via accretion-driven
654 2NOVEMBER2012 VOL338 SCIENCE www.sciencemag.org Connelley et al. 2012 Further reading / viewing
Williams & Cieza 2011 Annual Reviews Protoplanetary Disks and their Evolution https://ui.adsabs.harvard.edu/abs/2011ARA%26A..49...67W/abstract In the next lecture, we’ll study how dust grains grow, concentrate, collectively collapse into planetesimals… and ultimately form planets!
Key observational tests lie in high resolution images that reveal rings, gaps, cavities, and spirals https://bulk.cv.nrao.edu/almadata/lp/DSHARP/ Asynchronous lecture
Andrews 2020 Annual Reviews Observations of Protoplanetary Disk Structures
https://arxiv.org/abs/2001.05007
See class website for pdf; I’ll email you discussion questions