* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Chapter 8 Formation of Stars
Cassiopeia (constellation) wikipedia , lookup
International Ultraviolet Explorer wikipedia , lookup
Space Interferometry Mission wikipedia , lookup
Observational astronomy wikipedia , lookup
Rare Earth hypothesis wikipedia , lookup
Perseus (constellation) wikipedia , lookup
Cygnus (constellation) wikipedia , lookup
Extraterrestrial life wikipedia , lookup
History of Solar System formation and evolution hypotheses wikipedia , lookup
Formation and evolution of the Solar System wikipedia , lookup
Definition of planet wikipedia , lookup
Aquarius (constellation) wikipedia , lookup
Corvus (constellation) wikipedia , lookup
Stellar classification wikipedia , lookup
Planetary system wikipedia , lookup
Nebular hypothesis wikipedia , lookup
Type II supernova wikipedia , lookup
Planetary habitability wikipedia , lookup
H II region wikipedia , lookup
Stellar kinematics wikipedia , lookup
Timeline of astronomy wikipedia , lookup
Hayashi track wikipedia , lookup
Chapter 8 Formation of Stars Substantial direct and indirect information indicates that stars are born in the clouds of gas and dust that we call nebulae. • Basics are well understood, many details are not. • We shall have to gloss over various sticky points with assumptions that will be justified by the observation that stars exist and, therefore, something like our assumption must be correct. • Much of this gloss is associated with the general observation that clouds that collapse to form stars have too much kinetic energy and too much angular momentum to produce directly the stars that we see. Since nature makes stars in abundance, this indicates that there exist mechanisms for nascent stars to shed these excess quantities. It is the details of how this happens that we shall circumvent with appeals to observations. 267 268 CHAPTER 8. FORMATION OF STARS 8.1 O and B Associations and T-Tauri Stars • Observation of many hot O and B spectral class stars in and near nebulae is a rather strong indicator that stars are being born there. • These stars are so luminous that they must consume their nuclear fuel at a prodigious rate. • Their time on the main sequence is probably only a million years or so, therefore they cannot be far from their place of birth. • We also see, usually in association with stellar O and B complexes in dust clouds, T-Tauri variables. • These are red irregular variables (spectral class F– M), with a number of unusual characteristics. They exhibit emission lines of hydrogen, Ca+ , and some other metals. 8.1. O AND B ASSOCIATIONS AND T-TAURI STARS 269 Expanding Shel l Unshifted Emission Star Blue-Shifted Emission Red-Shifted Emission Broad Emission Peak To Observer Blue-Shifted Absorption Continuum P Cygni Profile λ Absorption Minimum Figure 8.1: Origin of P Cygni profiles in Doppler shifts associated with expanding gas shells. • The spectral lines for T-Tauri stars often exhibit P Cygni profiles, as illustrated in Fig. 8.1, which indicate the presence of expanding shells of low-density gas around the stars. • They are more luminous than corresponding mainsequence stars, implying that they are larger. • They exhibit strong winds (T-Tauri winds), often with bipolar jet outflows having velocities of 300– 400 km s−1. CHAPTER 8. FORMATION OF STARS 270 Hidden Young Star j et j et Hidden Young Star Wobbling Jet Figure 8.2: Jets and Herbig–Haro objects associated with outflow from young stars near the Orion Nebula. In the top image, the star responsible for the jets is hidden in the dark dust cloud lying in the center of the image. The entire width of this image is about one light year. The Herbig–Haro objects are designated HH-1 and HH-2, and correspond to the nebulosity at the ends of the jets. In the bottom image, a complex jet about a half light year long emerges from a star hidden in a dust cloud near the left edge of the image. The twisted nature of the jet suggests that the star emitting it is wobbling on its rotation axis, perhaps because of interaction with another star. The Herbig-Haro object HH-47 is the nebulosity on the right of the image. It is about 1500 light years away, lying at the edge of the Gum Nebula, which may be an ancient supernova remnant. • Herbig–Haro Objects are often found in the directions of these jets. • Two examples of outflow from young stars and associated Herbig–Haro objects are shown in Fig. 8.2. 8.1. O AND B ASSOCIATIONS AND T-TAURI STARS 271 4 6 8 10 V Giant 12 14 Main Sequence 16 18 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 B-V Figure 8.3: HR diagram for the young open cluster NGC2264. Horizontal bars denote stars with Hα line emission; vertical bars denote variable stars. These considerations indicate that T-Tauri stars are still in the process of contracting to the main sequence. • They are less massive than the O and B stars that often accompany them, so they will have contracted more slowly and many will not yet have had time to reach the main sequence when the more rapidlyevolving O and B stars have done so. • The HR diagram for a young cluster is illustrated in Fig. 8.3, where we see many young stars that have not yet reached the main sequence. • Stars marked with horizontal and vertical bars in this figure have observational properties of T-Tauri stars. 272 CHAPTER 8. FORMATION OF STARS • The bipolar outflows could in principle be explained by an accretion disk around the young T-Tauri stars that would form as a result of conservation of angular momentum for the infalling matter. • Then, if there are strong winds emanating from the star, they would tend to be directed in bipolar flows perpendicular to the plane of the accretion disk. • However, it is difficult to explain the tight collimation of the jets (as good as 10% over one parsec) by such a mechanism, and the source of the energy driving the winds is also not explained by such a simple model. • The Herbig–Haro objects are likely the result of shocks formed when the matter flowing out of the T-Tauri star interacts with clumps of matter, or when clumps of matter ejected from the star interact with low density gas clouds. These observations suggest that we must look to the nebulae to produce the stars. They also suggest that the life of protostars contracting to the main sequence may be more complex (and violent) than the following simple considerations would leave us to believe.) 8.2. CONDITIONS FOR GRAVITATIONAL COLLAPSE 8.2 Conditions for Gravitational Collapse Let us investigate the general question of gravitational collapse to form stars by considering a spherical cloud • composed primarily of hydrogen, • that has a radius R, a mass M, and a uniform temperature T • that consists of N particles of average mass µ . We shall assume that the question of stability is one of competition between gravitation, which would collapse the cloud, and gas pressure, which would expand the cloud. 273 CHAPTER 8. FORMATION OF STARS 274 8.2.1 The Jeans Mass and Jeans Length • The gravitational energy is of the form Ω = −f GM 2 , R where the factor f = 53 if the cloud is spherical and of uniform density, and larger if the density increases toward the center. • We take the thermal energy to be that of an ideal gas, U = 23 NkT. • From the virial theorem, which describes a gravitating gas in equilibrium, we expect that the static condition for gravitational instability is 2U < |Ω|, implying that the system is unstable if it has a mass M with 3/2 3kT 3 1/2 3kT , R= M > MJ ≡ f Gµ mH f Gµ mH 4πρ where N = M/µ mH R = (3M/4πρ )1/3 have been employed. • The quantity MJ = 3kT f Gµ mH 3/2 3 4πρ 1/2 appearing in this condition is termed the Jeans mass. 8.2. CONDITIONS FOR GRAVITATIONAL COLLAPSE 275 • The Jeans mass MJ = 3kT f Gµ mH 3/2 3 4πρ 1/2 defines a critical mass beyond which the system becomes unstable to gravitational contraction. • Since the Jeans mass is proportional to T 3/2 ρ −1/2 , it will be smaller for colder, denser clouds. This makes physical sense: it is easier to collapse a cloud of a given mass gravitationally if the cloud is cold and dense than if it warm and diffuse. • We may also solve the preceding equation for the Jeans length, MJ ≡ 3kT R f Gµ mH → RJ = f Gµ mH MJ. 3kT • This defines the characteristic length scale associated with the Jeans mass and thus characterizes the minimum size of gravitationally unstable regions. CHAPTER 8. FORMATION OF STARS 276 8.2.2 The Jeans Density • It is often more useful to express the Jeans criterion in terms of a critical density for gravitational collapse (the Jeans density) 3/2 3 3 1/2 3kT 3 3kT → ρJ = . MJ = f Gµ mH 4πρ 4π M 2 f µ mH G • Notice that the critical density is lowest (and thus more easily achieved) if the mass is large and the temperature low, as we would expect on intuitive grounds. Example: Consider a cold cloud of molecular hydrogen at T = 20 K that has a mass of 1000 M⊙ ; the associated Jeans density is only ρJ = 10−22 g cm−3. On the other hand, a molecular hydrogen cloud at the same temperature but containing only 1 solar mass has a Jeans density that is 6 orders of magnitude larger. • The Jeans criterion is simple because it is a static condition that says nothing about gas dynamics and it neglects potentially important factors influencing stability such as magnetic fields, dust formation and vaporization, and radiation transport. • Nevertheless, the Jeans criterion is an extremely useful starting point for understanding how stars form from clouds of gas and dust that become gravitationally unstable. 8.3. FRAGMENTATION OF COLLAPSING CLOUDS 277 Figure 8.4: Fragmentation into gravitationally unstable subclouds. 8.3 Fragmentation of Collapsing Clouds From the foregoing collapse of more massive clouds is favored, but most stars contain less than 1M⊙ of material. • The solution to this dilemma is thought to lie in fragmentation, as illustrated in Fig. 8.4. • As we shall see, the initial collapse is expected to occur at almost constant temperature. Therefore, from MJ = 3kT f Gµ mH 3/2 3 4πρ 1/2 the Jeans mass decreases in the initial collapse • We speak loosely: The Jeans criterion assumes a cloud near equilibrium, not one already collapsing. • Hence as large clouds, which have the smallest Jeans density, begin to collapse their average density increases; • At some point subregions of the original cloud may exceed the critical density and become unstable in their own right toward collapse. 278 CHAPTER 8. FORMATION OF STARS • If there are sufficient perturbations present in the cloud, these subregions may separate and pursue independent collapse. • Within these subclouds the same sequence may be repeated: as the density increases, subregions may themselves become gravitationally unstable and begin an independent collapse. • By such a hierarchy of fragmentations, it is plausible that clusters of protostars might be formed that have individual masses comparable to that of observed stars 8.4. STABILITY IN ADIABATIC APPROXIMATION 8.4 Stability in Adiabatic Approximation To understand further the behavior of gravitationally unstable clouds, let us consider the adiabatic contraction (or expansion) of a homogenous cloud. • Real clouds will exchange energy with their surroundings and so are not completely adiabatic. • However the results obtained in this limit will often be instructive in understanding more realistic situations. 279 CHAPTER 8. FORMATION OF STARS 280 (a) (b) T ~ ρ2/3 (γ = 5/3) Expansion Collapse Halts 1 /3 T ~ρ e ps la l o C log T log T Expansion e ps olla C s ou nu i t n Co 1 /3 T ~ρ Contraction Contraction T~ ρ1/3 (γ = 4/3) log ρ log ρ Figure 8.5: Gravitational equilibrium in temperature–density space. • From the Jeans density 3 ρJ = 4π M 2 3kT f µ mH G 3 . equilibration of gravity and pressure requires the temperature T and density ρ be related by T ∝ ρ 1/3 . • In Fig. 8.5, this divides the T –ρ plane into – A region above the line T ∼ ρ 1/3 where the system is unstable toward expansion, and – a region below the line where the system is unstable toward contraction. 8.4. STABILITY IN ADIABATIC APPROXIMATION (a) (b) T ~ ρ2/3 (γ = 5/3) Expansion Collapse Halts 1 /3 ~ρ e ps la ol C log T log T Expansion T 281 1 /3 T ~ρ e ps olla C us uo tin n Co Contraction Contraction T ~ ρ1/3 (γ = 4/3) log ρ log ρ • For points above the stability line (in the unshaded area), pressure forces are larger than gravitational forces and the system is unstable to expansion. • For points below the stability line (in the shaded area), pressure forces are weaker than gravitational forces and the system is unstable with respect to contraction. CHAPTER 8. FORMATION OF STARS 282 (a) (b) T ~ ρ2/3 (γ = 5/3) Expansion Collapse Halts 1 /3 T log T log T Expansion ~ρ e ps la l o C 1 /3 T ~ρ se llap o sC ou nu i t n Co Contraction Contraction T~ ρ1/3 log ρ (γ = 4/3) log ρ 8.4.1 Dependence on Adiabatic Exponents • First consider a monatomic ideal gas, for which the adiabatic exponent is γ = 35 . • Since ρ ∝ V −1 and an adiabatic equation of state is TV γ −1 = constant, T ∝ ρ γ −1 → T (γ = 35 ) ∼ ρ 2/3 . • This corresponds to the dashed line in the left figure above, which is steeper than the equilibrium line and therefore crosses it. 8.4. STABILITY IN ADIABATIC APPROXIMATION (a) (b) T ~ ρ2/3 (γ = 5/3) Expansion Collapse Halts 1 /3 ~ρ e ps la l o C log T log T Expansion T 283 1 /3 T ~ρ se llap o sC ou nu i t n Co Contraction Contraction T ~ ρ1/3 (γ = 4/3) log ρ log ρ • A cloud that is unstable to gravitational contraction (corresponding to a point on the dashed line in the shaded area of the left figure) will follow the dashed line to the right as it collapses, as indicated by the arrow (right is increasing density). • But in this case the collapse will be halted at the point where the dashed line reaches the stability line (point labeled “Collapse Halts”). • Likewise, a cloud unstable to expansion (corresponding to a point on the dashed line lying in the unshaded area of the left figure) will follow the dashed line to the left as it expands (left is decreasing density). • This expansion halts at the stability line. • Thus, γ = 35 is gravitationally stable. CHAPTER 8. FORMATION OF STARS 284 (a) (b) T ~ ρ2/3 (γ = 5/3) Expansion Collapse Halts 1 /3 T ~ρ e ps la l o C log T log T Expansion 1 /3 T ~ρ se llap o sC ou nu i t n Co Contraction Contraction T~ ρ1/3 log ρ (γ = 4/3) log ρ • Now consider the right figure above, where we assume that the cloud has an adiabatic exponent γ = 34 . • In this case, the contraction (or expansion) follows an adiabat for which T ∝ ρ γ −1 ∝ ρ 1/3 . • Since this adiabat is parallel to the stability line, the two lines never cross and a system lying on the dashed line collapses and continues to collapse adiabatically. • This will also be the case if γ < 34 . • Likewise, a system with γ = 43 that is above the stability line will continue to expand adiabatically as long as γ = 43 . • γ ≤ 34 is gravitationally unstable. 8.4. STABILITY IN ADIABATIC APPROXIMATION 8.4.2 Physical Interpretation Physical meaning of the preceding discussion: • A gas is less able to generate the pressure differences required to resist gravity if the energy released by gravitational contraction can be absorbed into internal degrees of freedom. • This energy is not available to increase the kinetic energy of the gas particles. • The parameter γ is relevant because it is related to the heat capacities for the gas (γ = CP /CV ). • Typical sinks of energy that can siphon off energy internally are 1. Rotations of molecules 2. Vibrations of molecules 3. Ionization 4. Molecular dissociation. Such internal degrees of freedom are energy sinks that lower the resistance of the gas to gravitational compression. 285 CHAPTER 8. FORMATION OF STARS 286 • In the large clouds of gas and dust that are candidates for stellar birthplaces, γ can be reduced to a value of 43 = 1.33 or less by 1. The presence of polyatomic gases with more than five degrees of freedom, since 1 + s/2 s/2 γ= → γ (s = 5) = 1 + 5/2 7 = = 1.4, 5/2 5 2. By ionization of hydrogen around 10,000 K 3. By the dissociation of hydrogen molecules around 4,000 K. • The large molecules required for the first situation are relatively rare in the interstellar medium but their presence enhances the chance of gravitational collapse for a cloud. • In hydrogen ionization or molecular dissociation zones, – Typically γ ≤ 43 and this causes an instability until the ionization or dissociation is complete. – Then γ will return to normal values (γ ≃ 35 ) and collapse on the corresponding adiabat will reach the equilibrium line and stabilize the collapse (recall the following figure) (a) (b) T ~ ρ2/3 (γ = 5/3) Expansion Collapse Halts 1 /3 T ~ρ e ps la ol C log T log T Expansion 1 /3 T ~ρ e ps olla C s ou nu nti o C Contraction Contraction T ~ ρ1/3 (γ = 4/3) log ρ log ρ 8.5. THE COLLAPSE OF A PROTOSTAR 8.5 The Collapse of a Protostar The preceding introduction sweeps much under the rug, but we shall assume that the existence of stars implies that protostars form by some mechanism similar to the one outlined above and follow the consequences of the gravitational collapse of such a protostar. • Let us consider the collapse of a one solar mass protostar. • From the Jeans criterion, ρJ ≃ 10−16 g cm−3 for T = 20 K and M = 1M⊙. • Thus, we expect that a 1 M⊙ cloud can collapse if this average density is exceeded. • The size of this initial cloud may be estimated by assuming the density to be constant and distributed spherically, implying that R ∼ 1.6 × 1016 cm ∼ 1000 AU. • Thus, the initial protostar has a radius approximately 25 times that of the present Solar System. 287 288 CHAPTER 8. FORMATION OF STARS 8.5.1 Initial Free-Fall Collapse • We may assume that the initial collapse is free-fall and isothermal, as long as the gravitational energy released is not converted into thermal motion of the gas and thereby into pressure. • This will be the case as long as the energy not radiated away is largely taken up by 1. dissociation of hydrogen molecules into hydrogen atoms 2. ionization of the hydrogen atoms. • The dissociation energy for hydrogen molecules is εd = 4.5 eV • The ionization energy for hydrogen atoms is εion = 13.6 eV. • The energy required to dissociate and ionize all the hydrogen in the original cloud is then E = N(H2 )εd + N(H)εion M M εd + εion , = 2mH mH where N denotes the number of the corresponding species and mH is the mass of a hydrogen atom. • For the case of a protostar of one solar mass, the requisite energy is approximately 3 × 1046 erg. 8.5. THE COLLAPSE OF A PROTOSTAR 289 • If the dissociation and ionization energy M M E= εd + εion , 2mH mH is supplied by gravitational contraction from radius R1 to R2, GM 2 GM 2 − R 2 | {z R1 } gravity = M M εd + εion |2mH {z mH } . dissociation and ionization • Solve for R2 assuming M = 1M⊙ and R1 = 1016 cm to give R2 = 1013 cm ≃ 100R⊙ ≃ 21 AU. • The corresponding timescale is (see Exercise 4.1) p tff = 3π /32Gρ ≃ 7000 yr. • Therefore, a solar-mass protostar collapses from about 25 times the radius of the Solar System to about half the radius of the Earth’s orbit in near free-fall on a timescale of about 7,000 years. • The collapse then slows because 1. All the hydrogen has been dissociated and ionized, 2. the photon mean free path becomes short and the cloud becomes opaque to its own radiation, 3. temperature increases as heat begins to be trapped, and 4. the resulting pressure gradients counteract the gravitational force and bring the system into near hydrostatic equilibrium. Thus, we may apply the virial theorem in near adiabatic conditions from this point onward. CHAPTER 8. FORMATION OF STARS 290 8.5.2 Homology From the expressions for free-fall collapse we see that the characteristic timescale for free fall is independent of the radius of the collapsing mass distribution. • This behavior is termed homologous collapse. • One consequence of homologous collapse is that if the initial density is uniform it remains uniform for the entire collapse. • Because successive configurations in homologous processes are self-similar (related by a scale transformation), homologous systems are particularly simple to deal with mathematically. • Therefore, reasonably good approximate treatments of the initial phases of gravitational collapse are often possible by making homology assumptions. 8.5. THE COLLAPSE OF A PROTOSTAR Adiabatic log T H ionization H 2 dissociation Thermonuclear ignition Isothermal 291 Free fall log ρ Figure 8.6: Schematic track in density and temperature for the collapse of a gas cloud to form a star. 8.5.3 A More Realistic Picture The preceding picture is oversimplified but contains essential features supported by more sophisticated considerations. A more realistic variation of temperature and density for star formation is illustrated in Fig. 8.6. • In this more realistic picture the cloud begins to heat and deviate from free-fall behavior once it traps significant heat. • When the temperature is sufficient to dissociate and then ionize hydrogen, the cloud again collapses ∼ isothermally for a time. • Once all hydrogen has been dissociated and ionized, the collapse returns to one governed by approximately adiabatic conditions in near hydrostatic equilibrium. CHAPTER 8. FORMATION OF STARS 292 8.6 Onset of Hydrostatic Equilibrium The temperature at which hydrostatic equilibrium sets in may be estimated as follows. • From the virial theorem we have that 2U + Ω = 0. • The gravitational energy Ω is M M GM 2 =− Ω ≡ Ω(R2 ) = − εd + εion , R2 2mH mH where Ω(R1 ) has been neglected compared with Ω(R2 ). • From U = 23 NkT , the internal energy for the hydrogen ions and electrons in the fully ionized gas is approximately U ≃ 23 (NH + Ne)kT = 3NH kT = 3M kT, mH where NH is the number of hydrogen ions and Ne ∼ NH is the number of free electrons. • Therefore, the virial theorem requires that 2U + Ω = M M 6M kT − εd − εion = 0, mH 2mH mH and solving this for T gives 1 εd εion 2.6 eV ≃ + ≃ 30, 000 K T= k 12 6 k for the onset of hydrostatic equilibrium. 8.6. ONSET OF HYDROSTATIC EQUILIBRIUM 10 3 10 4 10 6 293 10 2 100 M Luminosity (Solar Units) 10 4 10 5 10 4 10 M 10 2 10 10 3 6 3M 10 5 10 10 0 10 ZAMS 10 5 10 5 6 1M 10 7 0.5 M −2 0.1 M 63,000 25,000 10,000 4000 10 6 10 8 1600 Surface Temperature (K) Figure 8.7: Evolutionary tracks for collapse to the main sequence. Numbers on tracks are times in years. • Subsequent contraction of the protostar is in near hydrostatic equilibrium and is controlled by the opacities, which govern how fast energy can be brought to the surface and radiated. • As we have discussed in §4.6, this too is a consequence of the virial theorem and leads to the Kelvin–Helmholtz timescale of about 107 years for a star of one solar mass. • The evolutionary tracks for protostars of various masses to collapse to the main sequence are shown in Fig. 8.7. 294 CHAPTER 8. FORMATION OF STARS 8.7 Termination of Fragmentation • Earlier we indicated that gravitational collapse of large clouds is likely to fragment into a hierarchy of sub-collapses, and we invoked this argument to explain why we observe many lowmass stars. • However, this argument is incomplete (even leaving aside that we gave no quantitative justification for it), because we must ask the question of what stops the fragmentation in the vicinity of 0.1–1 solar masses. • The most plausible answer is that the transition from isothermal to adiabatic collapse implies a modification of the Jeans criterion and that this dictates a lower limit for the mass of the fragments produced by a hierarchical collapse. • Substitution of the condition T ∝ ρ γ −1 for adiabats in 3/2 3kT 3 1/2 MJ = , f Gµ mH 4πρ implies that for adiabatic clouds MJ ≃ ρ (3γ −4)/2 . For γ = 5/3 then, the Jeans mass is proportional to ρ 1/2 and in adiabatic collapse the Jeans mass increases. • This implies that the transition from isothermal to adiabatic collapse sets a lower bound on possible Jeans masses. Plausible calculations based on this idea show that the fragmentation is likely to end in the vicinity of one solar mass, in rough agreement with observations. 8.8. HYASHI TRACKS 295 Hyashi Track Log Luminosity Star fully convective Radiative core develops Main Sequence Hydrogen fusion begins Region forbidden to stars of mass M and composition c (M,c) Log T Figure 8.8: Evolution of protostars to the main sequence. 8.8 Hyashi Tracks A more detailed understanding of the collapse to the main sequence follows from a fundamental result first obtained by Hyashi: A star generally cannot reach hydrostatic equilibrium if its surface is too cool. • This implies that there exists a region in the righthand portion of the HR diagram that is forbidden to a given star while it is in hydrostatic equilibrium. • This region, the boundaries of which depend on the mass and composition of the star, is called the Hyashi forbidden zone; it is illustrated in Fig. 8.8 for a star of mass M and composition c. CHAPTER 8. FORMATION OF STARS 296 8.8.1 Fully Convective Stars Stars contracting to the main sequence • Must have large surface areas and (relatively) high surface temperatures, so they have large luminosities. • Furthermore, once hydrogen is ionized they have high opacities. • The combination of high opacity with large luminosity ensures that the temperature gradients necessary to transport the energy radiantly exceed the adiabatic one. • Thus such collapsing stars are expected to be almost completely convective . • Recall from Ch. 3 that completely convective stars are approximately described by a γ = 35 polytrope: For a completely ionized star, fully mixed by convection with negligible radiation pressure, P = K ρ γ (r) = K ρ 1+1/n (r) = K ρ 5/3, if γ = 35 , where the polytropic index n is parameterized by n = 1/(γ − 1), and where K is constant for a given star. 8.8. HYASHI TRACKS 297 Hyashi Track Log Luminosity Star fully convective Radiative core develops Main Sequence Hydrogen fusion begins Region forbidden to stars of mass M and composition c (M,c) Log T By examining fully convective stars with a thin radiative envelope, Hyashi showed that • Contracting protostars follow an almost vertical HR trajectory called the Hyashi track for the star. • If the star is fully convective, the Hyashi track is essentially defined by the left boundary of the Hyashi forbidden zone, as illustrated in the above figure. • The forbidden zone can be understood simply: – A star on its Hyashi track is fully convective and radiates as a blackbody. – Thus, no other configuration could lose energy more efficiently in hydrodynamic equilibrium. – Thus no such stars can exist to the right of the Hyashi track in the HR diagram (which corresponds to lower surface temperatures). 298 CHAPTER 8. FORMATION OF STARS 8.8.2 Development of a Radiative Core As the collapsing star descends the Hyashi track its central temperature is increased by the gravitational contraction. • This decreases the central opacity (recall that for the representative Kramers opacity, κ ∼ ρ T −3.5 ). • Eventually this lowers the temperature gradient in the central region sufficiently that it drops below the critical value for convective stability. • A radiative core develops. • As contraction proceeds the radiative core begins to grow at the expense of the convective regions, which are eventually pushed out to the final subsurface regions characteristic of stars like the Sun. • (In more massive stars the subsurface convective zones are eliminated completely but the core may become convective if the power generation is sufficiently large after the star enters the main sequence.) 8.8. HYASHI TRACKS 299 Hyashi Track Log Luminosity Star fully convective Radiative core develops Main Sequence Hydrogen fusion begins Region forbidden to stars of mass M and composition c (M,c) Log T • While the fully convective star is on the Hyashi track, its luminosity decreases rapidly because of shrinking surface area. • However, as the opacity decreases over more and more of the interior because of the increasing temperature, the luminosity begins to rise again because more energy can flow out radiantly. • Since at this point the star is shrinking as its luminosity increases, its surface temperature must increase and the star begins to follow a track to the left and somewhat upward in the HR diagram (above figure). • Finally, the onset of hydrogen fusion causes the track to bend over and enter the main sequence, as also illustrated in the above figures. CHAPTER 8. FORMATION OF STARS 300 Hyashi Track Log Luminosity Star fully convective Radiative core develops Main Sequence Hydrogen fusion begins Region forbidden to stars of mass M and composition c (M,c) Log T Thus, the contraction to the vicinity of the main sequence is composed of two basic periods: 1. A vertical descent in the HR diagram for fully convective stars, followed by 2. a drift up and to the left as the interior of the star becomes increasingly radiative at the expense of the convective envelope. 8.8. HYASHI TRACKS (b) Reduced metals Increasing mass log L (a) 301 1M (c) Ma in s Normal log T equ enc log T 0.1 M 20 M e log T Figure 8.9: Dependence of Hyashi tracks on (a) composition and (b) mass. The solid portions of each curve in (c) represent the descent on the Hyashi track. 8.8.3 Dependence of Track on Composition and Mass Hyashi tracks depend weakly on the mass and composition, as illustrated in Fig. 8.9. • For more massive stars of fixed composition the Hyashi tracks are almost parallel to each other but increasingly shifted to the left in the HR diagram (Fig. 8.9b). • The Hyashi tracks also depend on stellar composition, because this can influence the opacity. • For example, a metal-poor star of a given mass will generally have a Hyashi track to the left of an equivalent metal-rich star because of lower opacity (Fig. 8.9a). CHAPTER 8. FORMATION OF STARS 302 (b) Reduced metals 1M (c) Increasing mass Ma in s Normal equ enc log T 0.1 M 20 M log L (a) log T e log T • Finally, the transition from convective to radiative interiors, and the corresponding transition from downward motion to more horizontal leftward drift on the HR diagram, is generally faster in more massive stars because of more rapid interior heating. • Therefore, as illustrated in Fig. (c) above and the following, 10 10 3 10 4 6 10 2 100 M Luminosity (Solar Units) 10 4 10 5 10 4 10 M 10 2 10 0 10 10 3 6 3M 10 5 10 10 ZAMS 5 10 5 6 1M 10 7 0.5 M 10 −2 0.1 M 63,000 25,000 10,000 4000 10 6 10 8 1600 Surface Temperature (K) massive stars leave the Hyashi track quickly and approach the main sequence almost horizontally . • Conversely, the least massive stars never leave the Hyashi track as they collapse and drop almost vertically to the main sequence They are thought to remain completely convective, even after entering the main sequence. 8.9. LIMITING LOWER MASS FOR STARS 303 8.9 Limiting Lower Mass for Stars A contracting protostar will become a star only if the temperature increases sufficiently in the core to initiate thermonuclear reactions. • The results of Exercise 4.7 show that, for an idealized star composed of a monatomic ideal gas having uniform temperature and density, the temperature varies with the cube root of the density: 2/3 M T = 4.09 × 106 µ ρ 1/3 . M⊙ • However, this behavior assumes an ideal classical gas; the temperature will no longer increase with contraction if the equation of state becomes that of a degenerate gas. • In Exercise 4.7, the critical temperature and density for onset of electron degeneracy was estimated by setting kT equal to the fermi energy; we found that the critical density is defined by ρ ≃ 6 × 10−9 µeT 3/2 g cm−3. • Inserting this into the preceding equation gives for the temperature at which the critical density is reached in the contracting protostar, M 2/3 1/3 7 . T ≃ 5.6 × 10 µ µe M⊙ CHAPTER 8. FORMATION OF STARS 304 • For M ∼ M⊙ and µ µe1/3 ∼ 1, we obtain from 2/3 M . T ≃ 5.6 × 107 µ µe1/3 M⊙ that T ∼ 107 K. • Thus, a solar mass protostar can produce an average temperature of 10 million K by contraction, which is more than enough to ignite hydrogen fusion before the electrons in the core become degenerate and stop the growth of temperature with pressure. • On the other hand, as the mass of the protostar is decreased we will eventually reach a mass where the core will become degenerate before the temperature rises to the required values to support hydrogen fusion. • Detailed calculations indicate that this limiting mass is approximately 0.08M⊙ − 0.10M⊙. • What of collapsing clouds with less than this critical mass required to form stars? • For them the growth in temperature is halted by electron degeneracy pressure before fusion reactions can begin and no star is formed. • It is speculated that many such brown dwarfs may exist in the Universe, supported hydrostatically by electron degeneracy pressure and radiating energy left over from gravitational contraction. 8.9. LIMITING LOWER MASS FOR STARS 305 It is useful to tie together several threads of discussion involving hydrostatic equilibrium, the virial theorem, stellar energy production, and gravitational collapse to form stars by asking a simple question: “ Why are stars hot?” • The popular perception is that stellar cores are hot because they correspond to enormous thermonuclear furnaces. • But the core of the Sun is in fact a very low density power source (Exercise 5.18: a few hundred watts per cubic meter in the core). • We have seen further in this chapter that it is the release of gravitational energy by contraction and the partial trapping of that energy by high stellar opacity that raises the protostar interior to the temperature (and density) required for hydrogen fusion. • The role of hydrogen fusion is not to heat stars (gravity and the virial theorem can do that nicely); it is to sustain the luminosity over much longer periods than would be possible otherwise. • Triggering thermonuclear reactions replaces the Kelvin–Helmholtz timescale for sustained luminosity from gravitational contraction with the much longer nuclear burning timescale. • This enables the Sun to radiate its present power for billions of years rather than millions of years. • So the source of sustained luminosity and sustained high interior temperatures for main sequence stars is indeed hydrogen fusion, but the cores of those stars were heated by gravitational contraction, not fusion, to their present temperatures. • They maintain those high temperatures and luminosities by slow fusion reactions under conditions of hydrostatic equilibrium. CHAPTER 8. FORMATION OF STARS 306 8.10 Brown Dwarfs Brown dwarfs collapse out of hydrogen clouds, not out of protoplanetary disks (like stars). • But they radiate energy only by gravitational contraction, not from hydrogen fusion (like planets). • Their masses are expected to range from several times the mass of Jupiter to a few percent of the Sun’s mass. • The cooler brown dwarfs may resemble gas giant planets in chemical composition, • Hotter ones may begin to look chemically more like stars. • They are difficult to detect since they are small and of very low luminosity. 8.10. BROWN DWARFS 8.10.1 Spectroscopic Signatures The first brown dwarf discovered was Gliese 229B. • Gliese 229B appears to be too hot and massive to be a planet, but too small and cool to be a star. • The IR spectrum of GL229B looks like the spectrum of a gas giant planet. • Most telling is evidence of methane gas, common in gas giants but not found in stars because methane can survive only in atmospheres having temperatures lower than about 1500 K. 307 308 CHAPTER 8. FORMATION OF STARS In addition to searching for gases like methane that should not be present in stars, searches for brown dwarfs have also looked for evidence of the element lithium. • Hydrogen fusion destroys lithium in stars. • At temperatures above about 2 × 106 K, a proton encountering a lithium nucleus has a high probability to react with it, converting the lithium to helium. • The amount of lithium that can survive is a function of how strongly the material of the star is mixed down to the core fusion region by convection. • Protostars are convective, so stars start off with a strongly mixed interior, but the initial core temperature in the protostar is not high enough to burn lithium. • The lightest stars (red dwarfs) remain convective once on the main sequence, so lithium is mixed down to the fusion region and destroyed in red dwarfs. • Because these stars are cool, it takes some time to burn the lithium, but calculations indicate that lithium could survive no longer than about 2 × 108 years in the lightest true star. 8.10. BROWN DWARFS No Lithium 309 Molecular Hydrogen and Helium Lithium, Methane Fully c onvective Solid metallic core Thermonuclear reactions Red Dwarf No thermonuclear reactions Brown Dwarf Gas Giant Figure 8.10: Contrasting interiors of a red dwarf, a brown dwarf, and a gas giant planet. Generally stars initiate thermonuclear reactions but brown dwarfs and planets do not. Thus, lithium is destroyed in stars. The presence of methane is also an indication that the temperatures are too low for the object to be a star. Gas giant planets can also contain lithium and methane, but their upper interiors tend to be dominated by molecular hydrogen and helium. The basic interior structures expected for stars, brown dwarfs, and gas giant planets are summarized schematically in Fig. 8.10. CHAPTER 8. FORMATION OF STARS 310 Sun Gliese 229A Teide 1 Gliese 229B Jupiter 5800 K 3800 K 2700 K 900 K 180 K G2 Star Red Dwarf Brown Dwarf Planet Figure 8.11: Size and surface temperature trends for stars, brown dwarfs, and gas giant planets. 8.10.2 Stars, Brown Dwarfs, and Planets Figure 8.11 summarizes the size and surface temperature trend from middle main sequence stars like the Sun, through the lowest mass stars (red dwarfs), through brown dwarfs, and finally to planets. • Brown dwarfs can have surface temperatures comparable to that of the lowest mass stars, but atmospheric compositions similar to large planets. • The challenge is to distinguish brown dwarfs from stars and gas giants at interstellar distances. • A number of brown dwarf candidates have now been identified but in many cases there is uncertainty about whether they are brown dwarf companions of stars, or giant planets orbiting stars. 8.11. LIMITING UPPER MASS FOR STARS 8.11 Limiting Upper Mass for Stars As we have discussed in the preceding section, a limiting lower mass for stars is set by the requirement that sufficient temperature be generated by gravitational collapse to commence the fusion of hydrogen to helium in the core. • An upper limiting mass for stars is thought to exist because of the opposite extreme: if the star is too massive, the intensity of the energy production makes the star unstable to disruption by the radiation pressure. • The pressure associated with the radiation grows as the fourth power of the temperature and thus will be most important for very hot stars. • It is instructive to ask what photon luminosity is required such that the magnitude of the force associated with the radiation field is equivalent to the magnitude of the gravitational force. • This critical luminosity, which defines limiting configurations that are stable gravitationally with respect to the pressure of the photon flux, is termed the Eddington luminosity. 311 CHAPTER 8. FORMATION OF STARS 312 8.11.1 Eddington Luminosity • As shown in Exercise 8.1, the force per unit volume associated with a photon gas is given by the gradient of the radiation pressure, 1 dPr 4 3 dT Fr = − = aT . V dr 3 dr • If the magnitude of this force is equated to the magnitude of the gravitational force, we obtain an expression for the Eddington luminosity LEdd = 4π cGM . κ • We may expect that stars exceeding this luminosity can blow off surface layers by radiation pressure. • The ejection of material may also be aided by stellar pulsations that result from pressure instabilities at high luminosity, and may be influenced by the presence of stellar rotation and magnetic fields. • Both the force from the radiation pressure and that from the gravitational field are proportional to ρ /r2, so the dependence on r and ρ cancels from the Eddington luminosity and the critical luminosity is determined completely by the mass of the star and an appropriate opacity for regions near the surface. 8.11. LIMITING UPPER MASS FOR STARS 8.11.2 Estimate of Limiting Mass The Eddington luminosity may be expressed as M LEdd ≃ 3.5 × 104 , L⊙ M⊙ if we estimate the opacity κ by the Thomson formula. • We may use this equation to estimate a radiationpressure mass limit by – assuming that the most luminous stars observed (L ∼ several × 106L⊙ ) are at the Eddington limit, and – approximating the relevant opacity by the Thomson formula. • As shown in Exercise 8.1, this suggests a maximum stable mass of order 100M⊙. This is a very crude estimate but calculations, and observations, suggest that the most massive stars indeed have masses of this order. 313 314 CHAPTER 8. FORMATION OF STARS 8.11.3 Envelope Loss from Massive Young Stars As we discuss in the next section, there is strong observational evidence that very massive stars eject large amounts of material from their envelopes early in their lives. Radiation pressure and pulsational instabilities are thought to play a leading role in these mass-loss processes. 8.11. LIMITING UPPER MASS FOR STARS Massive stars may go through early stages where they expel large portions of their envelopes into space at velocities as large as 1000 km s−1. • In such stars, the timescale for mass loss τloss ∼ M/Ṁ, where M is the mass and Ṁ the rate of mass loss, may be shorter than their main sequence timescales. • One class of stars exhibiting large mass loss while on the main sequence is that of Wolf–Rayet stars, which are main sequence stars characterized by 1. large luminosity, 2. envelopes strongly depleted in hydrogen, and 3. high rates of mass loss. • Most Wolf–Rayet stars have masses of 5 − 10M⊙. • They are thought to be the remains of stars initially more massive than 30M⊙ that have ejected all or most of their outer envelope, exposing the hot helium core (as a result, they are very strong UV emitters). • The envelopes of Wolf–Rayet stars typically contain 10% or less hydrogen by mass, with individual stars exhibiting different levels of envelope stripping. 315 CHAPTER 8. FORMATION OF STARS 316 (a) (b) HD56925 η Carinae Figure 8.12: (a) Wolf–Rayet star HD56925 surrounded by remnants of its former envelope. (b) η Carinae, surrounded by ejected material. • Figure 8.12a shows the nebula N2359, a wind-blown shell of gas that has been expelled from the Wolf– Rayet star HD56925 (marked by the arrow.) • The nebula contains shock waves generated by interaction of the wind and interstellar medium, and is glowing from excitation of expelled material. • Figure 8.12b shows an extreme example of mass loss: the supermassive, highly unstable star, η Carinae. • Elemental abundances in the nebula around η Carinae are consistent with this being the supergiant phase of a 120 M⊙ star that has evolved with very large mass loss on the main sequence and afterwards. • (Some evidence suggests that η Carinae may be a binary star, which would complicate interpretation). 8.12. PROTOPLANETARY DISKS 317 Figure 8.13: Schematic model of an accretion disk and bipolar outflow. 8.12 Protoplanetary Disks • In the final stages of protostar collapse, we may expect that because of conservation of angular momentum the matter from the nebula will continue to accrete onto the star from an equatorial accretion disk. • There is also strong empirical evidence that young stars produce very strong stellar winds that are focused perpendicular to the equatorial accretion disk. • Thus, accretion disks and strong bipolar outflow may be a rather common feature of stars collapsing to the main sequence. Figure 8.13 illustrates. • The strong wind blowing from young stars is not completely understood. One idea is that it is caused by the matter drawing a magnetic field inward as it falls into the accretion disk. • The outward flowing wind is partially blocked by the accretion disk around the equator of the forming star and so escapes along the polar axis, thus producing bipolar outflows from the young star. CHAPTER 8. FORMATION OF STARS 318 Solar System Beta Pictoris Vega Fomalhaut Solar System Solar System Figure 8.14: Radio frequency maps indicating the presence of dusk disks around three young stars. • However, this simple picture cannot explain the evidence for the rather narrow width of the outflow observed in some cases. Presumably the full mechanism is more complex, perhaps involving the effect of magnetic fields to focus the ejected material. Figure 8.14 shows submillimeter radio frequency observations of three nearby, young stars. • The maps are false-color illustrations of the intensity of radio emission, with the emission thought to originate primarily from dust in the vicinity of the stars. • The location of the star is indicated by the black star symbol. • The size of our Solar System is illustrated by a vertical bar. • In all three cases, there is strong evidence for dust disks around these young stars. 8.13. EXTRASOLAR PLANETS 8.13 Extrasolar Planets The dust disks observed around a number of young stars suggests that planetary formation may be taking place in these systems. • Indeed, over the past few years impressive evidence has accumulated for ∼1000 extrasolar planets. • These planets are difficult to observe directly at their great distance. They are detected primarily through their gravitational influence on the parent star. • In principle they could be detected by the wobble in angular position on the celestial sphere of the parent star caused by the gravitational tug of the planet as it moves about its orbit. • In practice, it has proven easier to instead detect the wobble corresponding to the small periodic Doppler shifts for the radial motion of the parent star induced by this motion, and (in favorable cases) by small variations in light output caused by planetary transits of the parent star. 319 CHAPTER 8. FORMATION OF STARS 320 Center of Mass Period = P Radial Velocity Blueshift Redshift Full Amplitude = 2K Observer Time Figure 8.15: The Doppler wobble method for detecting extrasolar planets. 8.13.1 The Doppler Wobble Method The Doppler wobble method is illustrated in Fig. 8.15. • It requires that changes in radial velocity for the parent star of order 10 m s−1 be detected (see Exercise 8.4). • The semiamplitude K of the stellar radial velocity caused by an orbiting planet is given by 2π G 1/3 mp sin i , K= P (M∗ + mp )2/3 (1 − e2)1/2 where i is the tilt angle of the orbit, mp is the mass of the unseen companion, M∗ is the mass of the star, the orbital eccentricity is e, and the orbital period P is given by Kepler’s third law, 4π 2 a3 , P = G(M∗ + mp ) 2 where a is the semimajor axis of the relative orbit. 8.13. EXTRASOLAR PLANETS 321 51 Pegasi Radial Velocity (m/s) +100 M = 2.11 x 1030 kg +50 0 -50 0 20 40 60 80 100 120 140 160 Time (hours) Figure 8.16: Doppler shift induced by planetary companion of 51 Pegasi. • Generally, one does not know the tilt angle i, so masses are uncertain by a factor sin i in the absence of further information. • Fits of the preceding equations to the 51 Pegasi data in Fig. 8.16 indicate that the wobble of the parent star is caused by a planetary companion of 0.44 Jupiter masses in an almost circular orbit with only a 4.2 day period. • Thus the planet would orbit well inside the orbit of Mercury in our Solar System. 322 CHAPTER 8. FORMATION OF STARS 8.13.2 Precision of the Method In the general case, the precision attainable with present technology restricts the method to detection of massive planets (Jupiters and Saturns), and is more sensitive for planets relatively near their parent star. • For example, with current precision a planet of one Jupiter mass would be detectable for a = 1 AU but not at Jupiter’s actual distance of 5.2 AU. • Improvements in the Doppler technology are anticipated that would allow a Saturn-mass planet to be detected for a < 1 AU and a Jupiter-mass planet at 5 AU. • It is thought that the ultimate limits to the technique correspond to a precision of about 3 m s−1 , dictated by intrinsic stability limitations for the stellar photospheres that the method must observe. • The only Earth-mass extrasolar planets detected so far are for a special case of planets orbiting a pulsar, where one can exploit techniques depending on the precise timing associated with pulsars to detect lower-mass planets. • In principle, the transit method discussed below could also be used to detect Earth-mass planets. Relative Flux (Brightness) 8.13. EXTRASOLAR PLANETS 323 1.01 1.00 0.99 HD209458 0.98 -0.1 0.0 0.1 Time (days) Figure 8.17: Lightcurve for transit of HD209458 by an extrasolar planet. 8.13.3 Transits of Extrasolar Planets In cases where the geometry is favorable for an eclipse as seen from Earth, it is possible to detect the transit of extrasolar planets across the face of their parent star through the periodic reduction in light output for the system. • In Fig. 8.17 we show data corresponding to the transit of the star HD209458 by a planet. • Such data allow the tilt angle i of the orbit to be constrained to near π2 (since the geometry must lead to an eclipse), and from the eclipse timing the radius of the planet can be estimated. • This information, coupled with the information from Doppler analysis of the system, allows a rather full picture to be constructed: a detailed orbit of the planet, its mass, its size, and its density, and information about the atmosphere of the planet. CHAPTER 8. FORMATION OF STARS 324 Relative Intensity 1.010 Secondary eclipse 1.005 1.000 0.995 0.990 0.46 0.48 0.50 0.52 0.54 Phase Figure 8.18: Lightcurve for Secondary IR eclipse of the extrasolar planet of HD209458b. 8.13.4 Secondary Eclipses of Extrasolar Planets Infrared measurements from above the atmosphere are sensitive enough to detect the secondary eclipse of extrasolar planets (the eclipse of the planet by the parent star). • Such a secondary eclipse is illustrated in Fig. 8.18 for a planet in an eclipsing orbit around the star HD209458b. • These measurements permit the IR flux associated with the planet to be deduced from the total flux decrease during the secondary eclipse. • By fitting such data to models of planetary atmospheres one can begin to determine the properties of the planet from the IR eclipse data. • In this case it was concluded that the planet has an atmospheric temperature of about 1100 K. 8.13. EXTRASOLAR PLANETS 325 Mercury Venus Earth Mars Inner Solar System Tau Bootis 3.8 MJ 51 Peg 0.47 MJ Upsilon Andromedae 0.68 MJ 0.84 MJ 55 Cancri 2.1 MJ Gliese 876 1.1MJ Rho Cr B HD114762 10 MJ 6.6 MJ 70 Vir 16 Cyg B 1.7 MJ 47 UMa 2.4 MJ 4.0 MJ Gleise 614 0 1 2 3 Orbital Semimajor Axis (AU) Figure 8.19: Orbit radius and mass data for some extrasolar planets. 8.13.5 The Hot Jupiter Problem • Around normal stars, the planets discovered so far tend to have Jupiter-like masses and lie rather close to their parent star. • Data for some extrasolar planets are compared with corresponding quantities for our Solar System in Fig. 8.19. • There we see that many extrasolar planets of gas-giant mass have orbital semimajor axes well inside 1 AU (implying orbital periods of only a few days). • This is a situation very different from our Solar System, where massive Jupiter-like planets are at large distances from the Sun. CHAPTER 8. FORMATION OF STARS 326 Mercury Venus Earth Mars Inner Solar System Tau Bootis 3.8 MJ 51 Peg 0.47 MJ Upsilon Andromedae 0.68 MJ 0.84 MJ 55 Cancri 2.1 MJ Gliese 876 1.1MJ Rho Cr B HD114762 10 MJ 6.6 MJ 70 Vir 16 Cyg B 1.7 MJ 47 UMa 2.4 MJ 4.0 MJ Gleise 614 0 1 2 3 Orbital Semimajor Axis (AU) • This anomaly has been called the hot Jupiter problem, since in these extrasolar systems we see gas giants that are near their parent star and therefore strongly heated. • The hot Jupiter problem presents a challenge to our standard ideas of how planets form. • If this pattern remains in further study of extrasolar planets, it may require us to revise our understanding of the history of gas giants in our own Solar System. • Observation of many massive planets very near their parent stars has made popular various planetary migration theories. • In these theories, planets formed originally in outer, colder regions of the solar nebula migrate inward over time because of dynamical effects in the nebula.