Miras, Mass-Loss, and the Ultimate Fate of the Earth

L. A. Willson & G. H. Bowen, Iowa State University

Fire and Ice:

Some say the world will end in fire,

Some say in ice.
From what I've tasted of desire
I hold with those who favor fire.
But if it had to perish twice,
I think I know enough of hate
To know that for destruction ice
Is also great
And would suffice

Robert Frost

Whether the Earth ends in fire or ice depends how, when,

and how fast the Sun sheds mass.

Early or extended mass loss: Earth survives the red

giant stages of the Sun and ends "in ice".

Late, abrupt mass loss: Earth is engulfed in the bloated

Sun near the end of its red giant evolution.

1
1. Stages of Solar Evolution
Main Sequence
1010 years (=10 Gyr) of
4H He in the core, then Red Giant
Branch (RGB)
4H He around a core of He
to maximum L~2000Lnow
Helium Core
abrupt start to He to C and O Flash

core He to C and O at ~ 100LnowV Horizontal Branch

Alternating H and He "burning"

Asymptotic Giant
until the C+O core ~ 0.6 MSun, when
mass loss removes the rest. Branch (AGB)

The post-main sequences stages leading up to "the end" take

an additional 2x109 years.
Mass loss
10,000
Post-AGB

1000 Asymptotic Giant

Branch Red Giant
Branch
100
L/Lnow Pre-main
10 sequence

1 Main sequence

6000 5000 4000 3000

Surface Temperature, Kelvins

2
 Increasing Solar Power => increasing T for the planets

Radiative equilibrium temperatures for the planets are

approximately given by

TRE = 290K (L/now)1/4 / (distance in AU)1/2

600

°C
Mercury
200
400
Venus
Earth 100
T,
Mercury

Kelvins
0
200
Mars
-100

-200
0
0 0.1 0.2 0.3
1 1.5 2
log(luminosity)
luminosity / luminosity now

The evolution of planetary temperatures during the main

sequence (now to about 5x109 years ahead)

Habitable zone (approximate):

3
 Extending this to the end of the AGB:
Rocky planet vaporization zone

Mercury
2000
Venus

Earth
Jupiter
Mercury

Mars
1000 Saturn

Uranus
Neptune

Habitable zone
0
0 1 log(L/now) 2 3 4
1 10 100 1000 10,000
log(luminosity)
L/now

time-> RGB tip

Now End of
Main Horizontal
Sequence Branch shell Mass Loss
flashing ends the
starts AGB

The maximum temperature a planet achieves depends on

its distance from the Sun and the maximum luminosity of
the Sun - both of these, in turn, depend on the mass loss
that ends the AGB.

4
Most astronomers have been using Reimers' Relation as a
recipe for mass loss

= -dM/dt = η 4x10-13 (L/LSun) (R/RSun)/(M/MSun) MSun/year

where LSun, RSun, and MSun refer to our present-day Sun.

This relation was derived from a fit to observations of mass

loss from red giants, with the adjustable parameter added
later when it was found that the original relation killed stars
too soon.

-4

-5 η=
logM 1
-6 0.6

-7

-8
5.8 6.0 6.2 6.4 6.6 6.8
logLR/M

Sackmann, Boothroyd and Kraemer (1993, Ap. J. 418, 457)

modeled future solar evolution using Reimers' Relation with
η=0.6.

5
Evolutionary stages in the Sackmann, Boothroyd and
Kraemer models:

10,000
shell flashing
and mass loss

Asymptotic
1000
Giant Branch

Horizontal
Branch
100
Red Giant
Branch
L/LSun

10
Pre-main
sequence

Now
1

6000 5000 4000 3000

Surface Temperature, Kelvins

6
Compared with the radius of Earth's orbit, the Sun is small
until it reaches nearly the tip of the red giant or asymptotic
giant branch:
10,000
100 RSun
AGB
RGB

1000 100 RSun

30 RSun
100

L/LSun 10 RSun

10

3 RSun

6000 5000 4000 3000

Surface Temperature, Kelvins

7
In their models Earth escapes its fiery fate mainly because
0.275 solar masses are removed on the first ascent of the red
giant branch.

Their RGB mass loss is entirely due to the Reimers' relation -

there is no mass loss "event" associated with the helium core
flash.

8
Given an evolutionary track of the form
R = function (M, L, Z),
any mass loss law (L, R, M, Z)
may be expressed as (R, M, Z).

-4 Bowen

1.0 Vassiliadis &

0.7
-5 Wood

-6
log ,
MSun/yr 0.7
Reimers
-7
1.0

-8 0.7 Earth's
orbit if
MSun = 0.7
-9
Earth's
orbit now

150 200 300 400

Radius / current solar radius

Evolution ends near dlogL/dt = -dlogM/dt - closer to this for

steeper mass loss laws, but always close. This is indicated
by the squares (0.7MSun) and dots (1MSun).

9
Using the Bowen model results, one finds that AGB evolution
ends with a "cliff"
logM = -10 -8
4 -6
4 -4
2.8
2
2
M 1.4
Chandrasekhar
1 limit
1
0.7
core mass

1000 10,000
luminosity
whose position depends on mass and metallicity.

0.6

0.4
Chandrasekhar
logM
0.01 limit
0.2
0.1
0.3 0.001
0.0
Z/Z = 1.0
core mass
-0.2

3.0 3.4 3.8 4.2 4.6

logL
10
 How can the empirical relation (Reimers' relation)
and the theoretical mass loss law (Bowen's results)
be so very different? Is one of these wrong?

A steep mass loss law => severe selection effects:

-4
Reimers' Reln. for
η=1 is shown as a
-5 dashed red line

logM 2.8 4 The cliff is the bold

-6 1.4 2
0.7 1 black line, with
-7 mass labels

-8 Evolution follows the

blue arrows
5.8 6.0 6.2 6.4 6.6 6.8
logLR/M

Reimers' relation reinterpreted: It tells us which stars

are losing mass, not how a star loses mass.

(Analogous to: The main sequence is not an evolutionary

track but the location of stars "burning" hydrogen in their
cores.)

11
The "cliff" edge stars are the Miras:

5
logL 1

4
0.7

3 "Cliff" edge
stars

2 2.2 2.4 2.6 2.8

log(Period, days)

The Sun will end up between the 0.7 and the 1 solar mass
points in this plot, depending on how much mass it loses at
the helium core flash.

With a very steep mass loss relation, it is less likely than for
the "Reimers' mass loss formula" case that the Sun will lose
much mass on the first ascent of the red giant branch. The
most likely cases are "all" or "nothing" with the possible
exception of a finite M ejected at the core flash.

12
Averaged over the pulsation cycle,

The density is
-9 rho(170)
10 increasing with time; rho(180)
the green curve rho(190)
-11 H
10 describes the density
near the end of the
density

10 -13 AGB.

10 -15

10 -17

10 -19
0 1 2 3 4 5 6
distance (AU)

10 -8 Earth's orbit
rho(170)
rho(180)
10 -10 rho(190)
Mars's orbit H

10 -12
density

10 -14

10 -16

10 -18
0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6
distance (AU)

13
 Assuming the Sun starts the AGB with a mass about the
same as it is today, we find:

time-averaged
x The Sun's
density = 10 gm/cm3, x = -15
mass
-13 decreases, and
1.5 Mars Mars escapes
R,
AU -11
max R =
1.0 -9 228 x now
Earth

Venus
R=200xnow

0.5 R=150xnow

2000 3000 4000LSun

0 1 2 3 4
time in millions of years

At the crash:
RSun = 180 Rnow (varying from 173 to 188 with P = 313 days)
LSun = 2790 Lnow

At the end:
LSun = 3950 is the maximum achieved.

14
 Shell flashes modify L and R on a scale of 1000-105 years*:
3.6
3.5
3.4 ~105 years
3.3
logL

3.2
3.1
3
2.9
2.8
1.647 10 8 1.648 10 8 1.649 10 8 1.65 10 8 1.651 10 8 1.652 10 8
t
time (years)

200
180
Radius/RSun

160
140
120
100
80
60 5 5 5 5 5 5 5 5 5
1.6 10 1.8 10 2 10 2.2 10 2.4 10 2.6 10 2.8 10 3 10 3.2 10
F
time (years)

200
180
~1000 years
Radius/RSun

160
140
120
100
80
60 5 5 5 5 5
1.6 10 1.65 10 1.7 10 1.75 10 1.8 10
F
time (years)

This will need to be included in future mass-loss calculations.

*
Models by S. Kawaler using ISUEVO, February 2000

15
The final fates of the planets, their moons, and the asteroids:

Object distance peak T Final fate

from Sun reached
(now)
(L<4000K)
Mercury 0.387 *** Into the Sun, RGB

Venus 0.723 *** Into the Sun, early

AGB
Earth 1.000 <2306> Into the Sun,
late AGB,
unless ∆MRGB>0.2 MSun
and its Moon Crashes into Earth
before Earth dies.
Mars 1.524 1868 Escapes
unless ∆MRGB ≈ 0
asteroids Small ones (<<100km)
spiral in to Sun after
Earth dies.
Jupiter 5.203 1011 Baked but not
destroyed
Saturn, Uranus, and Neptune Also baked but not
destroyed

16