= 1 cm2

Lecture 5

Basic Nuclear Physics – 3

j
Nuclear Cross Sections Flux per cm2 I
and Reaction Rates = nj v
Total area of target nuclei
per cm3 = nI σ I

Reaction rate per cm3

assuming no blocking = n jv nI σ I

The reaction rate for the two reactants, I and j

as in e.g., I (j,k) L is then:
nI n j σ I j v
which has units reactions cm-3 s-1
For example, a term in the rate equation
It is more convenient to write things
for 12C during the CNO cycle might look like
in terms of the mole fractions,
X
YI = I n I = ρ N A YI
AI ⎛ gm ⎞ ⎛ atoms ⎞ ⎛ Mole ⎞ dY(12 C)
so that the rate becomes
⎜⎝ 3 ⎟⎠ ⎜⎝
cm
⎟
Mole ⎠ ⎜⎝ gm ⎟⎠ = − ρY(12 C)Yp NA σ pγ (12 C)v +...
dt
( ρ N A ) 2 YI Y j σ I j v
and a term in a rate equation decribing the destruction of I might be
for the reaction 12C(p,γ )13 N
dYI Equivalent to
= − ρYIY j N A σ I j v +.... dnI
dt = − nI n j σ Ij v + ...
dt

Here denotes a suitable average over energies and angles

and the reactants are usually assumed to be in thermal equilibrium.
The thermalization time is short compared with the nuclear timescale.
The cross section for reaction is defined in the usual way:
number reactions/nucleus /second
σ=
number incident particles/cm 2 / second
Center of mass system – that coordinate system in which the
total initial momenta of the reactants is zero.
σ clearly has units of area (cm ) 2

For a Maxwell-Boltzmann distribution of reactant energies The energy implied by the motion of the center of mass
is not available to cause reactions.

Replace mass by the "reduced mass"

The average of the cross section times velocity is
1/2 3/2 ∞
⎛ 8 ⎞ ⎛ 1 ⎞ M1M 2
σ I jv = ⎜ ⎟ ⎜ ⎟
⎝ πµ ⎠ ⎝ kT ⎠ ∫σ
0
Ij
(E) E e− E/kT dE µ=
M1 + M 2
1/2 1/2
⎛ 2E ⎞ ⎛ 1⎞ ⎛ 2 ⎞
v =⎜ dv = ⎜ ⎟ ⎜ dE
where µ is the "reduced mass" ⎝ m ⎠⎟ ⎝ 2 ⎠ ⎝ mE ⎠⎟
3/2 1/2 Read Clayton – Chapter 4.1
M Im j ⎛ 2E ⎞
σ v 3 dv = σ ⎜
1⎛ 2 ⎞
dE
µ= ⎝ m ⎠⎟ 2 ⎝⎜ mE ⎠⎟
M I+ m j
2
= σ E dE
for the reaction I (j, k) L m2

For T in 109 K = 1 GK,  in barns (1 barn = 10-24 cm2), E6 in MeV, and The Cross Section
k = 1/11.6045 MeV/GK, the thermally averaged rate factor in cm3 s-1 is Area subtended by a
de Broglie wavelength
in the c/m system. How much the nucleus I+j looks
6.197 x 10-14
∞
 1 Characteristic quantum like the target nucleus I with j
= =
Â1/2 T93/ 2 ∫0 jk 6 6
σ jk v = σ (E ) E e−11.6045 E6 /T9 dE6 mechanical dimension
sitting at its surface. Liklihood
p k of the system
of staying inside R once you get
there.
AIA j
 = for the reaction I(j,k)L
AI + A j σ (E) = π  2 ρ Pl (E) Χ(E, A)
geometry penetration nuclear
term factor structure
If you know jk from the lab, or a calculation, in the center of mass (Cla 4-180)
frame, just put it in and integrate. The end probability a flux of
particles with energy E
The actual form of  may be very complicated and depends upon the at infinity will reach the
presence or absence of resonances and the reaction mechanism. nuclear surface. Must account
In general, however, it is of the form … for charges and QM reflection.

see Clayton Chapter 4

where  is the de Broglie wavelenth in the c/m system
π  2 π  2 0.656 barns Consider just the barrier penetration part (R < r < infinity)
π 2 = = =
µ 2 v 2 2 µ E Â E(MeV) Clayton p. 319ff shows that Schroedinger's
where 1 barn = 10-24 cm 2 is large for a nuclear cross section. equation for two interacting particles in a radial
Note that generally E(MeV) < 1 and  > Rnucleus but potential is given by (Cla 4-122) [see also our Lec 4]
 is much smaller than the interparticle spacing. χ l (r) m
Ψ(r, θ ,φ ) = Yl (θ ,φ ) *
r potential
where χ (r) satisfies
M1M 2 Z I Z j e2
µ= V (r) = r>R
M1 + M 2 ⎡ - d
2 2
l(l + 1) 2
⎤ r
⎢ + +V (r) − E ⎥ χ l (r) = 0
AA ⎣ 2 µ dr 2
2 µ r 2
⎦
 = 1 2 ~ 1 for neutrons and protons V (r) = Vnuc r<R
A1 + A2
(Clayton 4-103)
~ 4 for α -particles if A I is large
for interacting particles with both charge and angular
momentum. The angular momentum term represents the
known eigenvalues of the operator L2 in a spherical potential
*The 1/r cancels the r 2 when integrating Ψ * Ψ over
solid angles (e.g. Clayton 4-114). It is not part of
the potential dependent barrier penetration calculation.

To solve
⎡ - 2 d 2 l(l + 1) 2 ⎤
Classically, centrifugal force goes like ⎢ + +V (r) − E ⎥ χ l (r) = 0
⎣ 2 µ dr 2
2 µ r 2
⎦
mv2 m 2 v2 R 2 L2 divide by E and substitute for V(r) for r >R
Fc = = 3
=
R mR mR 3 ⎡ − 2 d 2 l(l + 1) 2 Z I Z j e2 ⎤
One can associate a centrifugal potential with this, ⎢ + + −1⎥ χ (r) = 0
⎢⎣ 2 µ E dr 2 µr 2 E
2
rE ⎥⎦
−L2
∫ Fc dR =
2mR 2
Change of radius variable. Substitute for r
2µ E 2µ E 2µ E 2
Expressing things in the center of mass system and ρ= r dρ → dr d 2ρ → d r
2 2 2
taking the usual QM eigenvaluens for the operator L2 and for Coulomb interaction chain rule
one has Z I Z j e2 2E
η= v=
−l(l + 1)  2 v µ
2 µR 2 ρ and η are dimensionless
to obtain
numbers
⎡ −d 2 l(l + 1) 2η ⎤
⎢ + + −1⎥ χ l ( ρ ) = 0
⎣ dρ
2
ρ 2
ρ ⎦
 multiply by -1
d 2χ 2η l(l + 1)
+ (1− − )χ = 0
dρ2 ρ ρ2
This is the solution for ρ Pl gives the probability of barrier penetration to the nuclear
R<r <∞ radius R with angular momentum l. In general,
has solutions (Abromowitz and Stegun 14.1.1)
http://people.math.sfu.ca/~cbm/aands/

χ = C1 Fl (η , ρ ) + C2 Gl (η , ρ ) C1 = 1 C2 = i ρ
ρ Pl = where Fl is the regular Coulomb function
Fl 2 (η, ρ ) + Gl2 (η, ρ )
where F and G, the regular and irregular Coulomb functions and Gl is the irregular Coulomb function
are the solutions of the differential equation and the constants e.g., Illiadis 2.162
come from applying the boundary conditions
See Abramowitcz and Stegun, Handbook of Mathematical Functions, p. 537
The barrier penetration function Pl is then given by
These are functions of the dimensionless variables

2
χ l (∞) Fl 2 ( ρ = ∞) +Gl2 ( ρ = ∞) 1 Z I Z j e2
Pl = = = 2 η= = 0.1575Z I Z j  / E
χ l (R)
2
Fl 2 (η , ρ ) +Gl2 (η , ρ ) Fl (η , ρ ) +Gl2 (η , ρ ) contains all the charge dependence
v
2µ E contains all the radius dependence
Cla 4-115 ρ= R = 0.2187 ÂE R fm
For the one electron atom with 2
The “1” in the numerator corresponds to a a potential
Ze2
, one obtains the
r
purely outgoing wave at infinity from a same solution but the radial component
decaying state. is Laguerre polynomials.

Z I Z j e2
Physical meaning of η =
v
On the other hand,
The classical turning radius, r0 , is given by   
= = =
2µ E R p µv ⎛1 ⎞
ρ= R= 2 ( µ ) ⎜ µv 2 ⎟
2
 ⎝2 ⎠
1 2 ZI Z je
2 
µv =
2 r0
The de Broglie wavelength on the other hand is is just the size of the nucleus measured in de Broglie
wavelengths.
2Z I Z j e2 2
  ro = =η = 2η
= = µv 2 µv
p µv This enters in, even when the angular momentum and
charges are zero, because an abrupt change in potential
r0 at the nuclear surface leads to reflection of the wave
Hence η= nb., both η and ρ are dimensionless.
2 function.

The probability of finding the particle inside of its classical turning

radius decreases exponentially with this ratio.
 Z I Z j e2
For low interaction energy, (2η >>ρ , i.e., E << )
R
and Z j ≠ 0, ρ Pl has the interesting limit There exist other interesting limits for ρ Pl ,
for example when η is small - as for neutrons where it is 0
⎡ 2l(l + 1) ⎤ ρ ∝E1/2 ρ P0 = ρ
ρ Pl ≈ 2ηρ exp ⎢ −2πη + 4 2ηρ − ⎥ Abramowitz and Stegun,
⎢⎣ 2ηρ ⎥⎦ 14.6.7 ρ3 ρ << 1 for cases of interest
ρ P1=
where 1 + ρ2 for neutron capture
( ) ρ5
1/2 independent of E and l
2ηρ = 0.2625 Z I Z j ÂR fm ρ P2 =
9 + 3ρ 2 + ρ 4
which is independent of energy but depends on nuclear size.
This implies that for l = 0 neutrons
the cross section will go as 1/v. Z I Z j e2
Note: η= = 0
v
rapid decrease with smaller energy and increasing charge(η ↑ )
2µ E
rapid decrease with increasing angular momentum E 1/ 2 ρ= R = 0.2187 ÂE R fm
ZIZ j e2 i.e., π  ρ P0 ∝
2
∝ E −1/ 2 2
η= = 0.1575Z I Z j  / E E
v
The leading order term for l = 0 proportional to For low energy neutron induced reactions, the
2µ E
ρ= R = 0.2187 ÂE R fm
2 cross section times velocity, i.e., the reaction rate
ρ Pl ∝ exp ( −2πη ) term, is approximately a constant

For particles with charge, providing X(A,E) does not vary rapidly.
with energy (exception to come), i.e., the nucleus is "structureless"

e−2πη
σ (E) = π  2 ρ Pl X ( A, E) ∝
E

This motivates the definition of an "S-factor"

S(E)= σ (E) E exp(2πη )

η = 0.1575Z I Z j  / E
AI Aj
 =
AI + Aj
This S-factor should vary slowly with energy. The first order
effects of the Coulomb barrier and Compton wavelength have been
factored out.
For those reactions in which S(E) is a slowly varying function of energy
in the range of interest and can be approximated by its value at the energy
where the integrand is a maximum, E0 ,
S(E0 ) Illiadis – Fig. 3.12
σ (E)  exp(−2πη )
E
1/2 ∞
⎛ 8 ⎞ ⎛ 1 ⎞
3/2
Maxwell Boltzmann factor and
NA σ v ≈ NA ⎜ ⎟ ⎜⎝ ⎟⎠ S(Eo ) ∫ exp(−E / kT − 2πη (E))dE
⎝ πµ ⎠ kT 0
barrier penetration factor vs
energy for the reaction
where η (E) =0.1575 Â / E(MeV ) Z I Z j 12C(,)16O at T =2. The product
8
e-E/kT times e-2 is shown on
The quantity in the integral looks like a logarithmic and a linear
scale. The Gamow peak is
at 0.32 MeV which is much
greater than kT = 17.2 keV.
The left axis shows probability
in arbitrary units.

For accurate calculations we would just enter the

energy variation of S(E) and do the integral numerically.
However, Clayton shows (p. 301 - 306) that
where E o is the Gamow Energy
⎛ −E ⎞
exp ⎜ − 2πη ⎟ can be replaced to good accuracy by
⎝ kT ⎠
⎛ − ( E − E )2 ⎞
( ) T9
2/3
C exp ⎜ 0 E 0 = πη E1/2 kT ; η E1/2 =0.1575 Â Z I Z j ; kT =
2 ⎟
, i.e. a Gaussian with the same maximum and
⎝ ( Δ / 2) ⎠ 11.6045
( )
1/3
second derivative at maximum E o = 0.122 Z I2 Z 2j ÂT92 MeV

and Δ is its full width at 1/e times the maximum

4
(
( Eo kT )1/2 = 0.237 Z I2 Z 2j ÂT95 )
1/6
Δ= MeV
3
'
Δ is approximately the harmonic mean of kT and E 0
and it is always less than E 0
e.g. 3He(α ,γ )7Be at 1.5 x 107 K In that case, the integral of a Gaussian is analytic Clayton 4-54ff
uses S in keV b,
4.34 ×10 8
otherwise the same
NA σ v = S(E0 ) τ 2 e−τ cm 3 / (Mole s)
( ) answer.
1/3
2 2 2
Eo = 0.122 Z Z ÂT I j 9
MeV Â Z I Z j

 =
(3)( 4) = 1.714; T = 0.015; ZI = Z j = 2
where S(E 0 ) is measured in MeV barns. If we define
3+4 9
λ jk = N A σ jk v

((2) (2) (1.71)(0.015) )

1/3
2 2 2
Eo = 0.122 MeV then a term in the rate equation for species I such as Yj ρλ jk has units
Different people use different
= 0.02238 MeV = 22.4 keV ⎛ Mole ⎞ ⎛ gm ⎞ ⎛ cm 3 ⎞ −1
⎜⎝ gm ⎟⎠ ⎜⎝ cm 3 ⎟⎠ ⎜⎝ Mole s ⎟⎠ = s
conventions for  which sometimes
do or do not include  or NA. This
Similarly defines mine. Clayton does not innclude

( ) Note that τ here is

1/6 NA.'
Δ = 0.237 ZI2Z j2 ÂT95 = 0.0124 MeV = 12.4 keV
1/3
3E0 ⎛ Z I2 Z 2j  ⎞
τ= = 4.248 ⎜ ⎟
differs from Clayton which
See the plot of the S-factor a few slides back kT ⎝ T9 ⎠ measures T in 106 K

2
⎛ −E ⎞ ⎛ E − E0 ⎞
exp ⎜ − 2πη ⎟ ≈ e −τ exp ⎜
⎝ kT ⎠ ⎝ Δ / 2 ⎠⎟ 1/2
λ ⎛ 2⎞ 4
⎛ 8⎞
1/2
⎛ 1⎞
3/2 ⎡ ⎛ E − E ⎞2⎤
∞
= τ 2e −τ S(E0 ) MeV1/2 amu −1/2 barn
e ∫ S(E) exp ⎢ − ⎜ ⎥ dE
NA ⎜⎝ µ ⎟⎠
−τ
λ ≈ NA ⎜ ⎟ 0

⎝ πµ ⎠ ⎝⎜ kT ⎠⎟ 0 ⎢ ⎝ Δ / 2 ⎠⎟ ⎥
⎣ ⎦
9 3 π (0.1575 ZI Z j  )
1/2
⎛ 8⎞ ⎛ 1⎞
3/2 ∞ ⎡ ⎛ E − E ⎞2⎤
= NA ⎜ ⎟ ⎜ e −τ S(E0 ) ∫ exp ⎢ − ⎜ 0 ⎥ dE
⎝ πµ ⎠ ⎝ kT ⎟⎠ ⎢ ⎝ Δ / 2 ⎟⎠ ⎥
0
⎣ ⎦ 7.2 ×10 −16 2 −τ
⎛ E − E0 ⎞ 2dE Δ dx = τ e S(E0 ) cm3 s −1 (Clay 4 − 56)
Let x = ⎜
⎝ Δ / 2 ⎟⎠
dx =
Δ
so dE =
2
ÂZI Z j
−2E0
Can replace lower bound to intergral E =
Δ
by E = - ∞ with little loss of accuracy (footnote 4.34 ×108
λ = NA σ v = S(E0 ) τ 2e −τ cm3 / (Mole s)
Clayton p 305) so that ÂZI Z j
1/2 3/2 ∞
⎛ 8⎞ ⎛ 1⎞ Δ
λ = NA ⎜ ⎟ ⎜⎝ kT ⎟⎠ e −τ S(E0 ) ∫ exp ⎡⎣ −x 2 ⎤⎦ dx
⎝ πµ ⎠ 2 −∞
1/2 3/2
⎛ 8⎞ ⎛ 1⎞ Δ
= NA ⎜ ⎟ e −τ S(E0 ) π
⎝ πµ ⎠ ⎝⎜ kT ⎠⎟ 2
1/2 3/2 nb. The unit conversion factor is 10 −24 * (6.02× 10 23 i 1.602× 10 −6 )1/2
⎛ 2⎞ ⎛ 1⎞ −τ
= NA ⎜ ⎟ e Δ S(E0 )
⎝ µ⎠ ⎜⎝ kT ⎟⎠ µ = Â amu
⎛ Δ ⎞ 4
⎜⎝ (kT )3/2 ⎟⎠ = τ2
9 3π ηE 1/2
 Adelberger et al, RMP, (1998(
A dτ A τ
f = τ 2 e −τ τ= = − 4/3 = −
T 1/3 dT 3T 3T

df dτ dτ
= 2τ e −τ − τ 2e −τ
dT dT dT

T ⎛ df ⎞ T τ T τ
= (2τ e −τ )(− ) − 2 −τ (τ 2e −τ )(− )
f ⎜⎝ dT ⎟⎠ τ 2e −τ 3T τ e 3T

⎛ d ln f ⎞ τ − 2
=⎜ =
⎝ d lnT ⎟⎠ 3

τ −2
∴ f ∝Tn n=
3

For example, 12C + 12C at 8 x 108 K Thus non-resonant reaction rates will have a
temperature dependence of the form
1/3
⎛ 62 62 12⋅12 ⎞
τ = 4.248 ⎜ 12+12
⎟
⎜⎝ 0.8 ⎟⎠
Constant constant
λ exp(− ) τ 2 e− τ
T 2/3 T1/ 3
= 90.66
This is all predicated upon S(Eo ) being constant, or at
90.66 − 2 least slowly varying. This will be the case provided:
n= = 29.5
3
i) E << ECoul ; l = 0
ii) All narrow resonances, if any, lie well outside the
p + p at 1.5 x 107 K
1/ 3
Gamow window
⎛ 1⋅1 ⎞
⎜ 1⋅1⋅ 1 + 1 ⎟ Eo ± Δ / 2
τ = 4.248 ⎜ ⎟
⎜ 0.015 ⎟ That is there are no resonances or there are
⎝ ⎠
= 13.67 very many overlapping resonances

13.67 - 2 iii) No competing reactions (e.g., (p,n), (p,) vs (p,))

n= = 3.89 open up in the Gamow window
3
 Resonant Reactions Reaction Mechanisms
In general, there are four categories of strong and electromagnetic
reactions determined by the properties of resonances through which 1)  Direct Capture - an analogue of atomic radiative capture
each proceeds The target nucleus and incident nucleon (or nucleus) react
without a sharing of energy among all the nucleons. An example
be the direct radiative capture of a neutron or proton and
S(E) ~ const •  Truly non-resonant reactions (direct capture and the immediate ejection of one or more photons. The ejected photons
the like) are strongly peaked along the trajectory of the incident projectile.
The reaction time is very short, ~ R/c ~10-21 s.

S(E) ~ const
• Reactions that proceed through the tails of broad This sort of mechanism dominates at high energy (greater than
distant resonances about 20 MeV, or when there are no strong resonances in or near the
Gamow window. It is especially important at low energies in light
S(E) highly • Reactions that proceed through one or a few nuclei where the density of resonances is very low.
variable narrow resonances within the “Gamow window” The S-factor for direct capture is smooth and featureless.

S(E)~ const • Reactions that have a very large number of Examples:

resonances in the Gamow window
3
He(α ,γ )7 Be, 2 H(p,γ )3He, 3He( 3He, 2p)4 He
12
C(n,γ )13C, 48
Ca(n,γ )49Ca

Treating the incoming

particle as a plane wave The process involves a a single matrix element and is thus
distorted by the nuclear
a single step process. Direct capture is analogous to
potential results in the
Distorted Wave Born bremsstrahlung in atoms.
Approximation often used
to calculate direct reactions. Direct capture provides a mechanism for reaction in
the absence of resonances. Usually DC cross sections are
Here the incoming particle
is represented as a plane wave much smaller than resonant cross sections on similar
which goes directly to a standing nuclei - if a resonance is present.
wave with orbital angular
momentum l in the final nucleus.
 2) Resonant Reaction:

A two step reaction in which a relatively long-lived excited I. Direct reactions (for example, direct radiative capture)
state of the compound nucleus is formed – the resonance .
This state decays statistically without any memory (other than energy
direct transition into bound states
and quantum numbers) of how it was produced. The outgoing En '
particles are not peaked along the trajectory of the incident particle. Sn
(This is called the Bohr hypothesis or the hypothesis of A+n
nuclear amnesia ). The presence of a resonance says that the
internal structure of the nucleus is important and that a long-lived B
state is being formed.
II. Resonant reactions (for example, resonant capture)
Resonances may be broad or narrow. The width is given by the
Step 1: Compound nucleus formation Step 2: Compound nucleus decay
(inverse of the ) lifetime of the state and the uncertainty principle. (in an unbound state)
ΔE Δt  
Generally states that can decay by emitting a neutron or proton will En ' '
be broad (if the proton has energy greater than the Coulomb barrier. Sn '
A+n
Resonances will be narrow if they can only decay by emitting a
photon or if the charged particle has energy << the Coulomb barrier..
B B

Other (non-radiative) channels

Step 1: Compound nucleus formation Step 2: Compound nucleus decay One or more resonances may be present in the Gamow energy
(in an unbound state) non-radiative channel window, in which case their contributions are added, or there
may be a broad resonance just outside the Gamow energy
En ' ' window, either above or below.
Sn
I+n The S-factor will be smooth in this latter case. In the case
of one or a few narrow resonances it will definitely not be
K Sα (K)
Nucleus I + n smooth. In the case of many broad overlapping resonances,
Nucleus I + n
it will be smooth again.

Resonances may be broad if they can decay by emitting a neutron,

Not all reactions emit radiation and stay within the original compound nucleus. proton or alpha-particle. For example, the 2.366 MeV (1/2+)
One may temporarily form a highly excited state that decays by ejecting excited state of 13N is broad because it can emit an energetic
e.g., n, p, or alpha-particle. E.g., I(n,)K: proton. That same state can serve as a resonance for the
reaction 12C(p,)13N which has a Q-value, Qp = 1.944 MeV
S(E) 13C(p,)14N is similar

1−
J π (13 Cground ) =
2

2.366 Excitation energy

13
(422) = 457 - 1.944 Q value for (p)
12
0.422 MeV Threshold c/m
The energy scale is given in the center of mass
fram (422 keV) needs to be converted to the lab
frame to compare with lab data. Multiply by
(A1+A2)/(A1A2)
what l-waves contribute?

For both cases the S factor is slowly varying in the Gamow “window”. Consider, however, the reaction 24
Mg(p,γ )25 Al
Say hydrogen burning at 2 x 107 K, or T9 = 0.020 This reaction might be of interest either in hot hydrogen burning
at 30 million K or in carbon burning at 800 million K. Consider the
12
C(p,γ )13N latter.
1/ 3 1/ 3
⎛ 12 ⋅ 1 ⎞ ⎛ 24 ⋅ 1 ⎞
EGamow = 0.122 ⎜ 62 12 0.022 ⎟ = 0.0289 MeV = 28.9 keV EGamow = 0.122 ⎜ 122 12 0.82 ⎟ = 0.543 MeV
⎝ 12 + 1 ⎠ ⎝ 24 + 1 ⎠

1/ 6 1/ 6
⎛ 12 ⋅ 1 ⎞ ⎛ 24 ⋅ 1 ⎞
Δ = 0.237 ⎜ 62 12 0.02
5
⎟⎠ = 0.0163 MeV = 16.3 keV Δ = 0.237 ⎜ 122 12 5
0.8
⎟⎠ = 0.447 MeV
⎝ 12 + 1 ⎝ 25 + 1

Note on the previous pages, there is no data at energies this low. Now three resonances and direct capture contribute.
As is generally the case, one must extrapolate the experimental
date to lower energies than are experimentally accessible. The
S-factor is useful for this.
Another Example: … and the corresponding S-factor Note varying widths and
RESONANT PLUS
effects for E >>  !

Not constant S-factor

for resonances
(log scale !!!!)

~ constant S-factor
for direct capture

Resonance contributions are on top of direct capture cross sections

How to calculate?
Decaying states in general have an an energy If a reaction is dominated by narrow resonances, its
distribution given by the Breit-Wigner or Cauchy cross section will be given by the Breit-Wigner equation
distribution (Clayton 3-103)*. The normalized probability (see page 347 Clayton, also probs. 3-7 and eq, 3-103).
that the state has energy E is

Γ jΓ k 2J r +1
Γ / 2π dE σ jk (E) = π  2ω ω=
P(E)dE =
(E − ε ) (2J I + 1)(2J j + 1)
2
( E − ε r )2 + ( Γ / 2 )2 r
+ Γ tot
2
/4
where
 The  s are the partial widths (like a probability but with
Γ= nb. units of energy dimensions of energy) for the resonance to break up into
τ but rather like a rate
various channels. These now contain the penetration factors.
and τ is the lifetime The lifetime of a resonance is

τ= Γ tot = ∑ Γ k  = 6.582 ×10−22 MeVsec
* Solve wave function for a quasistationary state
Γ tot
2
subject to the constraint that ∫ ψ k = exp(-t/τ ). Take This cross section will be sharply peaked around r, with a width tot
Fourier transform of ψ (t) to get ϕ (E) and normalize.
 The cross section contribution due to a single resonance is given by the
Breit-Wigner formula:

Γ tot = ∑ Γ i Γ1Γ 2
σ (E) = π  2 ⋅ ω ⋅
(E − Er )2 + (Γ / 2)2

Usual geometric factor

∝ Γ1 Partial width for decay of resonance
0.656 1 by emission of particle 1
= barn = Rate for formation of Compund
 E
nucleus state

Spin factor: ∝ Γ2 Partial width for decay of resonance

by emission of particle 2
= Rate for decay of Compund nucleus
2J r + 1
ω= into the right exit channel
(2J1 + 1)(2J 2 + 1)
 = Total width is in the denominator as
a large total width reduces the maximum
probabilities (on resonance) for
decay into specific channels.

Rate of reaction through a narrow resonance

Narrow means: Γ << ΔE

Γ1Γ 2
In this case, the resonance energy must be near the relevant energy range σ = π  2ω
E to contribute to the stellar reaction rate.
pull out front
( E − Er )2 +  ( Γ / 2 )2
∞
Recall: 8 1 −
E

3/2 ∫
< σ v >= σ (E)E e kT dE ∞ ∞
πµ (kT ) 0 dE
∫ σ (E)dE ≈ π  rω Γ1 (Er ) Γ 2 (Er ) ∫
2

0 0
(E − Er ) + (Γ r / 2)2
2

Γ1 (E)Γ 2 (E)
and σ (E) = π  ω 2

(E − Er )2 + (Γ(E) / 2)2
2π
Γr
For a narrow resonance assume:
E
−
M.B. distribution Φ(E) ∝ E e kT constant over resonance Φ(E) ≈ Φ(Er )
All widths ( ) ! constant over resonance Γ i (E) ≈ Γ i (Er )
2 constant over resonance
Then one can carry out the integration analytically (Clayton 4-193) and finds: Illiadis Table 4.12

For the contribution of a single narrow resonance to the stellar reaction rate:
−11.605 E r [MeV]
cm 3
N A < σ v >= 1.54 ⋅10 (AT9 ) −3/2
ωγ [MeV]e
11 T9

s mole

The rate is entirely determined by the resonance strength ωγ

!

2J r +1 Γ1Γ 2
ωγ =
(2J j +1)(2J I +1) Γ

Which in turn depends mainly on the total and partial widths of the resonance at
resonance energies.
Γ1Γ2
Often Γ = Γ1 + Γ 2 Then for Γ1 << Γ2 ⎯
⎯→ Γ ≈ Γ2 ⎯
⎯→ ≈ Γ1
Γ
ΓΓ
Γ 2 << Γ1 ⎯⎯
→ Γ ≈ Γ1 ⎯⎯
→ 1 2 ≈ Γ2
Γ
And reaction rate is determined by the smaller one of the widths !

As one goes to heavier nuclei and/or to higher excitation

energy in the nucleus, the number of excited states, and hence
the number of potential resonances increases exponentially.
As one goes up in
excitation energy many
more states and many Why? The thermal energy of a non-relativistic, nearly degenerate
more reactions become gas (i.e., the nucleus) has a leading term that goes as T2 where
accessible.
T is the nuclear temperature. The energy, E, of a degenerate gas
from an expansion of Fermi integrals is:
E = f(ρ ) + a(kT)2 + b (kT)4 + .... here  is the
density and  is
the partition function
One definition of temperature is
1 ∂ ln Ω 1 ∂S
= = S = k ln Ω defines T
kT ∂E T ∂E
where Ω is the number of states (i.e., the partition function)
∂ ln Ω ∂ ln Ω ∂E
=
∂T ∂E ∂T
 What is the cross section when the density of resonances is large?
⎛ ∂E ⎞
d ln Ω ~
1
kT
⎜⎝
∂T
⎟⎠ dT ~
1
kT
(
2ak 2 T   dT ) Take N (>>1) equally spaced identical resonances in an energy interval E.
For example, assume they all have the same partial widths.
ln Ω ~ 2ak ∫ dT = 2akT + const
Generate an energy averaged cross section
Ω ~ C exp ( 2akT)
D
and if we identify the excitation energy, E x ≈ a(kT)2 , E+ΔE

E ∫ σ (E) dE
1
E+ΔE N
ω Γ j Γ k dE
i.e., the first order thermal correction to the internal energy, then σ = E
ΔE
∝
ΔE ∫ ∑E 1 (E − ε r )2 + Γ r 2 / 4
Ex
( kT)2 ~
The number of excited states ω Γ jΓ k ∞
dE
a (resonances) per unit excitation
energy increases exponentialy
D << E ≈
ΔE
N ∫ (E − ε ) 2
+ Γr2 / 4
( )
0 r

Ω = C exp 2 aE x with excitation energy.

∞
dE 2π N 1
Empirically a ≈ A/9. There are corrections to a for shell ∫ (E − ε )
0
2
+ Γ 2
/ 4
=
Γr
=
ΔE D
r r
and pairing effects. In one model (back-shifted Fermi gas)
Γ jΓk T jTk
0.482 σ = 2π 2  2ω = π  2ω
C = 5/6 3/2 Γr D Ttot
A Ex Γj
where T j = 2π
D

This gives the Hauser-Feshbach formula for estimating

cross sections where the density of resonances is high.

π 2 T jl (J π , E)Tkl (J π , E)
σ jk (E) = ∑ ( 2J + 1)
(
( 2J I + 1) 2J j + 1 ) all
J rπ
r
Ttot (J π , E)

Expressions for the transmission functions for n, p, , and '

are given in Woosley et al, ADNDT, 22, 378, (1978). See also
the appendix here. A transmission function is like an average
strength function for the reaction over the energy range of interest.
It includes the penetration function. It is dimensionless and
less than 1.

This formula has been used to generate thousands of cross sections

for nuclei with A greater than about 24. The general requirement
is many ( > 10) resonances in the Gamow window.
 More levels to make
Q2
transitions to at higher
Q1 Q and also, more
phase space for the
outgoing photon.
E γ3 for electric dipole

Tγ (Q 2 ) > Tγ (Q1 )
and as a result
Tn Tγ
σ nγ ∝ ≈ Tγ
Tn + Tγ
is larger if Q is larger

Summary of reaction mechanisms

I(j,k)L
The Q-value for capture on nuclei that are tightly
bound (e.g., even-even nuclei, closed shell nuclei)
is smaller than for nuclei that are less tightly bound
(e.g., odd A nuclei, odd-odd nuclei).

As a result, nuclear stability translates into smaller

cross sections for destruction - most obviously for
nuclei made by neutron capture, but also to some
extent for charged particle capture as well.

This is perhaps the chief reason that tightly bound

nuclei above the iron group are more abundant in nature
than their less abundant neighbors.
 Summary of reaction mechanisms
Special Complications in Astrophysics
I(j,k)L
• Low energy = small cross section – experiments are hard.
• Add to Q-value and look inside nucleus I+j
• Very many nuclei to deal with (our networks often include
•  Any resonances nearby or in window 1600 nuclei; more if one includes the r-process)

No Yes •  The targets are often radioactive and short lived so that
the cross sections cannot be measured in the laboratory
Right spin and parity?
(56Ni, 44Ti, 26Al, etc)
No Yes
• Sometimes even the basic nuclear properties are not know
Tail of A few Many - binding energy, lifetime. E.g., the r-process and the rp-
Broad Narrow Overlapping process which transpire near the neutron and proton-
Direct Extrapolate Breit- Hauser-
drip lines respectively.
Capture S-factor Wigner Feshbach
• Unknown resonances in many situations

• Target in excited state effects – in the laboratory the

target is always in its ground state. In a star, it may not be
 • Electron screening
Nuclei are always completely ionized – or almost
completely ionized at temperature in stars where
nuclear fusion occurs. But the density may be
sufficiently high that two fusing nuclei do not
experience each others full Coulomb repulsion.
This is particularly significant in Type Ia supernova
ignition.
Electron screening is generally treated in two limiting
cases.
Weak screening: (Salpeter 1954)
The electrical potential of the ion is adjusted to
reflect the presence of induced polarization in the background
electrons. The characteristic length scale for this screening
is the Debye length

1/2
⎛ kT ⎞
           RD = ⎜ ζ = ∑ (Zi2 + Zi )Yi
⎝ 4π e ρ N Aς ⎟⎠
2

Clayton 2-238 and discussion before

This is the typical length scale for the clustering of charge

in the plasma. Weak screening holds if RD >> nZ-1/3

Roughly the ion sphere is the volume over

which a given ion can "polarize" the surrounding
electron cloud when that cloud has a thermal The modified Coulomb potential is then
energy ~kT. Its size is given by equating thermal
kinetic energy to electrical potential energy. e2 Z
The charge within such a cloud is (Volume)(ne e) V= exp(−r / RD )
and the charge on each ion is Ze. The volume is 4/3 π R3D
r
and ne = ZnZ . So
⎛4 3⎞
Clayton eq. 4-215 and discussion leading up to it
⎜⎝ π RD ⎟⎠ ( ZnZ e ) (Ze) In general must include
shows that, in the limit that RD >> the inter-ion
3 more than one kind of
PE =  kT
RD ions and the interaction separation, then the effect of screening is an overall
among electrons and
⎛ ⎞
1/2
among ions, not only
reduction of the Coulomb potential by an energy
⎜ kT ⎟ between ions and electrons,
RD ~ ⎜ Compare with Clayton 2-235
4 2 2 ⎟ Z I Z j e2
⎜⎝ π e Z nZ ⎟⎠ Uo =
3 These “Coulomb correction”
Differs by 3
affect the pressure and energy RD
of a gas, not just reaction rates
ρ N AYZ = nZ
This potential does not vary greatly over the region where
the rate integrand is large (Gamow energy)
 The leading order term in the screening correction
e.g., the screening
for p+p at the (after considering Mawell Boltzmann average)
is Clayton 2-262, following Salpeter (1954)
solar center is then (Clayton 4-221; see also Illiadis 3.143) shows that the total potential energy of the ion sphere,
about 5% - Illiadis
P 210
including both the repulsive interaction of the electrons
Uo among themselves and the attractive interaction with
U0 << kT f ≈ 1− = 1+ 0.188 Z I Z j ρ1/2ς 1/2 T6−3/2
kT the ions, is

Strong screening: Salpeter (1954); Salpeter and 9 ⎛ ( Ze) ⎞

2

= − 17.6 Z 5/3 ( ρYe ) eV << Gamow energy E 0

1/3
U= −
van Horn (1969) 10 ⎜⎝ RZ ⎟⎠
If RD becomes less than the inter-ion spacing, and the correction factor to the rate is exp(-U o / kT )>> 1 with
then the screening is no longer weak. Each ion of
(
−U 0 =17.6 ( ρYe ) ⎡⎢ Z I + Z j ) ⎤
1/3 5/3
charge Z is individually screened by Z electrons. − Z I5/3 − Z 5/3
j ⎥ eV (Cla 4-225)
⎣ ⎦
The radius of the “ion sphere” is
1/3 More accurate treatments are available, but this can
⎛ 3Z ⎞ 4π RZ3 clearly become very large at high density. See Itoh et al.
RZ = ⎜ i.e. ne = Z
⎝ 4π ne ⎟⎠ 3 ApJ, 586, 1436, 2003

Reflection at a Potential Change

For simplicity consider the case where the incident particle has no
charge, i.e., a neutron, and take angular momentum, l = 0.
In QM there exists reflection
Energy
whether V increases or
incident
decreases
Appendix: reflected
E
Perfectly
0 p2
Barrier Penetration x<0 absorbing – E=
E+Vo what gets in 2µ
and Transmission Functions stays in
2µ E p 2π 1
= = = ≡k
  λ 
-Vo V(x)
x>0

2µ E
Wave number for incident particles k= x<0

2 µ (E + Vo ) 2 µVo
inside well K= ≈
 
Ψ(x) = Ae ikx + Be −ikx x <0 Incident wave plus reflected wave
= Ce iKx x >0 Wave traveling to the right
Though for simplicity we took the case
Ψ(x), Ψ ′(x)continuous implies at x=0, A+B=C l = 0 and Z = 0 here, the result can be generalized
ikA − ikB =iKC to reactants with charge and angular momentum
K
B 1− k
⇒ = For Z= 0 ρ P0 = ρ l=0
A K
1+
k ρ 3
ρ P1= l =1
1 + ρ2
K K
B
2 (1+ ) 2 − (1− ) 2 4K / k 4Kk ρ5
T =1 − = k k = = The fraction that penetrates ρ P2 = l=2
A K 2 K 2 (k + K ) 2 to the region with the new 9 + 3ρ 2 + ρ 4
(1+ ) (1+ ) potential.
k k
and if E << Vo
For Z > 0
4k 4π kR 4πρ ρ
T= = = = 4π S f ρ P0 ρ Pl =
K π KR π KR Z I Z j e2 Aˆ Fl 2 (η, ρ ) + Gl2 (η, ρ )
1 recall ρ P0 = ρ = kR η=
v
= 0.1575 Z I Z j
E ( MeV )
where S = is the "black nucleus strength function"
π KR 2µ E ˆ R ( fm)
ρ= R0 = 0.2187 AE
f corrects empiricaly for the fact that the nucleus is 2
0

not purely absorptive at radius R

It is customary to define the transmission function for particles But actually the strength function is parameterized in
(not photons) as terms of the black nucleus approximation used in the
T = 4π S f ( ρ Pl ) transmission function calculation. Unknown parameters
are fit to data.

where S, the strength function, could be thought of in terms For nuclei A < 65
of resonance properies as
R = 1.25 A1/3 + 0.1 fm for n,p
Γj3 2 θ j
2

S= = (see 3 pages ahead) 1.09 A1/3 + 2.3 fm for alpha particles

D µ R2 D
which is a constant provided that θ j2 ∝ D, the level spacing. 1 2 µ Vo
S= K= Vo ≈ 60 MeV
This is consistent with the definition π KR 2
Γ
T = 2π This is what is used in the Hauser Feshbach formalism
D
Here f is the reflection factor , empirically 2.7 for n and p
and 4.8 for alpha-particles, which accounts for the fact that the
reflection is less when the potential does not have infinitely
sharp edges at R. Hence the transmission is increased.
 Analogously the photon transmission function is defined as: Semi-empirical ’s

Γγ Typically  ~ eV – larger for large E in the transition; smaller if

Tγ = 2π = Strength function * phase space factor a large J is required or E is small.
D
For nucleons and alpha particles it can be shown (Clayton 330 – 333)
3 that
Phase space  Eγ for dipole radiation
Eγ 5 for quadrupole radiation ⎛ 3 2 ⎞ 2 125.41 MeV 2
Γ lj = ⎜ 2⎟
θ j ρ Pl = θ j ρ Pl
⎝ µR ⎠ ÂR 2 ( fm)
The strength function is usually taken to be a constant
or else given a ``Giant Dipole (Lorentzian) form. where j2 is the dimensionless reduced width which must be
evaluated experimentally, but is between 0 and 1 (typically 0.1).
The transmission functions to the ground state and each excited
state are calculated separately and added together to get a total photon The resulting widths are obviously very energy sensitive (via Pl)
transmission function. but for neutrons and protons not too much less than the Coulomb
energy, they are typically keV to MeV.

The decay rate of the state is qualitatively given by (Clayton p 331) aside:
Very approximate estimates for '
λ ≡ probability/sec for particle from decaying system to cross large Typically  ~ eV – larger for large E in the transition; smaller if
spherical shell a large J is required or E is small.
1
λ = = velocity at infinity * penetration factor * probability per unit dr For nucleons and alpha particles it can be shown (Clayton 330 – 333
τ
and appendix to this lecture) that
that the particle is at the nuclear
radius ± dr use this only in the
⎛ 3 2 ⎞ 2 125.41 MeV 2 absence of any
Γ =⎜
l
2⎟
θ j ρ Pl = θ j ρ Pl experimental data
⎝ µR ⎠
j
Γ 3 2 ρ 3 3 ÂR 2 ( fm)
= = v Pl θ = Pθ2 = ρ Pθ 2
 R µR R l µ R2 l
3 4π R 2 dr where j2 is the dimensionless reduced width which must be
where = is the probability per unit radius evaluated experimentally, but is between 0 and 1 (typically 0.1).
R 4 / 3π R 3 d (volume)
volume
See appendix to this lecture (last page)
for finding the nucleon if the density is constant
The resulting widths are obviously very energy sensitive (via Pl)
θ 2 = dimensionless constant < 1 but for neutrons and protons not too much less than the Coulomb
energy, they are typically keV to MeV.
µv 2µ E
ρ = kR = R= R
 2