Lecture 8: β Decay

•Basic process & energetics

•Fermi theory
•ft-values
•Electron-capture
•Parity violation
•Special cases

Lecture 8: Ohio University PHYS7501, Fall 2021, Z. Meisel ([email protected])

What is β decay? e- can be from atomic shells
(terrestrial cases) or from
a) Nucleus perspective: surrounding electron gas
𝛽𝛽− : 𝐴𝐴𝑍𝑍𝑋𝑋 → 𝑍𝑍+1𝐴𝐴𝑋𝑋𝑋 + 𝑒𝑒 − + ν�𝑒𝑒 (e.g. neutron star outer crust)
𝛽𝛽+ : 𝐴𝐴𝑍𝑍𝑋𝑋 → 𝑍𝑍−1𝐴𝐴𝑋𝑋𝑋 + 𝑒𝑒 + + ν𝑒𝑒
𝐸𝐸𝐸𝐸: 𝑒𝑒 − + 𝐴𝐴𝑍𝑍𝑋𝑋 → 𝑍𝑍−1𝐴𝐴𝑋𝑋𝑋 + ν𝑒𝑒

b) Nucleon perspective: I’m going to let you finish,

Free neutron decay takes ~15min
𝛽𝛽− : 𝑛𝑛 → 𝑝𝑝 + 𝑒𝑒 − + ν� 𝑒𝑒 (though there is some controversy), but the process really has to
𝛽𝛽+ : 𝑝𝑝 → 𝑛𝑛 + 𝑒𝑒 + + ν𝑒𝑒 while t½(p)>1034yr. do with quarks and bosons…
𝐸𝐸𝐸𝐸: 𝑒𝑒 − + 𝑝𝑝 → 𝑛𝑛 + ν𝑒𝑒

c) Quark perspective:
𝛽𝛽− : 𝑑𝑑 → 𝑢𝑢 + 𝑒𝑒 − + ν�𝑒𝑒
𝛽𝛽+ : 𝑢𝑢 → 𝑑𝑑 + 𝑒𝑒 + + ν𝑒𝑒
𝐸𝐸𝐸𝐸: 𝑒𝑒 − + 𝑢𝑢 → 𝑑𝑑 + ν𝑒𝑒

2
β decay energetics
• β-decay can proceed if energetics allow it (𝑄𝑄𝛽𝛽 > 0):
Atomic mass excesses have the electron
• 𝑄𝑄𝛽𝛽 = 𝑀𝑀𝑀𝑀(𝑍𝑍, 𝐴𝐴) − 𝑀𝑀𝑀𝑀(𝑍𝑍 + 1, 𝐴𝐴)
−
masses included, so it’s only for positron
• 𝑄𝑄𝛽𝛽+ = 𝑀𝑀𝑀𝑀(𝑍𝑍, 𝐴𝐴) − 𝑀𝑀𝑀𝑀(𝑍𝑍 − 1, 𝐴𝐴) − 2𝑚𝑚𝑒𝑒 emission that we have to take into account
that we will not be gaining an electron with
• 𝑄𝑄𝐸𝐸𝐸𝐸 = 𝑀𝑀𝑀𝑀 𝑍𝑍, 𝐴𝐴 − 𝑀𝑀𝑀𝑀 𝑍𝑍 − 1, 𝐴𝐴 the atom, but instead effectively losing one.

• For several cases near stability, more than one type is possible, e.g.
• An estimate for β instability for the nuclear landscape
can be determined by finding the minimum 𝑍𝑍 for a
given 𝐴𝐴 using the semi-empirical mass formula
(See homework #3)
• Unsurprisingly, the agreement with the
valley of β stability is excellent:

3
“9 out of 10 professional nuclei agree,
β decay is the preferred mode of disintegration”

4
β decay in nuclear astrophysics (selected examples)
r-process nucleosynthesis
Type I X-ray bursts (Woosley et al. ApJS 2004) (Mumpower et al. Prog. Nucl. Part. Phys (2016))

*all important half-lives for the rp-process have been

determined experimentally to sufficient precision
5
β decay spectrum, spin conservation, and the neutrino
R. Evans, The Atomic Nucleus (1955)

• Early experiments investigating the “β ray” showed that it was

not emitted with a singular energy, like the “α ray”,
but rather in a continuum of energies
• Though the maximum energy is equal to the decay Q-value
• Furthermore, the reaction 𝑛𝑛 → 𝑝𝑝 + 𝑒𝑒 − doesn’t conserve spin!
• 𝐽𝐽𝑛𝑛 = 𝐽𝐽𝑝𝑝 = 𝐽𝐽𝑒𝑒 = 12 … so 0 ≤ 𝐽𝐽𝑝𝑝 + 𝐽𝐽𝑒𝑒 ≤ 1 ≠ 12
• To remedy this issue, Pauli proposed the involvement of a 3rd
hypothetical particle, the neutrino ν Like a proper old-timey physicist, he made this proposal not in a
paper, but in a letter to physicist
• Given the above considerations, it was postulated that ν Lisa Meitner
is a spin-½ particle (“fermion”) that it is massless* and electrically neutral
(of course this isn’t quite true, but true enough for our purposes)

• In one of his last works before switching to primarily performing experimental work,
Fermi postulated (E. Fermi, Z. Phys. 1934) that nucleons could act as sources & sinks of electrons and
neutrinos, in analogy to charged particles acting as sources and sinks of photons in quantum
electrodynamics (the only successful theory of interactions between quantum particles at that point)
For what it’s worth, Nature rejected Fermi’s paper for being “too remote from physical reality” 6
Fermi theory of β decay
• Fermi posited that small perturbative interaction precipitated the β decay
• The initial state is described by the wave function of the parent nucleus in state 𝑗𝑗, Ψ𝑖𝑖 = ψ𝑝𝑝,𝑗𝑗 ,
whereas the final state is the product of the daughter nucleus in state 𝑘𝑘, ψ𝑑𝑑,𝑘𝑘 , the electron 𝜑𝜑𝑒𝑒 ,
and the neutrino* 𝜑𝜑ν , Ψ𝑓𝑓 = ψ𝑑𝑑,𝑘𝑘 𝜑𝜑𝑒𝑒 𝜑𝜑ν
*(of course it’s an antineutrino, but that’s cumbersome to say and it doesn’t matter here)
• The transition rate for such a case is derived by solving the Schrödinger equation to 1st order in
time-dependent perturbation theory, where the result is known as the Second Golden Rule:
(or often as Fermi’s Golden Rule …though he didn’t originally derive it)
2 ∗ 2
2𝜋𝜋 ′ 2𝜋𝜋 ′
λ= ћ
Ψ𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐻𝐻 Ψ𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝜌𝜌(𝐸𝐸) = ћ
∫ Ψ𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐻𝐻 Ψ𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑑𝑑𝜏𝜏 𝜌𝜌(𝐸𝐸𝑓𝑓 )
• The qualitative picture is that the transition rate is constant and depends only on two parts:
• the density of final states to which the decay can proceed
(i.e. # of ways e- and ν can share the decay energy 𝐸𝐸𝑓𝑓 )
• the “matrix element” describing the interaction
(i.e. description of initial state 𝑖𝑖 to one of many possible final states 𝑓𝑓)
• 𝜌𝜌(𝐸𝐸𝑓𝑓 ) is a phase-space factor described by kinematics
• The matrix element describes the wave-function overlap between initial and final states,
so a theoretical description requires calculating these wave functions
7
β decay phase space factor
𝑑𝑑𝑑𝑑
• The decay phase-space 𝜌𝜌 𝐸𝐸𝑓𝑓 = describes how many ways 𝐸𝐸𝑓𝑓 can be split between e- and ν
𝑑𝑑𝑑𝑑
• The number final states available is:
𝑑𝑑𝑑𝑑 = # 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑖𝑖𝑖𝑖 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑉𝑉 𝑜𝑜𝑜𝑜 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑤𝑤𝑤𝑤𝑤𝑤𝑤 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑝𝑝𝑒𝑒 𝑎𝑎𝑎𝑎𝑎𝑎 𝑝𝑝𝑒𝑒 + 𝑑𝑑𝑝𝑝𝑒𝑒
×
(# 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑖𝑖𝑖𝑖 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑉𝑉 𝑜𝑜𝑜𝑜 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑤𝑤𝑤𝑤𝑤𝑤𝑤 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑝𝑝ν 𝑎𝑎𝑎𝑎𝑎𝑎 𝑝𝑝ν + 𝑑𝑑𝑝𝑝ν )
• The number of available electron states between 𝑝𝑝𝑒𝑒 + 𝑑𝑑𝑝𝑝𝑒𝑒 is derived
by solving for the combination of momentum components 𝑝𝑝𝑥𝑥 , 𝑝𝑝𝑦𝑦 , 𝑝𝑝𝑧𝑧
that satisfy our momentum criterion:
(as we found for the nuclear level density when discussing the Fermi Gas Model)
𝑉𝑉 2
• 𝑛𝑛 𝑝𝑝𝑒𝑒 𝑑𝑑𝑝𝑝𝑒𝑒 = 4𝜋𝜋𝑝𝑝𝑒𝑒 𝑑𝑑𝑝𝑝𝑒𝑒
(2𝜋𝜋ћ)3
𝑉𝑉 2 𝑉𝑉 2 16𝜋𝜋2 2 2
• Similarly, 𝑛𝑛 𝑝𝑝ν 𝑑𝑑𝑝𝑝ν = 4𝜋𝜋𝑝𝑝ν 𝑑𝑑𝑝𝑝ν …so 𝑑𝑑𝑑𝑑 = 𝑝𝑝 𝑝𝑝 𝑑𝑑𝑝𝑝𝑒𝑒 𝑑𝑑𝑝𝑝ν
(2𝜋𝜋ћ)3 (2𝜋𝜋ћ)6 𝑒𝑒 ν
𝐾𝐾𝐾𝐾 𝑄𝑄−𝐾𝐾𝐾𝐾𝑒𝑒 𝑑𝑑𝑑𝑑
• For 𝑚𝑚ν = 0, 𝑝𝑝ν = ν = …so, for a fixed 𝑝𝑝𝑒𝑒 , 𝑑𝑑𝑑𝑑ν =
2
𝑐𝑐 2 𝑐𝑐 𝑐𝑐 L. van Dommelen, Quantum Mechanics for Engineers (2012)
𝑉𝑉 16𝜋𝜋 2 2
• As such, 𝑑𝑑𝑑𝑑 = 𝑄𝑄 − 𝐾𝐾𝐾𝐾𝑒𝑒 𝑝𝑝𝑒𝑒 𝑑𝑑𝑝𝑝𝑒𝑒 𝑑𝑑𝑄𝑄
(2𝜋𝜋ћ)6 𝑐𝑐 3
𝑑𝑑𝑑𝑑 𝑉𝑉 2 16𝜋𝜋2 2 2
• Meaning the change in the # of final states for a change in 𝑄𝑄 is: = 6 3 𝑄𝑄 − 𝐾𝐾𝐾𝐾 𝑒𝑒 𝑝𝑝𝑒𝑒 𝑑𝑑𝑝𝑝𝑒𝑒
𝑑𝑑𝑑𝑑 (2𝜋𝜋ћ) 𝑐𝑐
• Which is the 𝜌𝜌 𝐸𝐸𝑓𝑓 we were after 8
β decay phase space factor & the β energy spectrum
• Considering the β decay rate for an electron momentum within 𝑝𝑝𝑒𝑒 + 𝑑𝑑𝑝𝑝𝑒𝑒 ,
2
2𝜋𝜋 ′
λ(𝑝𝑝𝑒𝑒 )𝑑𝑑𝑑𝑑𝑒𝑒 = ћ Ψ𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐻𝐻 Ψ𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝜌𝜌(𝐸𝐸)
• The matrix element is just some number, so the functional form is from 𝜌𝜌(𝐸𝐸)
• Therefore, we expect λ 𝑝𝑝𝑒𝑒 𝑑𝑑𝑑𝑑𝑒𝑒 ∝ 𝑄𝑄 − 𝐾𝐾𝐾𝐾𝑒𝑒 2 𝑝𝑝𝑒𝑒2

k Wu & Albert, Phys. Rev (1949)

Not too bad, but what effect are we forgetting that

will cause positrons and electrons to behave differently? Coulomb repulsion! 9
Coulomb distortion to the β spectrum
• Fermi realized that the protons in the daughter nucleus would repel e+ and attract e-,
modifying the resultant spectrum by a factor 𝐹𝐹 𝑍𝑍𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 , 𝑝𝑝𝑒𝑒
(in an unfortunate convention, this coulomb distortion factor is often called the Fermi function …or a better term is the Fermi screening factor)

• 𝐹𝐹 𝑍𝑍𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 , 𝑝𝑝𝑒𝑒 is pretty nasty to calculate, so numerical tables are generally used instead
(See e.g. J.Reitz, Phys. Rev. (1949))

• A non-relativistic approximation that works for nuclides with 𝑍𝑍𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 ≪ 2 � 137 is

(See R. Evans, The Atomic Nucleus (1955) for the relativistic version that works well for all Z)
2𝜋𝜋𝜋𝜋
𝐹𝐹 𝑍𝑍𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 , 𝑝𝑝𝑒𝑒 ≈ ,
1−𝑒𝑒 −2𝜋𝜋𝜋𝜋
±𝑍𝑍𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝛼𝛼
where 𝑦𝑦 ≡ 𝐸𝐸𝑒𝑒 , 𝛼𝛼 ≈ 1/137,
𝑝𝑝𝑒𝑒 𝑐𝑐
𝐸𝐸𝑒𝑒 = 𝑚𝑚𝑒𝑒 𝑐𝑐 2 + 𝐾𝐾𝐾𝐾𝑒𝑒 ,
and + is for β- decay and − is for β+ decay
• An improvement to this approximation is to multiply
it by the factor 𝑝𝑝𝑒𝑒2 1 + 4𝛾𝛾 2 − 1 𝑆𝑆 , where 𝛾𝛾 = 𝛼𝛼𝑍𝑍,
𝑆𝑆 = 1 − 𝛼𝛼 2 𝑍𝑍 2 − 1, with 𝛼𝛼 as the fine-structure
constant. It is “accurate to about 1%”
(Venkataramaiah et al. JPG 1985)
10
 𝟐𝟐

β decay matrix element λ = 2𝜋𝜋

ћ
∗
� 𝜳𝜳𝒇𝒇𝒇𝒇𝒇𝒇𝒇𝒇𝒇𝒇 𝑯𝑯′ 𝜳𝜳𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊 𝒅𝒅𝝉𝝉 𝜌𝜌(𝐸𝐸𝑓𝑓 )

• Though the phase-space factor and Coulomb distortion give us the β spectrum,
we still need to evaluate the initial/final wave function overlap to get the decay rate
∗ 2 ∗ 2
′ ∗ ∗
• ∫ 𝛹𝛹𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐻𝐻 𝛹𝛹𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑑𝑑𝜏𝜏 = ∫ ψ𝑑𝑑,𝑘𝑘 𝜑𝜑𝑒𝑒 𝜑𝜑ν 𝐺𝐺𝐹𝐹 ψ𝑝𝑝,𝑗𝑗 𝑑𝑑𝜏𝜏 , where 𝐺𝐺𝐹𝐹 is the constant describing the
perturbation (the Fermi coupling constant 𝐺𝐺𝐹𝐹 ≈ 8.9 × 10−5𝑀𝑀𝑀𝑀𝑀𝑀 � 𝑓𝑓𝑓𝑓3 )
The dimensionless form of this coupling constant is GF*(MWc2)2/(√2*4π(ћc)3) ≡ αW ~ 10-3*αfs …so it’s “weak”
1 −𝑖𝑖 𝑝𝑝⃗ν �𝑟𝑟
• The neutrino free-streams out, so it is treated as an outgoing plane wave 𝜑𝜑ν∗ = 𝑒𝑒 ћ
𝑉𝑉
• The electron has more of an interaction with the nucleus, but the Coulomb part is taken care of
𝑟𝑟
1 −𝑖𝑖 ⃗
𝑝𝑝 �
by the Coulomb distortion from earlier, so it is also treated as a plane wave 𝜑𝜑𝑒𝑒∗ = 𝑒𝑒 𝑒𝑒 ћ
𝑉𝑉
𝑟𝑟 i.e. no angular
−𝑖𝑖(𝑝𝑝⃗ν +𝑝𝑝⃗𝑒𝑒 )�
• To first order in a Taylor expansion, 𝑒𝑒 ћ ≈ 1 − −𝑖𝑖 𝑝𝑝⃗ν + 𝑝𝑝⃗𝑒𝑒 � + ⋯ ≈ 1, 𝑟𝑟
ћ
momentum transfer
(we’ll consider higher orders later)
∗ ∗ ∗ 2 𝐺𝐺𝐹𝐹2 ∗ 2 𝐺𝐺𝐹𝐹2 2
so: ∫ ψ𝑑𝑑,𝑘𝑘 𝜑𝜑𝑒𝑒 𝜑𝜑ν 𝑔𝑔ψ𝑝𝑝,𝑗𝑗 𝑑𝑑𝜏𝜏 ≈ ∫ ψ𝑑𝑑,𝑘𝑘 ψ𝑝𝑝,𝑗𝑗 𝑑𝑑𝜏𝜏 ≡ 𝑀𝑀𝑓𝑓𝑓𝑓
𝑉𝑉 2 𝑉𝑉 2
2
• 𝑀𝑀𝑓𝑓𝑓𝑓 is the overlap between wave-function for state 𝑗𝑗 of the parent nucleus and the wave-
function for state 𝑘𝑘 of the daughter nucleus and is known as the nuclear matrix element
11
 2

β decay rate per electron momentum 𝜆𝜆 = 2𝜋𝜋

ћ
∗
� 𝛹𝛹𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐻𝐻 ′ 𝛹𝛹𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑑𝑑𝜏𝜏 𝜌𝜌(𝐸𝐸𝑓𝑓 )

• Combining all the pieces, You may have worried that things
2𝜋𝜋 𝐺𝐺𝐹𝐹2 2 𝑉𝑉 2 16𝜋𝜋2 depended on the nuclear volume
λ 𝑝𝑝𝑒𝑒 𝑑𝑑𝑑𝑑𝑒𝑒 = 𝑀𝑀𝑓𝑓𝑓𝑓 𝐹𝐹 𝑍𝑍𝑑𝑑 , 𝑝𝑝𝑒𝑒 𝑄𝑄 − 𝐾𝐾𝐾𝐾𝑒𝑒 2 𝑝𝑝𝑒𝑒2 𝑑𝑑𝑝𝑝𝑒𝑒 V, but now you can see it cancels
ћ 𝑉𝑉 2 (2𝜋𝜋ћ)6 𝑐𝑐 3
𝑮𝑮𝟐𝟐𝑭𝑭 𝟐𝟐
• Cancelling & consolidating: 𝝀𝝀 𝒑𝒑𝒆𝒆 𝒅𝒅𝒅𝒅𝒆𝒆 = 𝑴𝑴𝒇𝒇𝒇𝒇 𝑭𝑭 𝒁𝒁𝒅𝒅 , 𝒑𝒑𝒆𝒆 𝑸𝑸 − 𝑲𝑲𝑲𝑲𝒆𝒆 𝟐𝟐 𝒑𝒑𝟐𝟐𝒆𝒆 𝒅𝒅𝒑𝒑𝒆𝒆
𝟐𝟐𝝅𝝅𝟑𝟑 ћ𝟕𝟕 𝒄𝒄𝟑𝟑
• This provides us with a neat way to ascertain the decay Q-value
R. Evans, The Atomic Nucleus (1955)
• From the bold equation,
𝝀𝝀 𝒑𝒑𝒆𝒆
we see ∝ 𝑴𝑴𝒇𝒇𝒇𝒇 𝑸𝑸 − 𝑲𝑲𝑲𝑲𝒆𝒆
𝒑𝒑𝟐𝟐𝒆𝒆 𝑭𝑭 𝒁𝒁𝒅𝒅 ,𝒑𝒑𝒆𝒆

• This yields a straight line which intersects the

horizontal axis at 𝑲𝑲𝑲𝑲𝒆𝒆 = 𝑸𝑸
• This type of plot is known as a Kurie plot Q=1.989MeV

• Note that this simple form only applies to

“allowed transitions”
…which we’ll discuss more about in a moment
12
Aside: ν mass from the Kurie plot
• It turns out that incorporating a non-zero neutrino mass into the phase-space factor calculation
𝐺𝐺𝐹𝐹2 2 𝑚𝑚ν2 𝑐𝑐 4
yields 𝜆𝜆 𝑝𝑝𝑒𝑒 𝑑𝑑𝑑𝑑𝑒𝑒 = 𝑀𝑀𝑓𝑓𝑓𝑓 𝐹𝐹 𝑍𝑍𝑑𝑑 , 𝑝𝑝𝑒𝑒 𝑄𝑄 − 𝐾𝐾𝐾𝐾𝑒𝑒 2 𝑝𝑝𝑒𝑒2 1− 𝑑𝑑𝑝𝑝𝑒𝑒
2𝜋𝜋3 ћ7 𝑐𝑐 3 𝑄𝑄−𝐾𝐾𝐾𝐾𝑒𝑒 2 E. Fermi, Z.Phys. (1934)

• Fermi realized one could possibly use this to determine 𝑚𝑚ν

• The best shot of doing this would be for small 𝑄𝑄,
for which tritium (𝑄𝑄 ≈ 18.6𝑘𝑘𝑘𝑘𝑘𝑘) is the best case
• Limits from such measurements place R. Evans, The Atomic Nucleus (1955)
𝑚𝑚ν𝑒𝑒 < 1.1𝑒𝑒𝑒𝑒 (KATRIN Collaboration, M. Aker et al. PRL 2019))
This is complementary to the more stringent
limit placed by astronomical observations:
Σmν<0.3eV
(A. Goobar et al. JCAP (2006))

13
Aside: nuclear mass from the Kurie plot
• Of course the end-point of the Kurie plot can also be used to obtain nuclear masses, assuming
the mass of the less-exotic nucleus in the decay is known
• This is known as the “β end-point method”
…which has fallen out of favor lately due to systematic discrepancies with higher-precision techniques
S. Rinta-Antila et al. PRC (2004)
I. Dillman et al. PRL (2003)

β end-point

Penning Trap

14
Total β decay rate and the 𝑓𝑓𝑓𝑓 value
• All that’s left is to integrate over the momentum distribution
𝐺𝐺𝐹𝐹2 2 𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚
𝜆𝜆 = 𝑀𝑀𝑓𝑓𝑓𝑓 ∫0 𝐹𝐹 𝑍𝑍𝑑𝑑 , 𝑝𝑝𝑒𝑒 𝑄𝑄 − 𝐾𝐾𝐾𝐾𝑒𝑒 2 𝑝𝑝𝑒𝑒2 𝑑𝑑𝑝𝑝𝑒𝑒 , where 𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚 𝑐𝑐 = 𝑄𝑄2 − 𝑚𝑚𝑒𝑒2 𝑐𝑐 4
2𝜋𝜋3 ћ7 𝑐𝑐 3
• Unfortunately, life is hard and so is that integral Evaluating the Fermi integral shows
• The dimensionless Fermi integral is defined as: λ∝Q5, which is known as “Sargent’s Rule”
1 𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚 2 2
𝑓𝑓 𝑍𝑍𝑑𝑑 , 𝑄𝑄 ≡ ∫ 𝐹𝐹 𝑍𝑍 , 𝑝𝑝
𝑑𝑑 𝑒𝑒 𝑄𝑄 − 𝐾𝐾𝐾𝐾𝑒𝑒 𝑝𝑝𝑒𝑒 𝑑𝑑𝑝𝑝𝑒𝑒
𝑚𝑚𝑒𝑒 𝑐𝑐 3 𝑚𝑚𝑒𝑒 𝑐𝑐 2 2 0
and numerical integration or tables of solutions are used to evaluate it
• Our tidy expression for the total decay rate (a.k.a. the decay constant) is therefore,
𝐺𝐺𝐹𝐹2 𝑚𝑚𝑒𝑒5 𝑐𝑐 4 2 ln(2)
𝜆𝜆 = 𝑀𝑀𝑓𝑓𝑓𝑓 𝑓𝑓 𝑍𝑍𝑑𝑑 , 𝑄𝑄 =
2𝜋𝜋3 ћ7 𝑡𝑡½
• Since 𝑄𝑄 and 𝑡𝑡½ can be determined experimentally, the “comparative half-life” 𝑓𝑓 𝑍𝑍𝑑𝑑 , 𝑄𝑄 𝑡𝑡½ is
2ln(2)𝜋𝜋3 ћ7 1
used to determine the matrix element 𝑓𝑓𝑓𝑓 = 2 5 4 2
𝐺𝐺𝐹𝐹 𝑚𝑚𝑒𝑒 𝑐𝑐 𝑀𝑀𝑓𝑓𝑓𝑓
• Alternatively, a calculation of the wave-function overlap can be used to determine 𝑡𝑡½ if
experimental data or theoretical calculations are available for 𝑄𝑄
• Because 𝑡𝑡½ (β) for nuclei spans orders of magnitude for nuclei, log10 (𝑓𝑓𝑓𝑓) is often quoted 15
 R. Evans, The Atomic Nucleus (1955) …a similar open-source plot is in L. van Dommelen’s Quantum Mechanics for Engineers

The Fermi integral

In case you find yourself
needing to calculate the
f(Zd,Q), use a graph such
as this one, which was
determined numerically.

Could instead reproduce

the calculations of
Towner & Hardy PRC
2015, pgs. 015501.

16
 Fundamental physics with super-allowed transitions
• For mirror nuclei (swapped 𝑁𝑁 & 𝑍𝑍), decays between states with identical 𝐽𝐽𝜋𝜋 (always 0+ to 0+) the initial
2
and final state wave functions are expected to be nearly identical, so 𝑀𝑀𝑓𝑓𝑓𝑓 = 1
• Such transitions are called “super-allowed”
• E.g. 14O(0+)→14N(0+)+e-+ νe, Q=2.831MeV
•From the Fermi integral plot on the previous page, 𝑓𝑓 ≈ 102.5
several corrections are responsible for 𝑓𝑓 → F
2ln(2)𝜋𝜋3 ћ7 1 2 ln 2 𝜋𝜋3 ћ𝑐𝑐 7 1
•So 𝑡𝑡½ = ≈ ≈ 20𝑠𝑠 Hardy & Towner, PRC (2015)
𝐺𝐺𝐹𝐹2 𝑚𝑚𝑒𝑒5 𝑐𝑐 4 𝑓𝑓 𝐺𝐺𝐹𝐹2 𝑚𝑚𝑒𝑒 𝑐𝑐 2 5 𝑐𝑐 𝑓𝑓
•The actual 𝑡𝑡½ ≈ 70𝑠𝑠 …which isn’t half bad
for reading off of a steeply logarithmic plot!
• It turns out that, given adequate corrections,
super-allowed transitions all seem to obey
2
𝑀𝑀𝑓𝑓𝑓𝑓 = 1, which is something known as
the conserved vector current (CVC)
hypothesis.
• Any deviation from this would imply
physics beyond the standard model 17
Gamow-Teller β decay transitions
2

𝜆𝜆 = 2𝜋𝜋
ћ
∗
� 𝛹𝛹𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐻𝐻 ′ 𝛹𝛹𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑑𝑑𝜏𝜏 𝜌𝜌(𝐸𝐸𝑓𝑓 )

• Earlier, when we made the approximation for the electron and neutrino 𝑟𝑟
wave functions that
we could take the first term of the Taylor expansion: 𝑒𝑒 −𝑖𝑖(𝑝𝑝⃗ν+𝑝𝑝⃗𝑒𝑒 )�ћ ≈ 1,
we consequently made the result that the β decay would result in no angular momentum
transfer (because we ignored the terms with 𝑝𝑝⃗ � 𝑟𝑟)
⃗
• And yet, some observed transitions with ∆𝐽𝐽 = 1 (or with ∆𝐽𝐽 = 0 from a non-0+ state) yield
lifetimes as if they’re not reduced by the amount one would expect from the angular
momentum transfer
• These cases, Gamow-Teller decays, happen due to the coupling of spin between the electron
and the neutrino, which can be anti-aligned (𝑆𝑆 = 0) or aligned (𝑆𝑆 = 1)
• For the anti-aligned case, zero angular momentum is carried away and parity is conserved
…these are the Fermi decays we’ve covered thus far
• For the aligned case, 1 or 0 units of angular momentum can be carried away and parity is still
conserved ….these are Gamow-Teller decays
• Note that GT decays cannot happen from 0+ to 0+,
since spin 0 (from the beta-decaying nucleus) and 1 (from 𝑆𝑆 = 1) can only combine to be 1
• For GT decays, the wavefunction overlap won’t be as good as for 0+ to 0+, so these decays are
“allowed” but never “super allowed” 18
Decay selection rules and “forbidden” decays
2

𝜆𝜆 = 2𝜋𝜋
ћ
∗
� 𝛹𝛹𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐻𝐻′ 𝛹𝛹𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑑𝑑𝜏𝜏 𝜌𝜌(𝐸𝐸𝑓𝑓 )
𝑟𝑟
−𝑖𝑖(𝑝𝑝⃗ν +𝑝𝑝⃗𝑒𝑒 )�ћ
• As was just alluded to, ignoring higher-order terms of the 𝑒𝑒 Taylor expansion omits
the possibility for angular momentum transfer
• If angular momentum transfer is to occur, higher-order terms need to be included and it will
2
no longer be the case that 𝑀𝑀𝑓𝑓𝑓𝑓 is independent of 𝑝𝑝𝑒𝑒
2
• In fact, for these cases ∆𝐽𝐽 > 1 and/or ∆𝜋𝜋 = 𝑦𝑦𝑦𝑦𝑦𝑦, the leading-order overlap 𝑀𝑀𝑓𝑓𝑓𝑓 = 0
and so a higher-order term will be necessary
• The order that’s required will correspond to the angular momentum transfer of the decay ∆𝐽𝐽
• This combined with whether or not parity is changed is referred to as how “forbidden” a
transition is…even though it’s just a hindrance
• 0+ − 0+ → "𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 − 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎"
• 0+ − 1+ 𝑜𝑜𝑜𝑜 ∆𝐽𝐽 = 0 𝑜𝑜𝑜𝑜 1 𝑎𝑎𝑎𝑎𝑎𝑎 ∆𝜋𝜋 = 𝑛𝑛𝑛𝑛 → "𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
• ∆𝐽𝐽 = 0 𝑜𝑜𝑜𝑜 1, ∆𝜋𝜋 = 𝑦𝑦𝑦𝑦𝑦𝑦 → "𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓"
• ∆𝐽𝐽 = 2, ∆𝜋𝜋 = 𝑛𝑛𝑛𝑛 → "𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓
• For a given transition type, 𝑓𝑓𝑓𝑓 will typically be within an order of magnitude of some value
19
Empirical 𝑓𝑓𝑓𝑓 value distribution
W. Meyerhof, Elements of Nuclear Physics (1967)

Loveland, Morrissey, & Seaborg, Modern Nuclear Chemistry (2006)

Empirical ranges for log(ft) given a particular ∆𝐽𝐽, ∆𝜋𝜋 are available in compilations,
e.g. B. Singh et al. Nuclear Data Sheets (1998) .
For a given ∆𝐽𝐽, ∆𝜋𝜋 plausible values of log(ft) often range within ±1.5 of the median, i.e. this gets you a ballpark value.
20
Tabular method to find 𝑓𝑓𝑓𝑓 value
• Supposing you know the (state-to-state) Q-value and half-life, 𝑓𝑓𝑓𝑓 can be evaluated using a graphical/tabular
method that summarizes the result of a large number of analytic calculations (S.Moszkowski, Phys.Rev. (1951))
• log 𝑓𝑓𝑓𝑓 = log 𝑓𝑓0 𝑡𝑡 + log 𝐶𝐶 + ∆log(𝑓𝑓𝑓𝑓), where log 𝑓𝑓0 𝑡𝑡 is from (1), log 𝐶𝐶 from (2), and ∆log(𝑓𝑓𝑓𝑓) from (3)
(1) (2) (3)

p=branching %

21
NNDC’s online calculator for 𝑓𝑓𝑓𝑓 values
• If you already have experimental information about the decay, such as the Q-value and half-life, you can directly
calculate the ft-value with the NNDC’s online calculator at https://www.nndc.bnl.gov/logft/
Using 𝑓𝑓𝑓𝑓 for spectroscopy
• Since we established 𝑓𝑓𝑓𝑓 is linked to a particular ∆𝐽𝐽, ∆𝜋𝜋, a measurement of 𝑓𝑓𝑓𝑓 can constrain 𝐽𝐽𝜋𝜋 of
an unknown state if the β decay proceeds from a state with known 𝐽𝐽𝜋𝜋
• Nuclear masses combined with γ ray energy I. Dillman et al. PRL (2003)
measurements provide the transition Q-value,
i.e. 𝑓𝑓
• The % of the time the decay proceeds
through a given state (the branching ratio)
gives the partial half-life for that decay
𝑡𝑡½
(𝑡𝑡½ ′ = ), i.e. 𝑡𝑡
𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵
• Given knowledge of the parent 𝐽𝐽𝜋𝜋
(always 0+ for even-even nuclei), one can infer
𝐽𝐽𝜋𝜋 for the daughter state
Loveland, Morrissey, & Seaborg, Modern Nuclear Chemistry (2006)

23
λ for Electron capture
2

𝜆𝜆 = 2𝜋𝜋
ћ
∗
� 𝛹𝛹𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐻𝐻′ 𝛹𝛹𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑑𝑑𝜏𝜏 𝜌𝜌(𝐸𝐸𝑓𝑓 )

• Rather than a nucleon undergoing transmutation by its lonesome, instead e--capture can occur
• This is either due to a capture of a low-lying (usually the “K-shell”) electron or due to the electron
Fermi energy in an electron-degenerate environment being high enough to overcome the
electron-capture Q-value
• The decay constant for electron-capture decay is a bit different than for β decay, because the
final state only consists of a nucleon and a neutrino … i.e. 𝐾𝐾𝐾𝐾𝑒𝑒 = 0 and Ψ𝑓𝑓 = ψ𝑑𝑑,𝑘𝑘 𝜑𝜑ν
𝐺𝐺𝐹𝐹2 2 2
• The decay constant is then: 𝜆𝜆 = 𝜑𝜑𝐾𝐾 (0) 2 ,
𝑀𝑀𝑓𝑓𝑓𝑓 𝑇𝑇ν
2𝜋𝜋3 ћ3 𝑐𝑐 3
where 𝜑𝜑𝐾𝐾 (0) is the wave-function for the inner-most atomic electron (the one in the “K-shell”)
3/2
1 𝑍𝑍𝑚𝑚𝑒𝑒 𝑒𝑒 2
• You may recall from your Quantum class, 𝜑𝜑𝐾𝐾 0 =
𝜋𝜋 4𝜋𝜋𝜀𝜀0 ћ2
λ𝐾𝐾
• As such, the ratio of electron-capture to β+ decay for a nucleus goes as ∝ 𝑍𝑍 3
λ β+
(of course, 𝑄𝑄𝛽𝛽 > 2𝑚𝑚𝑒𝑒 is a requirement for β+ decay to be possible in the first place)
Since EC decay only emits a neutrino, which will be almost impossible for us to detect,
how do you figure EC decay is usually detected? X-ray and Auger electron emission due to
atomic electrons filling the vacated orbital 24
Urca cooling: EC-β- cycling H.Schatz et al Nature (2014)

• In extremely dense environments, electrons are

degenerate, meaning that electrons are available with
an energy equal to the electron fermi energy 𝐸𝐸𝐹𝐹,𝑒𝑒
• When 𝐸𝐸𝐹𝐹,𝑒𝑒 ≈ 𝑄𝑄𝐸𝐸𝐸𝐸 , electron capture will proceed
• However, when 𝐸𝐸𝐹𝐹,𝑒𝑒 − 𝑄𝑄𝐸𝐸𝐸𝐸 ≲ 𝑘𝑘𝐵𝐵 𝑇𝑇, there is some
phase-space open near the Fermi surface for β- decay
to occur [note that normally an electron would be forbidden from re-entering
the environment due to Pauli exclusion]

• In these sweet spots, EC-β- cycling occurs, releasing two

neutrinos with each cycle
• The neutrinos carry away energy with them and hence
the phenomenon is known as Urca cooling
𝑄𝑄5 𝑇𝑇 5 𝑓𝑓𝑓𝑓𝐸𝐸𝐸𝐸 +𝑓𝑓𝑓𝑓𝛽𝛽
• 𝐿𝐿ν ∝ , where 𝑓𝑓𝑡𝑡 ′ = B. Paxton et al. ApJS 2016
𝑓𝑓𝑓𝑓 𝛽𝛽 𝑓𝑓𝑓𝑓 𝐸𝐸𝐸𝐸
𝑓𝑓𝑓𝑓𝑓 2
[The 𝑓𝑓𝑓𝑓-values are different because of the different spin- =
degeneracy of the parent state. This is usually negligible because 2𝐽𝐽𝛽𝛽 + 1 2𝐽𝐽𝐸𝐸𝐸𝐸 + 1
𝑓𝑓𝑓𝑓 is often uncertain by orders of magnitude.] where 𝐽𝐽 is for the parent state
25
Urca cooling in astrophysics (selected examples)
Light curves of neutron star cooling Core temperature evolution for
following accretion turn-off accreting ONe WDs
(Meisel & Deibel, Astrophys. J. (2017)) (Schwab, Bildsten, & Quataert MNRAS (2017))

26
Transition Strengths: B(F) and B(GT)
• Theoretical calculations of weak transition rates characterize such rates with transition strengths,
where B(F) and B(GT) are the Fermi and Gamow-Teller transition strengths, respectively
• Each is described by he modulus-square of the expectation value
of the relevant transition operator
[See e.g. : Lecture Notes in Nuclear Structure Physics (B.A. Brown)]
1
• Their sum is inversely proportional to the comparative half-life: 𝑓𝑓𝑓𝑓 ∝
𝐵𝐵 𝐹𝐹 +𝐵𝐵(𝐺𝐺𝐺𝐺)
• B(GT) is also handy because it can be deduced from charge-exchange measurements:
M. Sasano et al. PRL (2011)

27
Parity Non-Conservation
• For α-decay, parity was an important consideration for selection rules
• Transitions were only possible for which the parity change was ∆𝜋𝜋 = (−1)𝑙𝑙
• For weak transitions, this is not the case
• It was demonstrated by observing the 60Co β-decay
angular distribution for 60Co with its spin aligned
along and against a magnetic field
• If parity were conserved, reflecting the spatial
coordinates (by preparing 60Co as spin down
instead of spin up) shouldn’t change the β angular
distribution…but it did
• This showed weak transitions don’t conserve parity

nagualdesign
28
Further Reading
• Chapter 8: Modern Nuclear Chemistry (Loveland, Morrissey, Seaborg)
• Chapter 7: Nuclear & Particle Physics (B.R. Martin)
• Chapter 14, Section 19: Quantum Mechanics for Engineers (L. van Dommelen)
• Chapter 15: Introduction to Special Relativity, Quantum Mechanics, and Nuclear Physics for
Nuclear Engineers (A. Bielajew)
• Chapter 4: Lecture Notes in Nuclear Structure Physics (B.A. Brown)
• Chapter 17: The Atomic Nucleus (R. Evans)
• Chapter 16: Elementary Nuclear Theory (H. Bethe)

29