WSU-HEP-2402

May 8, 2024

Quasielastic Lepton-Nucleus Scattering and the

Correlated Fermi Gas Model
arXiv:2405.05342v1 [hep-ph] 8 May 2024

Bhubanjyoti Bhattacharya(a,b) , Sam Carey(b) , Erez O. Cohen(c) , Gil Paz(b)

(a)
Department of Natural Sciences,
Lawrence Technological University, Southfield, Michigan 48075, USA
(b)
Department of Physics and Astronomy,
Wayne State University, Detroit, Michigan 48201, USA
(c)
Physics Department,
Nuclear Research Center - Negev,
P.O. Box 9001, 84190 Beer-Sheva, Israel

Abstract

The neutrino research program in the coming decades will require improved precision. A
major source of uncertainty is the interaction of neutrinos with nuclei that serve as targets
for such experiments. Broadly speaking, this interaction often depends, e.g., for charge-
current quasi-elastic scattering, on the combination of “nucleon physics”, expressed by
form factors, and “nuclear physics”, expressed by a nuclear model. It is important to get
a good handle on both. We present a fully analytic implementation of the Correlated
Fermi Gas Model for electron-nucleus and charge-current quasi-elastic neutrino-nucleus
scattering. The implementation is used to compare separately form factors and nuclear
model effects for both electron-carbon and neutrino-carbon scattering data.
1 Introduction
Current and upcoming experiments involving lepton-nucleon scattering aim to precisely mea-
sure parameters in the Standard Model (SM) Lagrangian that describe leptonic interactions
as well as to uncover non-standard neutrino interactions (NSI). Such precision measurements
require a better control of systematic uncertainties in lepton-nucleus interactions. The cross-
section for a charged lepton or a charge-current quasi-elastic (CCQE) neutrino scattering off
a nucleus is determined by folding the lepton-quark interaction twice. Each folding gives rise
to a source of systematic uncertainty. The first involves nucleon form factors needed to fold
the lepton-quark interaction into the lepton-nucleon interaction. The result is the scattering
cross-section on a single nucleon. The second folding is needed in going from the nucleon to
the nuclear level, which involves a nuclear model and results in the cross-section on a nucleus
consisting of multiple nucleons. Ideally, we would like to have separately a good control for
each source of uncertainty.
It can be argued that the majority of the research in this field has focused on the nuclear
level, where the nucleon level has received less study. This seems plausible for uncertainties
arising from the electromagnetic form factors, since they can be extracted from electron-
nucleon scattering, see, for example, the parameterizations in [1] and [2]. Uncertainties from
the axial form factor are much harder to control. Historically, a dipole form factor was assumed
for the axial form factor. For example, the comparative study [3], based on [4], compared six
nuclear models: Benhar’s spectral function with and without the final-state interactions [5–8],
the Valencia spectral function [9–12], the GiBUU model [13, 14], and the local and global
Fermi gas models, all using the dipole model for the axial form factor [4].
The dipole model is not motivated from first principles and its usage can underestimate
the uncertainties. This issue was highlighted in the MiniBooNE analysis [15] that seemed
to require a higher axial mass than the preceding world-average value to explain the data.
In [16, 17] a z-expansion based parametrization was introduced for the axial form factor.
Assuming a definite nuclear model, namely, the relativistic Fermi gas (RFG) model [18], an
axial mass was extracted from the MiniBooNE data using z-expansion based parametrization
that was consistent with the preceding world-average values.
In the RFG model a nucleon within a nucleus is assumed to occupy momentum states
only up to the Fermi momentum. Data from the last two decades have shown that this is
not the case. Inclusive measurements involving large momentum transfer (Q2 > 1.5 GeV2 ),
showed that approximately 20% of nucleons have momentum greater than the Fermi momen-
tum [19–21]. Almost all high-momentum nucleons appear in short-range correlated (SRC)
pairs, predominantly neutron-proton ones, and contribute most of the kinetic energy carried
by nucleons in nuclei. Exclusive measurements on 12 C and 4 He have led to direct observations
of such SRC pairs [22–26]. An extension of the RFG model that takes into account the high
momentum tail beyond the nuclear Fermi momentum due to SRC pairing is the Correlated
Fermi Gas (CFG) model suggested in [27]. In the CFG model, the nucleon momentum dis-
tribution is constant below the Fermi momentum and has a small high-momentum tail above
the Fermi momentum.
The CFG model has been used to study the equation of state of nucleonic stars and its
effects on stellar properties such as maximum mass and particle fraction [28–31]. The CFG

1
model has also been implemented to study the EMC effect, where the quark-gluon structure of
a nucleon bound in an atomic nucleus is modified by the surrounding nucleons [32]. One goal of
this paper is to present a fully analytic implementation of the CFG model for electron-nucleus
and CCQE neutrino-nucleus scattering1 .
A second goal of this paper is to use this implementation to compare separately form factors
and nuclear model effects for both electron-carbon and neutrino-carbon scattering data. Lack
of such a control is both a long standing issue and a current topic. The first arises from
the fact that many explanations for the MiniBooNE data [15] assumed the dipole axial form
factor model and only tried to modify the nuclear effects. The studies of [16, 17] fixed the
nuclear model and allowed for a flexible axial form factor. The lack of control is a current
topic since in the last few years many extractions of the axial from factors from experimental
data [34, 35] and lattice QCD [36–40] became available. A better control on the axial form
factor will potentially allow to separate the nucleon and nuclear effects.
The structure of the paper is as follows. In section 2 we review the relation between
the cross section and the nuclear tensor, and present the construction of the nuclear tensor
using the nucleon form factors for the RFG and CFG models. In section 3 we present the
analytic implementation of the CFG model for lepton-nucleus scattering. We compare the
CFG predictions to electron-carbon data in section 4.1 and neutrino MiniBooNE data in
section 4.2. For both we consider separately form factor effects, by comparing different form
factor parameterizations, and nuclear effects, by comparing the RFG and CFG models. We
summarize our findings in section 5. Some more technical details of the paper are relegated
to the appendices.

2 Models
2.1 Lepton-nucleus cross section
We study processes in which an incoming lepton, neutral or charged, with four-momentum
k that scatters to a lepton with four-momentum k ′ off a target with four-momentum pT .
The differential cross section is expressed in terms of the nuclear tensor Wµν defined below.
Considering the possibility of both vector and axial currents, we can decompose Wµν as a
linear combination of scalar functions Wi as [16]

pTµ pTν iϵµνρσ pρT pσT qµ q ν pTµ qν + qµ pTν

Wµν = −gµν W1 + W 2 − W 3 + W 4 + W5 , (1)
m2T 2m2T m2T 2m2T
1
See [33] for a related implementation in GENIE v3.2 .

2
where q = k − k ′ and mT is the target mass. Defining Eℓ ≡ k ′ 0 and P⃗ℓ ≡ ⃗k ′ the energy and
momentum of the final state lepton, the (anti-)neutrino-nucleus cross section is [16]
(
ν
dσnuclear G2 |P⃗ℓ |
= F2 2(Eℓ − |P⃗ℓ | cos θℓ ) W1 + (Eℓ + |P⃗ℓ | cos θℓ )W2
dEℓ d cos θℓ 16π mT
)
1 h i m2 m2
± (Eℓ − |P⃗ℓ | cos θℓ )(Eν + Eℓ ) − m2ℓ W3 + 2ℓ (Eℓ − |P⃗ℓ | cos θℓ )W4 − ℓ W5 , (2)
mT mT mT

where the upper (lower) sign is for neutrino (anti-neutrino) scattering.

Neglecting the electron mass, the electron-nucleus cross section is
e
dσnuclear α2 E 2 h i
= 4 ℓ 2W1 (1 − cos θℓ ) + W2 (1 + cos θℓ ) . (3)
dEℓ d cos θℓ 2q mT
The nuclear tensor Wµν is formally related to matrix elements of the vector and axial current
between the initial and final nuclear states. We can relate it to the single nucleon tensor,
Hµν for both the RFG and CFG models by using a “statistical” approach presented explicitly
in [16]. Denote by ni (p) the momentum distribution of the initial nucleon momentum p. The
final-state nucleon momentum p′ phase space is limited by a factor of [1 − nf (p′ )] from the
Fermi-Dirac statistics. The nuclear tensor Wµν can be expressed in terms of the nucleon tensor
Hµν as follows,

d3 p mT d3 p′
Z Z
Wµν = 2V ni (p) (2π)4 δ 4 (p − p′ + q)Hµν [1 − nf (p′ )] , (4)
(2π)3 Ep (2π)3 2Ep′

see [16] for details. Accounting for two possible spin states, the number of nucleons N of a
certain type, namely, protons or neutrons, determines the normalization factor V as

d3 p
Z
N= 2V ni (p). (5)
(2π)3

In the following we use N = A/2, where A is the total number of nucleons.

Since we will not consider the kinematics of the final state nucleon in this paper, we can
integrate over d3 p′ using the three-momentum delta function. We also incorporate a binding
energy, ϵb , by the replacements

p0 → ϵp ≡ Ep − ϵb , p′0 → ϵ′p′ ≡ Ep′ , (6)

p
where Ep ≡ m2N + |p|2 . The resulting relation between Wµν and Hµν is [16]
R
Wµν ≡ d3 p f (p, q 0 , q)Hµν (ϵp , p; q 0 , q) , (7)

with
0 mT V δ(ϵp − ϵ′p+q + q 0 )
f (p, q , q) = ni (p)[1 − nf (p + q)] . (8)
4π 2 ϵp ϵ′p+q

3
We discuss f (p, q 0 , q) for the RFG and CFG cases below. The nucleon tensor Hµν is given by

Hµν = Tr[(p/′ + mp′ )Γµ (q)(p/ + mp )Γ̄ν (q)] , (9)

where Γµ (q) is defined via the matrix element of the electromagnetic or charged weak current
Jµ as

⟨p′ |Jµ |p⟩ = ū(p′ )Γµ (q)u(p) . (10)

Γµ (q) can be expressed in term of form factors:

i qµ
Γµ (q) = γµ F1 (q 2 ) + σµν q ν F2 (q 2 ) + γµ γ5 FA (q 2 ) + γ5 FP (q 2 ) , (11)
2mN mN
where we assume time-reversal invariance. For neutrino scattering we also assume isospin
symmetry, that allows to relate the neutron-to-proton F1,2 to the electromagnetic form factors
F1,2 of the proton and neutron. For charged-lepton scattering, the nucleon mass is replaced
by the mass of the proton (mp ) or neutron (mn ). For neutrino scattering we assume isospin
symmetry and take mN = mp = mn . Based on these symmetries, Hµν can be decomposed as

pµ pν ϵµνρσ qµ qν (pµ qν + qµ pν )
Hµν = −gµν H1 + 2
H2 − i 2 pρ q σ H3 + 2 H4 + H5 . (12)
mN 2mN mN 2m2N

The Hi ’s are expressed in terms of the form factors Fi as [16]

H1 = 8m2N FA2 − 2q 2 (F1 + F2 )2 + FA2

 

H2 = H5 = 8m2N F12 + FA2 − 2q 2 F22 ,

H3 = −16m2N FA (F1 + F2 ) ,
q2
F22 + 4FP2 − 2m2N F22 − 4m2N (F1 F2 + 2FA FP ) .


H4 = − (13)
2
Combining these expressions, the cross-section is expressed by the nuclear tensor Wµν that
is a convolution of the single nucleon tensor Hµν and a nuclear model parameterized by the
initial and final nucleon momentum distributions.
We now discuss the nucleon momentum distributions for the RFG and the CFG models.

2.2 Relativistic Fermi Gas Model

In the RFG model the distributions of neutrons and protons are

ni (p) = θ(pF − |p|) , nf (p′ ) = θ(pF − |p′ |) , (14)

where pF is a parameter of the model. Using equation (5) we find that the normalization
factor V is
d3 p A 3π 2
Z
A
= 2V n i (p) =⇒ V = . (15)
2 (2π)3 2 p3F

4
From these, explicit expressions can be derived for Wi . In particular,
1
W1 = a1 H1 + (a2 − a3 )H2 ,
2
ω2 ω2
   
ω 1
W 2 = a4 + 2 a3 − 2 a5 + 1 − 2 (a2 − a3 ) H2 ,
|q| |q| 2 |q|
 
mT ω
W3 = a7 − a6 H3 ,
mN |q|
m2T m2N
 
mN
W4 = 2 a1 H4 + a6 H5 + (3a3 − a2 )H2 ,
mN |q| 2|q|2
   
mT ω mT ω
W5 = a7 − a6 H5 + 2a5 + (a2 − 3a3 ) H2 , (16)
mN |q| |q| |q|

where ω = q 0 and
|p|2
Z Z
3
a1 = d p f (p, q) , a2 = d3 p f (p, q) ,
m2N
(pz )2 ϵ2p
Z Z
3 3
a3 = d p f (p, q) 2 , a4 = d p f (p, q) 2 ,
mN mN
ϵp pz pz
Z Z
3
a5 = d p f (p, q) 2 , a6 = d3 p f (p, q) ,
mN mN
Z
ϵp
a7 = d3 p f (p, q) . (17)
mN
see for example, [16]. For the RFG model the ai functions can be expressed in terms of
three master integrals bi over the initial nucleon energy. The limits of these master integrals
are determined by conservation of three-momentum and energy. We will encounter similar
features in the CFG model.

2.3 Correlated Fermi Gas Model

We follow the model of [27] with ρ = ρ0 and change the momentum variable from k to p. The
momentum distribution there is given by

A0
 |p| ≤ pF
SRC 4
nSN M (p) = c0 pF /p pF ≤ |p| ≤ λpF (18)

0 |p| ≥ λpF ,


where c0 = 4.16 ± 0.95 and λ ≈ 2.75 ± 0.25. For ρ = ρ0 , A0 is given by

3π 2
   
1 c0
A0 = 3 1 − 1 − , (19)
pF λ π2

5
and determined by the normalization
d3 p SRC
Z
1=2 n (p). (20)
(2π)3 SN M
To obtain n for the CFG model we change the normalization to A/2. Thus we define nCFG (p)
to be  
1 c0


 1− 1− 2
≡ α0 |p| ≤ pF
 λ4 π



nCFG (p) = c0 p F α1 (21)
2
≡ 4
pF ≤ |p| ≤ λpF
3π p |p|




|p| ≥ λpF .

0
Taking V = (A/2) 3π 2 /p3F , nCFG (p) satisfies

d3 p
Z
A
= 2V nCFG (p) . (22)
2 (2π)3

1.0
λ = 2.50, c0 = 3.21
V λ = 2.75, c0 = 4.16

0.8 λ = 3.00, c0 = 5.11

0.6
ni

I III IV
0.4

0.2

PF
λPF
II
0.0
0.0 0.2 0.4 0.6 0.8
p (GeV)

Figure 1: nCFG (p) as a a function of the magnitude of the 3-momentum. The different lines
correspond to extremal variation of λ and c0 . The labels for the regions I-V are explained in the
text.

nCFG (p) is plotted in figure 1. It exhibits a depleted Fermi gas region and a correlated
high-momentum tail. The dashed and dotted lines correspond to the maximum and minimum