Interpreting Hubble tension with a cascade decaying

dark matter sector

Quan Zhou, Zixuan Xu and Sibo Zheng

Department of Physics, Chongqing University, Chongqing 401331, China

arXiv:2507.08687v1 [astro-ph.CO] 11 Jul 2025

Abstract

Hubble tension can be alleviated by altering either early- or late-time ΛCDM. So far,
extensive studies have shown that only early dark energy or ad hoc combinations
of those two-fold effects can reduce the tension to 3σ or lower. In this work, we
improve the later solution by considering a cascade decaying dark matter sector, where
a parent particle produces relativistic dark matter in the early Universe and the dark
matter subsequently decays to neutrinos in the late-time Universe. We parametrize
this model with four model parameters, carry out a Markov Chain Monte Carlo fit to
Planck 2018+BAO+Pantheon data sets, and compare it to Big Bang Nucleosynthesis
limits, neutrino flux bounds, and Planck and SDSS data about matter power spectra.
Within parameter regions reducing the Hubble tension to 3.2σ and compatible with the
existing constraints, the main phenomenological features include a larger S8 ∼ 0.84,
smaller ns ∼ 0.95 and larger matter power spectra relative to the ΛCDM, which are
left for future tests.
Contents
1 Introduction 1

2 Cascade decaying dark matter sector: parametrization 3

2.1 Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

3 MCMC analysis 6
3.1 Data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.2 Numerical results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

4 Complementary tests 10
4.1 BBN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.2 Neutrino flux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.3 Matter power spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

5 Conclusion 14

1 Introduction
Hubble parameter H is a basic cosmological observable characterizing the expansion speed of
Universe. Assuming ΛCDM, there exists a ∼ 5.8σ tension1 between an indirect measurement
from the Planck 2018 data [1] about Cosmic Microwave Background (CMB) reporting a
value of H0 = 67.36 ± 0.54 km s−1 Mpc−1 at the 68% confidence level (CL), and a direct
measurement from the local experiment SH0ES [2–5] giving H0 = 73.01±0.92 km s−1 Mpc−1 .
Since both measurements are in high precision, it is unlikely to remove this tension in terms
of systematical uncertainties. As a result, this so-called Hubble tension is a robust signal of
new physics beyond the ΛCDM cosmology.
Looking for a clue to reduce the Hubble tension, one has to keep in mind the following
angle precisely measured by Planck,
rs (z∗ )
θs = , (1)
DA (z∗ )
R∞ Rz
where rs (z) = z cs dz ′ /H(z ′ ) is the sound horizon and DA (z) = 0 dz ′ /H(z ′ ) is the angular
distance, with z∗ the redshift at recombination. A nearly fixed value of θs in eq.(1) implies
1
The tension level depends on an explicit criterion adopted. In this work, we follow the DAMP criterion
[13] except mentioned otherwise.

1
that the solution to Hubble tension can be either early- or late-time one, corresponding
to lowering rs (z∗ ) and increasing DA (z∗ ) respectively. For comprehensive reviews on these
solutions, see e.g, [6–8], showing that either a single early- or late-time effect is insufficient
to reduce the Hubble tension to acceptable level of 3σ except an early dark energy [9–11].
Resolving the Hubble tension seems to favor an attendance of both the early- and late-
time effect as illustrated by e.g, me + ΩK [8, 12], which reduces the tension to ∼ 3σ; see [8]
for more examples. Instead of ad hoc combinations of those two-fold effects, in this work
we investigate non-ΛCDM models where the two-fold effects take place simultaneously with
the following criteria:

• Only one component of ΛCDM such as radiation, dark energy and dark matter (DM)
is altered, in order to avoid ad hoc combination of the two-fold effects.

• Model realization is possible or even natural from a viewpoint of new physics beyond
Standard Model (SM).

Explicitly, we consider to replace the cold dark matter (CDM) component of ΛCDM with a
cascade decaying DM (CDDM), where a decay of parent particle produces relativistic DM in
the early-time Universe, and the DM decays rather than is absolutely stable in the late-time
Universe.
Previously, altering the CDM component to reduce the Hubble tension has been studied
in the literature. For the relativistic effect of non-CDM on H0 , refs. [14–16] have shown that
Big Bang nucleosynthesis (BBN) limits prohibit a reduction of the tension down to 3σ. For
the late-time decay of DM into (dark) photons,2 refs.[17–23] have shown that Planck data
only allows a mild reduction on the tension. In this work, we investigate whether a CDDM
with these two-fold effects can reduce the Hubble tension to 3σ or lower, without violating
existing constraints.
The rest of this paper is structured as follows. Sec.2 is devoted to parametrize the
CDDM, which differs from the existing studies in two key points: (i) while contributing to
the effective neutrino number Neff , the DM energy density evolves non-trivially with redshift
before the DM fluid becomes non-relativistic, and (ii) DM decaying to neutrinos rather than
the (dark) photons allows relatively smaller DM lifetime to trigger larger late-time effect.
We derive the relevant background and perturbation equations prepared for later numerical
analysis. In Sec.3 we carry out a Markov Chain Monte Carlo (MCMC) analysis of CDDM
model to the latest data sets, showing a reduction of the tension down to 3.2σ. In Sec.4
we place Big Bang Nucleosynthesis (BBN) limits from the early Universe, and neutrino flux
bounds and limits on matter power spectra from the late Universe, respectively, illustrating
that resolving the tension at 3.2σ is compatible with these constraints. Finally, we conclude
in Sec.5.
2
Late-time DM decaying into other SM particles has been used to explain the XENON1T anomaly [27]
and the recent neutrino event [28].

2
2 Cascade decaying dark matter sector: parametriza-
tion
In this work we consider the CDDM composed of two species χm , χM with mass m and M
respectively. The χM particle, which is unstable with lifetime τM , decays to produce χm in
the early Universe as
χM → χm + X, (τM ≤ 104 sec), (2)
where X denotes a SM final state. This decay induced χm particles are initially relativistic,
but later become non-relativistic prior to the matter dominated epoch due to cosmic expan-
sion. Afterward, the χm particles, serving as a cold DM with lifetime τm , decay into the SM
neutrinos in late-time Universe as
χm → ν + ν̄, (τm ≥ 100 Gyr). (3)
In the rest of this section, we will show how to parametrize this CDDM, collect the set of
independent model parameters, and derive background and perturbation equations for later
numerical analysis.

2.1 Production
Regarding the early-time decay in eq.(2), the input parameters are composed of
{ρM,0 , τM , M, m, mX }, (4)
where ρM,0 denotes the present-day χM energy density if it does not decay and mX the X
mass. In order to produce the relativistic χm particles through this decay, one has to assume
M >> m and M >> mX . In this situation the mass parameter mX in eq.(4) can be simply
neglected. Moreover, as χm serves as the DM after the matter-dominated Universe, the
present-day χm relic density has to accommodate the observed DM relic density if its does
not decay in the late-time Universe, which means that ωm,0 = Ωm h2 = (ρm,0 /ρc )h2 is nearly
fixed with ρc the critical energy density and ρm,0 satisfying
ρm,0 m
= . (5)
ρM,0 M
Eq.(5) implies that ρM,0 can be fixed by adjusting the mass parameters m and M . As a
result, in eq.(4) we are left with the following three independent parameters
{τM , M, m}. (6)

To see how the parameters in eq.(6) affect the value of H0 , we now derive the explicit
forms of ρm and ρM as function of time t. First, for the non-relativistic χM with decay its
energy density evolves as
dρM −1
+ 3HρM = −τM ρM , (7)
dt
3
which gives

ρM (t) = ρM,0 e−t/τM . (8)

Second, the unperturbed Boltzmann equation for χm distribution function fm is given by

[24]
−1
∂fm aρM,0 τM
= δ(τ − τq ) (9)
∂τ 4πM Hq 3
where τ is the conformal time, H the conformal Hubble rate, q = a(τq )pmax the comoving
momentum with pmax ≈ M/2 the decay induced momentum without suffering from the
cosmic expansion. Given the distribution function the energy density reads as [25]
Z
1
ρm (t) = 4 q 2 dqdΩϵf1 (q) (10)
a (t)
p
where ϵ = q 2 + m2 a2 (t) is the comoving energy. Substituting the solution of eq.(9) into
eq.(10) gives us the explicit form of ρm (t).
Using eq.(5), we can rewrite eq.(10) as [26]
s 2
−1 t 
ρm,0 τM M2 aD
Z
−tD /τM
ρm (t) ≈ e 1+ dtD , (11)
a3 (t) 0 4m2 a(t)
where the subscript “D” refers to decay. Compared to the time of radiation-matter equality
with teq ≈ 1.6 × 1012 sec, if the decay takes place at time tD late enough to let the aD /aeq -
term dominate over unity, then one finds ρm scales as ∼ a−4 , suggesting that χm behaves
as a relativistic particle.3 Under this circumstance, it contributes to the effective neutrino
number
τM M
r
∆Neff (teq ) ≈ 1.19 × , (12)
teq m
at the time of radiation-matter equality, where we have used aD /aeq ≈ (tD /teq )1/2 during
the radiation-dominated epoch. Eq.(12) shows that a large value of H0 can be obtained by
choosing a large τM and a large mass ratio of M/m. For example, with τM ∼ 104 sec and
M/m ∼ 103 , we have ∆Neff (teq ) ∼ 0.1.
The early-time improvement on H0 is however limited. First, the Hubble rate enhanced
at the early-time Universe leads to a smaller value of sound horizon rs (z∗ ) with z∗ the redshift
at recombination, which is subject to the CMB constraints as discussed in Sec.3. Second,
the early-time decay induced energy injection into the SM thermal bath affects the light
element abundances including D, 4 He and 7 Li, which is therefore constrained by the BBN
limits. As discussed in Sec.4.1, the BBN limits are stringent for τM > 104 but become weak
for smaller values of τM . In this sense, we restrict to the parameter range of τM ≤ 104 in the
following analysis.
3
This relativistic behavior can be also interpreted in terms of equation of state of χm , see [26] for details.

4
2.2 Decay
Regarding the late-time decay in eq.(3), the inputs are composed of

{Br, τm }, (13)

where Br is the branching ratio of this decay channel. If one simply chooses Br = 1 as
we adopt here, τm is the only model parameters controlling the late-times deviations from
ΛCDM.
Due to the χm decay, the background equation of ρm and ρν is modified as [29]

a′
ρ′m + 3 ρm = −aτm −1
ρm ,
a
a′
ρ′ν + 4 ρν = aτm
−1
ρm , (14)
a
respectively, where primes denote derivatives with respect to the conformal time.
Moreover, the matter perturbation δm changes in synchronous gauge as

h′
δm = − , (15)
2
with h one of the two scalar modes in this gauge. This decay induced changes in the
perturbation equation of neutrino energy density δν in synchronous gauge evolve as [23]
4 2 −1 ρm
δν′ + θν + h′ = aτm (δm − δν ),
3 3 ρν
k2 −1 ρm
θν′ − (δν − 4σν ) = −aτm θν ,
4 ρν
4 2 4 3 −1 ρm
σν′ − θν − h′ − η ′ + kFν3 = −aτm σν ,
15 15 5 10 ρν
′ k
Fνℓ + [(ℓ + 1)Fνℓ+1 − ℓFνℓ−1 ] = 0, ℓ ≥ 3, (16)
2ℓ + 1
by using [30], where k is the wavenumber and Fνℓ are defined in [25, 29].
Similar to the early-time improvement, the late-time improvement on H0 is also con-
strained. On one hand, neutrinos due to the late-time decay of χm contribute to neutrino
flux constrained by existing data as discussed in Sec.4.2. On the other hand, the late-time
decay of χm results in a suppression (increase) on the matter power spectra at small (large)
scales [23, 29], which is constrained by the existing data about matter power spectra as
discussed in Sec.4.3.

5
 Model parameters I II III IV
M/m 1000 11000 11000 11000
τM [sec] 1000 2500 2500 2500
τm [Gyr] 150 150 300 1000
m [GeV] [10−3 , 1] [10−3 , 1] [10−3 , 1] [10−3 , 1]

Table 1: Four CDDM models with the model parameters composed of {M, τM , m, τm }, all
of which are fixed except the DM mass m.

3 MCMC analysis
Implementing the background and linear perturbation equations in Sec.2 to the Boltzmann
solver CLASS [31, 32], with a couple of modified branches therein, we now use MontePython
[33] to carry out the MCMC analysis on the CDDM model.

3.1 Data sets

We fit the CDDM model in Sec.2 to the following data sets.

• Planck: we constrain the CMB TT, TE, EE + lowE + lensing power spectra via
[34, 57].

• BAO: we consider the data from 6dFGS [36], BOSSDR7 [37] and DR12 [38], and
eBOSS DR16 [39].

• Pantheon: we impose the SN Ia luminosity data from [40].

To be clarify, we will not take into account the LSS data [41, 42] following the argument
of [8]. Given the inapplicability of the halofit prescription for calculating non-linear matter
power spectra in the late-time decaying DM models [29], we conservatively exclude the full
P (k) dataset from CFHTLenS [41] entering into the non-linear scales up to k ∼ 5 h/Mpc.
In practice, adding the LSS data to the above data sets can give rise to a larger best-fit value
of H0 .

3.2 Numerical results

We now present the results of four CDDM models I-IV as shown in Table.1, where all of
the model parameters are fixed except the DM mass m.
In terms of MCMC fit of the models I-IV to the Planck 2018+BAO+Pantheon data
sets, we show a short chain of parameters including H0 in fig.1, the best-fit values of the

6
 I
II
III
IV
ΛCDM

0.32
Ωm

0.3

0.28

0.87
S8

0.84

0.81

67.6 69.2 70.8 0.28 0.3 0.32 0.81 0.84 0.87

H0 Ωm S8

Figure 1: A short chain of MCMC fit of the CDDM models I-IV in Table.1 to the Planck
2018+BAO+Pantheon data sets, as compared to the ΛCDM (in gray), where H0 is in units
of km s−1 Mpc−1 .

Cosmological parameters I II III IV

102 ωb 2.238629 2.288454 2.298523 2.283133
ini
ωdm 0.1209978 0.1236136 0.1219745 0.1232426
100θs 1.041888 1.040730 1.040705 1.040793
ln (1010 As ) 3.055174 2.999885 3.000704 3.000614
ns 0.9643806 0.9434179 0.9460613 0.9422062
τreio 0.05683007 0.05249539 0.05459009 0.0539924
H0 68.10504 69.90806 70.26125 69.54282
S8 0.8042742 0.8382538 0.8443614 0.8746538
Ωm 0.2862790 0.2779738 0.2827627 0.2986182
rdrag 146.8210 144.5234 144.8464 144.6760

Table 2: The best-fit values of the cosmological parameters in the CDDM models I-IV with
respect to the Planck 2018+BAO+Pantheon data sets, where rdrag is units of Mpc.

7
cosmological parameters in Table.2, and a complete chain of parameters in fig.2. Several
comments are in order.

• The CDDM model I mimics the late-time effect, with the early-time effect negligible.
In this case with τm = 150 Gyr, one obtains the best-fit value of H0 = 68.10 km
s−1 Mpc−1 as shown in Table.2, which can be further increased by adopting a smaller
τm . However, such small value of τm violates the limits on matter power spectra as
discussed in Sec.4.3.

• The CDDM model II turns on both the early- and late-time effect. The model param-
eters related to the early-time effect in the model II are the same as in the model I.
As seen in Table.2, the best-fit value of H0 obviously increases to 69.91 km s−1 Mpc−1 ,
indicating the necessity of the two-fold effect in order to reduce the Hubble tension
down to 3σ.

• The necessity of the two-fold effect is further verified by the CDDM model III, which
differs from the model II by replacing τm = 150 Gyr with τm = 300 Gyr, giving a
value of H0 = 70.26 km s−1 Mpc−1 .

• The CDDM model IV mimics the same early-time effect as in the model III, but
turns off the late-time effect by choosing a large value of τm = 103 Gyr. In this case,
the best-fit value of H0 decreases from 70.26 to 69.54 in units of km s−1 Mpc−1 . While
such a decline in H0 can be compensated by choosing a larger values of τM > 104 sec,
which however violates the BBN limits as discussed in Sec.4.1.

As mentioned in the Introduction, we have followed the DMAP criterion to estimate the
tension level, which defines

∆χ2 = χ2CDDM,D+SH0ES − χ2CDDM,D , (17)

with D denoting the data sets of Planck 2018+BAO+Pantheon and SH0ES referring to the
SH0ES data set with H0 = 73.2 ± 0.9 km s−1 Mpc−1 . According to Table.2, we find that
the explicit tension level is {5.42, 3.19, 3.21, 3.48}σ with respect to the CDDM model I-IV ,
respectively. In this sense, the CDDM models II and III improve over the ad hoc two-fold
solutions such as me + ΩK [8, 12], which reduce the tension to the same order as the CDDM
but lack model interpretations.
As shown by II and III, the CDDM model within the parameter regions with respect
to M/m ∼ 1.1 × 104 , τM ∼ 2500 sec and τm ∼ 150 − 300 Gyr give a larger S8 ∼ 0.84 and
a smaller ns ∼ 0.95 compared to the ΛCDM, in order to compensate the two-fold effects
on the MCMC fit of the cosmological parameters to the Planck 2018+BAO+Pantheon data
sets.

8
 I
II
III
IV
ΛCDM
0.124
ωdm

0.121

0.118

1.042
100 θ s

1.041

1.04
ln(1010As)

3.08

3.03

2.98

0.97
ns

0.955

0.94

0.08
τ reio

0.06

0.04
log(m/GeV )

-0.5

-1.5

-2.5

70.8
H0

69.2

67.6

0.32
Ωm

0.3

0.28

147
rdrag

146
145

0.87
S8

0.84
0.81

2.22 2.27 2.32 0.118 0.121 0.124 1.04 1.041 1.042 2.98 3.03 3.08 0.94 0.955 0.97 0.04 0.06 0.08 -2.5 -1.5 -0.5 67.6 69.2 70.8 0.28 0.3 0.32 145 146 147 0.810.840.87
100 ωb ωdm 100 θs ln(1010As) ns τ reio log(m/GeV ) H0 Ωm rdrag S8

Figure 2: Same as in fig.1 but with the complete chain of the MCMC fit.

9
4 Complementary tests
In this section we place complementary constraints on the parameter regions of CDDM
model reducing the Hubble tension.

4.1 BBN
An electromagnetic energy release [43, 44] due to the decay of χM during the epoch of BBN
can alter the relic densities of light elements. As a result, the measured relic densities of
light elements place constraints [45, 46] on the χM decay lifetime and the electromagnetic
energy release parameter ζEM defined as

ζEM = ϵEM YM , (18)

where ϵEM is the initial electromagnetic energy released in each χM decay in eq.(2), and
YM = nM /nγ is the number density of χM before they decay, normalized to the number
density of background photons nγ .
In the situation with M >> m and M >> mX considered here, ϵEM ≈ M/2 for X =
{ℓ, γ} in eq.(2) with ℓ the SM leptons.4 Inserting the value of ϵEM into eq.(18) gives
 
−9 M
ζEM ≈ 3 × 10 GeV , (19)
m
where YM ≈ Ωdm ρc /(mnγ,0 ) has been used.
Fig.3 shows the BBN constraints on the plane of (τM , M/m) for the CDDM models I-
IV in Table.1, where eq.(19) has been used to transfer the BBN limits on ζEM to M/m
for an explicit τM . In this figure, the shaded regions are excluded by D/H < 1.3 × 10−5
and 7 Li/H < 0.9 × 10−10 [45] and the relativistic criterion M/m ≥ 10 adopted here. All
of the four models locate in the region being consistent with the BBN constraints. To
see which region being more capable of reducing the Hubble tension, we show contours
of ∆Neff = {0.05, 0.1, 0.15} (in dotted) at the time of radiation-matter equality therein.
Compared to these curves, ∆Neff (teq ) ≥ 0.1 can be achieved in the CDDM models II-IV ,
verifying that the early-time effects in these models are strong enough to uplift the value of
H0 without violating the BBN limits.

4.2 Neutrino flux

The late-time DM decay into neutrinos contributes to neutrino flux constrained by various
neutrino telescope data. The lower limit on τm increases from ∼ 1019 sec to ∼ 1027 sec with
respect to m from 10−3 GeV to 103 GeV, suggesting that smaller m range as considered in
Sec.3 is allowed to provide larger late-time effect.
4
If X is neutrino, the BBN limits can be relaxed, see e.g, [47].

10
 1 0 4
Ⅱ , Ⅲ , Ⅳ

7
L i lo w
1 0 3
Ⅰ

1 0 2
D N e ff = 0 .0 5
D N e ff = 0 .1 0
D N e ff = 0 .1 5
1 0 1 D lo w

R e la tiv is tic C rite rio n

1 0 0

1 0 2
1 0 3
1 0 4
1 0 5

Figure 3: The parameter regions of CDDM models I-IV compared to the BBN constraints
[45] and the relativistic criterion M/m ≥ 10 adopted. Here, the contours of ∆Neff (teq ) =
{0.05, 0.1, 0.15} at the time of radiation-matter equality are shown in dotted.

Fig.4 shows the planes of (m, τm ) extracted from fig.2 for the CDDM models I-IV , which
are compared to the existing bounds [48] including Borexino [49], KamLAND [50], and SK-
ν̄e [51], SK-Olivares [52] and SK atm [53]. As seen in this figure, the upper bound on τm
depends on the value of m. Specifically, τm ≤ 1019 sec allowing large late-time effect is
allowed in the mass range of m ≤ 5 MeV. In contrast, for m ≥ 8 × 10−3 (10−2 ) GeV, τm
has to be larger than ∼ 1020 (1021 ) sec, implying that the late-time effect on H0 due to the
DM decay is negligible. In other words, the CDDM models II and III that help reduce the
Hubble tension to 3.2σ are compatible with the existing neutrino telescope data with m ≤ 5
MeV.

4.3 Matter power spectra

Apart from contributing to the neutrino flux, the late-time DM decay also results in devia-
tions in the matter power spectra Pm (k) from that of ΛCDM, depending on the value of τm .
These deviations can be probed by a variety of cosmological experiments. Following [54],
ref.[55] has showed the constraints on Pm (k) at z = 0 over a wide range of k in terms of the
Planck 2018 temperature, polarization, and lensing reconstruction power spectra [56, 57] and
a sample of luminous red galaxies from the SDSS seventh data release (DR7) [58], among
others.5
5
We do not include data points entering in the non-linear regime with k > 0.2 Mpc−1 , where large
non-linear effects on Pm (k) have to be taken into account.

11
 1 0 2 2

S K − νe

K a m L A N D

1 0 2 1

1 0 2 0

1 0 1 9 Ⅲ

Ⅰ& Ⅱ
B o re x in o S K -O liv a re s S K a tm

1 0 − 3
1 0 − 2
1 0 − 1
1 0 0

Figure 4: Neutrino flux constraints [48] on the CDDM models I-IV in the mass range of
m ≤ 1 GeV including Borexino [49], KamLAND [50], and SK-ν̄e [51], SK-Olivares [52] and
SK atm [53].

Fig.5 shows deviations in the matter power spectra Pm (k) at z = 0 from the values of
ΛCDM.

• At large scales with k < keq ≈ 0.01 Mpc−1 , the values of Pm (k) is increased by smaller
values of Ωm due to the late-time effect.

• At intermediate scales with keq < k < 0.2 Mpc−1 , the deviations in Pm (k) are affected
by both the early- and late-time effect. The early-time effect imposes a correlation
between H0 and Pm (k) as shown in [14, 16], which requires an obvious increase in
Pm (k) for a larger H0 compared to the ΛCDM. In contrast, the late-time effect leads
to the smaller Ωm suppressing the growth factor D(a) [29] relative to the ΛCDM.
A combination of these two-fold effects finally gives rise to a larger Pm (k) at the
intermediate scales.

As a result, the sign of deviations ∆Pm = Pm (k) − Pm,ΛCDM (k) as shown in the bottom panel
of fig.5, is always positive over the whole range of k considered, serving a distinctive signal
of CDDM models, apart from those previously mentioned in Sec.3.2.

12
 104
Pm (k) [(Mpc/h) 3 ]

103

I
102 II
III
IV
CDM
101 Planck 2018 TT
Planck 2018 EE
Planck 2018
(k Mpc) 1.2 ∆Pm (k) [(Mpc/h) 3 ]

SDSS DR7 LRG

100
50
0
50
10-4 10-3 10-2 10-1
k [h/Mpc]
Figure 5: Top: Data points [55] collected from Planck 2018 CMB data and SDSS DR7 LRG,
compared to the matter power spectra Pm (k) of best-fit Planck 2018 ΛCDM model (in solid
black line) and of CDDM models I-IV in Table.1. Bottom: deviations of the CDDM models
from the ΛCDM compared to the data.

13
5 Conclusion
Hubble tension favors an early-time and/or late-time modification to the ΛCDM. So far,
reducing this tension to 3σ or lower can be achieved only by the early dark energy or ad hoc
combinations of the two-fold effects. In this work we have improved the later solution by
the CDDM model, in the sense that it reduces the tension to ∼ 3σ the same level as in ad
hoc combinations of the two-fold effects but is more realistic from the perspective of model
building.
We have made a systematic phenomenological analysis on the CDDM model. First,
we have parametrized the model with four model parameters composed of {M, m, τM , τm }.
After implementing the background and linear perturbation equations to CLASS, we have
made the MCMC fit of this model to the Planck 2018+BAO+Pantheon data sets. We
have shown that the parameter regions with respect to M/m ∼ 1.1 × 104 , τM ∼ 2500 sec,
τm ∼ 150 − 300 Gyr and m ≤ 5 MeV can reduce the tension to 3.2σ, without violating
the existing constraints from the BBN, neutrino flux bounds, and Planck and SDSS data
on the matter power spectra. Therein the main phenomenological features include a larger
S8 ∼ 0.84, smaller ns ∼ 0.95 and larger Pm (k) in the range of k ∼ 10−4 − 10−1 h/Mpc
compared to the ΛCDM, which are left for future tests.
There are a few interesting points not explored. For example, only the CDDM phe-
nomenology has been addressed, it is interesting to investigate model realizations of CDDM
models II and III. Moreover, implications of the smaller ns ∼ 0.95 to various inflation
models also deserve a detailed study.

References
[1] N. Aghanim et al. [Planck], Astron. Astrophys. 641, A6 (2020) [erratum: Astron. As-
trophys. 652, C4 (2021)], [arXiv:1807.06209 [astro-ph.CO]].

[2] A. G. Riess, S. Casertano, W. Yuan, L. M. Macri and D. Scolnic, Astrophys. J. 876,

no.1, 85 (2019), [arXiv:1903.07603 [astro-ph.CO]].

[3] A. G. Riess, et al. Astrophys. J. Lett. 908 (2021) no.1, L6, [arXiv:2012.08534 [astro-
ph.CO]].

[4] A. G. Riess, W. Yuan, L. M. Macri, D. Scolnic, D. Brout, S. Casertano, D. O. Jones,

Y. Murakami, L. Breuval and T. G. Brink, et al. Astrophys. J. Lett. 934, no.1, L7
(2022), [arXiv:2112.04510 [astro-ph.CO]].

[5] Y. S. Murakami, A. G. Riess, B. E. Stahl, W. D. Kenworthy, D. M. A. Pluck, A. Ma-

coretta, D. Brout, D. O. Jones, D. M. Scolnic and A. V. Filippenko, JCAP 11, 046
(2023), [arXiv:2306.00070 [astro-ph.CO]].

14
 [6] E. Di Valentino, O. Mena, S. Pan, L. Visinelli, W. Yang, A. Melchiorri, D. F. Mota,
A. G. Riess and J. Silk, Class. Quant. Grav. 38, no.15, 153001 (2021), [arXiv:2103.01183
[astro-ph.CO]].

[7] N. Schöneberg, G. Franco Abellán, A. Pérez Sánchez, S. J. Witte, V. Poulin and J. Les-
gourgues, Phys. Rept. 984, 1-55 (2022), [arXiv:2107.10291 [astro-ph.CO]].

[8] A. R. Khalife, M. B. Zanjani, S. Galli, S. Günther, J. Lesgourgues and K. Benabed,

JCAP 04, 059 (2024), [arXiv:2312.09814 [astro-ph.CO]].

[9] T. Karwal and M. Kamionkowski, Phys. Rev. D 94, no.10, 103523 (2016),
[arXiv:1608.01309 [astro-ph.CO]].

[10] E. Mörtsell and S. Dhawan, JCAP 09, 025 (2018), [arXiv:1801.07260 [astro-ph.CO]].

[11] V. Poulin, T. L. Smith, T. Karwal and M. Kamionkowski, Phys. Rev. Lett. 122, no.22,
221301 (2019), [arXiv:1811.04083 [astro-ph.CO]].

[12] T. Sekiguchi and T. Takahashi, Phys. Rev. D 103, no.8, 083507 (2021),
[arXiv:2007.03381 [astro-ph.CO]].

[13] M. Raveri and W. Hu, Phys. Rev. D 99, no.4, 043506 (2019), [arXiv:1806.04649 [astro-
ph.CO]].

[14] N. Blinov, C. Keith and D. Hooper, JCAP 06, 005 (2020), [arXiv:2004.06114 [astro-
ph.CO]].

[15] A. S. de Jesus, N. Pinto-Neto, F. S. Queiroz, J. Silk and D. R. da Silva, Eur. Phys. J.

C 83, no.3, 203 (2023), [arXiv:2212.13272 [hep-ph]].

[16] S. S. da Costa, D. R. da Silva, Á. S. de Jesus, N. Pinto-Neto and F. S. Queiroz, JCAP

04, 035 (2024), [arXiv:2311.07420 [astro-ph.CO]].

[17] K. Vattis, S. M. Koushiappas and A. Loeb, Phys. Rev. D 99, no.12, 121302 (2019),
[arXiv:1903.06220 [astro-ph.CO]].

[18] K. L. Pandey, T. Karwal and S. Das, JCAP 07, 026 (2020), [arXiv:1902.10636 [astro-
ph.CO]].

[19] S. J. Clark, K. Vattis and S. M. Koushiappas, Phys. Rev. D 103, no.4, 043014 (2021),
[arXiv:2006.03678 [astro-ph.CO]].

[20] G. Franco Abellán, R. Murgia and V. Poulin, Phys. Rev. D 104, no.12, 12 (2021),
[arXiv:2102.12498 [astro-ph.CO]].

[21] T. Simon, G. Franco Abellán, P. Du, V. Poulin and Y. Tsai, [arXiv:2203.07440 [astro-
ph.CO]].

15
[22] S. Alvi, T. Brinckmann, M. Gerbino, M. Lattanzi and L. Pagano, [arXiv:2205.05636
[astro-ph.CO]].

[23] Z. Xu, S. Xu, R. Zhang and S. Zheng, JHEP 09, 182 (2023), [arXiv:2304.02904 [hep-ph]].

[24] S. Aoyama, T. Sekiguchi, K. Ichiki and N. Sugiyama, JCAP 07, 021 (2014),
[arXiv:1402.2972 [astro-ph.CO]].

[25] C. P. Ma and E. Bertschinger, Astrophys. J. 455, 7-25 (1995), [arXiv:astro-ph/9506072

[astro-ph]].

[26] G. Blackadder and S. M. Koushiappas, Phys. Rev. D 90, no.10, 103527 (2014),
[arXiv:1410.0683 [astro-ph.CO]].

[27] S. Xu and S. Zheng, Eur. Phys. J. C 81, no.5, 446 (2021), [arXiv:2012.10827 [hep-ph]].

[28] K. Kohri, P. K. Paul and N. Sahu, [arXiv:2503.04464 [hep-ph]].

[29] V. Poulin, P. D. Serpico and J. Lesgourgues, JCAP 08, 036 (2016), [arXiv:1606.02073
[astro-ph.CO]].

[30] M. Kaplinghat, R. E. Lopez, S. Dodelson and R. J. Scherrer, Phys. Rev. D 60, 123508
(1999), [arXiv:astro-ph/9907388 [astro-ph]].

[31] J. Lesgourgues, [arXiv:1104.2932 [astro-ph.IM]].

[32] D. Blas, J. Lesgourgues and T. Tram, JCAP 07, 034 (2011), [arXiv:1104.2933 [astro-
ph.CO]].

[33] T. Brinckmann and J. Lesgourgues, Phys. Dark Univ. 24, 100260 (2019),
[arXiv:1804.07261 [astro-ph.CO]].

[34] N. Aghanim et al. [Planck], Astron. Astrophys. 641, A5 (2020), [arXiv:1907.12875

[astro-ph.CO]].

[35] N. Aghanim et al. [Planck], Astron. Astrophys. 641, A8 (2020), [arXiv:1807.06210

[astro-ph.CO]].

[36] F. Beutler, et al. Mon. Not. Roy. Astron. Soc. 416, 3017-3032 (2011), [arXiv:1106.3366
[astro-ph.CO]].

[37] A. J. Ross, et al. Mon. Not. Roy. Astron. Soc. 449, no.1, 835-847 (2015),
[arXiv:1409.3242 [astro-ph.CO]].

[38] S. Alam et al. [BOSS], Mon. Not. Roy. Astron. Soc. 470, no.3, 2617-2652 (2017),
[arXiv:1607.03155 [astro-ph.CO]].

[39] S. Alam et al. [eBOSS], Phys. Rev. D 103, no.8, 083533 (2021), [arXiv:2007.08991
[astro-ph.CO]].

16
[40] D. M. Scolnic, D. O. Jones, A. Rest, et al. Astrophys. J. 859, no.2, 101 (2018).
[arXiv:1710.00845 [astro-ph.CO]].

[41] C. Heymans, et al. Mon. Not. Roy. Astron. Soc. 432, 2433 (2013), [arXiv:1303.1808
[astro-ph.CO]].

[42] P. A. R. Ade et al. [Planck], Astron. Astrophys. 571, A20 (2014), [arXiv:1303.5080
[astro-ph.CO]].

[43] E. Holtmann, M. Kawasaki, K. Kohri and T. Moroi, Phys. Rev. D 60, 023506 (1999),
[arXiv:hep-ph/9805405 [hep-ph]].

[44] M. Kawasaki, K. Kohri and T. Moroi, Phys. Rev. D 63, 103502 (2001), [arXiv:hep-
ph/0012279 [hep-ph]].

[45] R. H. Cyburt, J. R. Ellis, B. D. Fields and K. A. Olive, Phys. Rev. D 67, 103521 (2003),
[arXiv:astro-ph/0211258 [astro-ph]].

[46] J. L. Feng, A. Rajaraman and F. Takayama, Phys. Rev. D 68, 063504 (2003),
[arXiv:hep-ph/0306024 [hep-ph]].

[47] S. Bianco, P. F. Depta, J. Frerick, T. Hambye, M. Hufnagel and K. Schmidt-Hoberg,

[arXiv:2505.01492 [hep-ph]].

[48] C. A. Argüelles, D. Delgado, A. Friedlander, A. Kheirandish, I. Safa, A. C. Vincent and

H. White, Phys. Rev. D 108, no.12, 123021 (2023), [arXiv:2210.01303 [hep-ph]].

[49] M. Agostini et al. [Borexino], Astropart. Phys. 125, 102509 (2021), [arXiv:1909.02422
[hep-ex]].

[50] S. Abe et al. [KamLAND], Astrophys. J. 925, no.1, 14 (2022), [arXiv:2108.08527 [astro-
ph.HE]].

[51] W. Linyan, Experimental Studies on Low Energy Electron Antineutrinos and Related
Physics, Ph.D. thesis, Tsinghua University (2018).

[52] A. Olivares-Del Campo, C. Bœhm, S. Palomares-Ruiz and S. Pascoli, Phys. Rev. D 97,
no.7, 075039 (2018), [arXiv:1711.05283 [hep-ph]].

[53] E. Richard et al. [Super-Kamiokande], Phys. Rev. D 94, no.5, 052001 (2016),
[arXiv:1510.08127 [hep-ex]].

[54] M. Tegmark and M. Zaldarriaga, Phys. Rev. D 66, 103508 (2002), [arXiv:astro-
ph/0207047 [astro-ph]].

[55] S. Chabanier, M. Millea and N. Palanque-Delabrouille, Mon. Not. Roy. Astron. Soc.
489, no.2, 2247-2253 (2019), [arXiv:1905.08103 [astro-ph.CO]].

17
[56] N. Aghanim et al. [Planck], Astron. Astrophys. 641, A1 (2020), [arXiv:1807.06205
[astro-ph.CO]].

[57] N. Aghanim et al. [Planck], Astron. Astrophys. 641, A8 (2020), [arXiv:1807.06210

[astro-ph.CO]].

[58] B. A. Reid, W. J. Percival, D. J. Eisenstein, L. Verde, D. N. Spergel, R. A. Skibba,

N. A. Bahcall, T. Budavari, M. Fukugita and J. R. Gott, et al. Mon. Not. Roy. Astron.
Soc. 404, 60-85 (2010), [arXiv:0907.1659 [astro-ph.CO]].

18