Axion Kinetic Misalignment

• Kinetic misalignment is a dark matter production mechanism where the axion field starts with a nonzero angular velocity, storing energy as rotational charge.
• It delays the onset of oscillations by allowing the field to traverse multiple periodic potential minima, decoupling relic abundance from the initial displacement.
• In the strong regime, fragmentation and resonance effects yield dense axion miniclusters, offering novel astrophysical probes for dark matter.

Kinetic misalignment, usually the axion kinetic misalignment mechanism, is a cosmological dark-matter production mechanism in which the axion field begins with a nonzero initial angular velocity rather than the zero-velocity initial condition of standard misalignment. In this framework the axion is the angular mode of a Peccei–Quinn (PQ) field, the early dynamics are organized by a conserved PQ charge density and its yield $Y_\theta$, and the field can traverse multiple minima of its periodic potential before becoming trapped. The onset of oscillations is therefore delayed relative to the conventional $3H \sim m_a(T)$ criterion, and the relic abundance is controlled primarily by the stored rotational charge rather than by an initial displacement angle alone (Co et al., 2019, Barman et al., 2021).

1. Field-theoretic definition

After spontaneous PQ symmetry breaking, the complex scalar is written in polar form as

$\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$

where $S$ is the radial saxion and $a$ is the axion, with $\theta \equiv a/f_a$. For homogeneous evolution, the axion equation of motion takes the standard damped form

$\ddot\theta+3H(T)\dot\theta+m^2(T)\sin\theta=0,$

with the periodic QCD potential conventionally written as $V_{\rm QCD}(a)=\chi(T)(1-\cos\theta)$ and $m(T)=\sqrt{\chi(T)}/f_a$ (Barman et al., 2021).

The defining change relative to standard misalignment is the initial condition. Standard misalignment assumes

$\dot\theta_i=0,$

whereas kinetic misalignment assumes

$3H \sim m_a(T)$0

The corresponding conserved angular charge can be written as

$3H \sim m_a(T)$1

in the broken-symmetry regime, or more generally as

$3H \sim m_a(T)$2

with $3H \sim m_a(T)$3 the entropy density (Barman et al., 2021, Yin et al., 11 Mar 2026).

This reformulation shifts the physical interpretation of the relic from “energy stored in an initial angle” to “energy stored in rotation in field space.” The original proposal explicitly compared the production of this angular motion to Affleck–Dine-like dynamics generated by early-universe PQ-breaking operators (Co et al., 2019).

2. Dynamical regimes and delayed trapping

The central dynamical question is whether the initial kinetic energy is large enough for the field to surmount the cosine barriers of the axion potential. In the original formulation, kinetic misalignment requires

$3H \sim m_a(T)$4

equivalently

$3H \sim m_a(T)$5

so that the field keeps rotating until the velocity redshifts to the trapping condition

$3H \sim m_a(T)$6

(Co et al., 2019).

Regime	Diagnostic condition	Characterization
Standard misalignment	$3H \sim m_a(T)$7; $3H \sim m_a(T)$8	Frozen field released by Hubble friction
Weak kinetic misalignment	$3H \sim m_a(T)$9; $\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$0	Effective initial condition modified, but onset not shifted
Strong kinetic misalignment	$\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$1; $\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$2	Barriers crossed repeatedly until delayed trapping

In the weak regime, the field initially behaves as a kination component with

$\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$3

but the oscillation temperature is unchanged. In the strong regime, the periodic potential is effectively irrelevant during the early rolling phase, oscillations begin only after the kinetic energy has sufficiently redshifted, and the late abundance becomes effectively independent of the initial angle $\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$4 (Barman et al., 2021).

A later ALP construction generalized the same logic by introducing a rapid transition from a tracking potential to a cosine potential during radiation domination. There the control parameter is

$\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$5

The three cases are $\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$6, immediate trapping without kinetic misalignment; $\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$7, onset near the top of the cosine potential; and $\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$8, a genuine kinetic-misalignment phase with $\Phi \sim \frac{1}{\sqrt{2}}(S+v_a)e^{i a/v_a},$9 and

$S$0

(Pérez-Poyatos, 30 Apr 2026).

3. Relic abundance and parameter space

The original analytic estimate expresses the final abundance in terms of the conserved charge yield,

$S$1

rather than in terms of a misalignment angle. A later generic-PQ-breaking treatment writes the same result numerically as

$S$2

with the factor of $S$3 in the axion yield traced to anharmonic effects near the top of the potential (Co et al., 2019, Yin et al., 11 Mar 2026).

This change in scaling is the reason kinetic misalignment opens parameter space that standard misalignment does not populate. Standard QCD-axion misalignment is usually associated with $S$4, whereas kinetic misalignment can efficiently operate for

$S$5

and is frequently discussed in the range

$S$6

(Co et al., 2019, Yin et al., 11 Mar 2026).

The strong-kinetic regime is especially distinctive because the relic abundance becomes linear in the conserved charge. In the UV-agnostic analysis,

$S$7

and the onset temperature can be delayed to $S$8 in the benchmark discussion, approximately independent of $S$9 once the observed dark matter abundance is imposed (Barman et al., 2021).

A plausible implication is that “axion mass” and “axion abundance” become much less tightly linked than in the conventional picture. This is the common thread behind the broad KMM literature: a conserved early-time angular momentum permits viable dark matter in regions that would otherwise be underabundant.

4. Microphysical realizations

Microphysically, kinetic misalignment requires a source of angular motion. The original proposal attributed this to explicit PQ-breaking operators active in the early universe, for example a higher-dimensional term such as

$a$0

which exerts a torque when the PQ field is displaced far from its vacuum value. As the field value later decreases, the explicit breaking becomes ineffective and the angular motion survives as approximately conserved PQ charge (Co et al., 2019).

Inflationary realizations embed that torque into a specific field-space geometry. In “Peccei-Quinn Inflation at the Pole,” the radial PQ field drives inflation near a pole in the kinetic term, while a PQ-violating angular potential generates a nonzero $a$1 during inflation; after inflation the approximately conserved Noether charge $a$2 carries this information into the reheating era (Lee et al., 2023). A DFSZ-type extension with two Higgs doublets and three right-handed neutrinos implements the same idea in a conformally coupled PQ sector, with successful inflationary predictions

$a$3

while the viable region is simultaneously constrained by reheating, non-restoration of PQ symmetry after reheating, and the axion quality problem (Lee et al., 2024).

UV-complete studies emphasize that the radial partner is indispensable. In explicit KSVZ-like constructions, the axion inherits kinetic energy from the radial mode, and damping of the radial-mode energy density is a necessary ingredient. One analysis identifies a nearly-quadratic radial potential, thermal damping from interactions in the plasma, and Higgs-portal interactions as especially natural ingredients, and argues that such implementations can place QCD axion dark matter in reach of IAXO while MADMAX, IAXO, and ALPS II can probe generic ALP dark matter from KMM (Eröncel et al., 2024).

Alternative realizations alter the origin of the kick rather than the late trapping criterion. “Trapped misalignment” uses a temperature-dependent $a$4 axion potential to force an early oscillation stage around the wrong minimum near $a$5; after the QCD transition, the velocity inherited from that stage can satisfy

$a$6

and dynamically source a kinetic-misalignment phase without the extra PQ-breaking operators used in the original proposal (Luzio et al., 2021). A post-inflationary variant instead invokes a stiff era with negative Ricci scalar, which drives a non-minimally coupled radial mode tachyonically to large amplitude; higher-dimensional $a$7-breaking operators then generate distinct charges in different domains, and the later axion potential converts the domain charges into dark matter even when the net global charge vanishes (Morgante et al., 15 Dec 2025).

A separate inflation-linked quartic-potential analysis tied viable KMM/parametric-resonance dark matter directly to the inflationary Hubble scale $a$8 and reheating temperature $a$9, finding

$\theta \equiv a/f_a$0

and highlighting rare kaon decays and suppression of structure formation on small scales as characteristic consequences of the allowed region (Co et al., 2020).

5. Fragmentation, resonance, and small-scale structure

Kinetic misalignment is not only a delayed-trapping mechanism. Once the background field rolls through many minima, the homogeneous condensate itself can become unstable. Linear fluctuations satisfy

$\theta \equiv a/f_a$1

which has Floquet instability bands because the background $\theta \equiv a/f_a$2 moves periodically through the cosine potential. A detailed semi-analytic treatment concludes that in the kinetic-misalignment regime the axion field is almost always entirely fragmented, so the energy of the homogeneous mode is redistributed into modes near

$\theta \equiv a/f_a$3

rather than remaining in a pure zero mode (Eröncel et al., 2022).

That analysis divides the evolution into four regimes: standard misalignment, kinetic misalignment with weak fragmentation, fragmentation after trapping, and fragmentation before trapping. The transition to complete fragmentation is controlled by $\theta \equiv a/f_a$4, with representative values near $\theta \equiv a/f_a$5 or $\theta \equiv a/f_a$6 for the boundary between weak and complete fragmentation, and around $\theta \equiv a/f_a$7 for fragmentation before trapping in the temperature-dependent case discussed there (Eröncel et al., 2022). Despite this dramatic restructuring of the field, the final relic abundance is found to change only at the order-one level rather than by orders of magnitude.

Fragmentation also changes the astrophysical predictions of axion dark matter. In the pre-inflationary scenario, the moving background gives sizeable adiabatic fluctuations already during the early kination-like phase, and the later resonance amplifies them into a spectrum peaked near

$\theta \equiv a/f_a$8

This leads to compact axion halos or miniclusters denser than those produced in standard misalignment, large misalignment, or the post-inflationary scenario, with lensing, pulsar-timing signatures, and gravitational-wave diffraction identified as relevant probes of the resulting small-scale structure (Eröncel et al., 2022).

A common misconception is that KMM simply reproduces the homogeneous-condensate phenomenology of standard misalignment with a shifted onset time. The fragmentation literature argues that this is generally not the correct late-time description: the charge-driven background is often a transient stage preceding a strongly inhomogeneous axion field.

6. Constraints, variants, and present status

The same explicit PQ-breaking terms that generate the initial kick also create the principal consistency problem of the mechanism. Generic operator analyses show that these terms shift the axion minimum away from the CP-conserving point and therefore reintroduce the PQ quality problem. In that language the induced effective CP angle must satisfy the neutron-EDM bound, the same breaking produces CP-even scalar couplings and axion-mediated fifth forces, and the resulting nonstandard cosmological history typically contains a brief early matter-dominated phase followed by kination (Yin et al., 11 Mar 2026). The same study finds that these epochs are so short that the gravitational-wave signal from global strings is highly suppressed and far below current experimental sensitivity.

Current cosmological fits do not generically favor large kinetic-misalignment effects. In an ALP scenario with a constant-$\theta \equiv a/f_a$9 pre-oscillatory phase, a transition to a cosine potential, and decay into dark radiation, Bayesian analysis finds that ALP-generated dark radiation does not solve the $\ddot\theta+3H(T)\dot\theta+m^2(T)\sin\theta=0,$0 or $\ddot\theta+3H(T)\dot\theta+m^2(T)\sin\theta=0,$1 tensions, favors negative pre-oscillatory equation of state with

$\ddot\theta+3H(T)\dot\theta+m^2(T)\sin\theta=0,$2

and constrains

$\ddot\theta+3H(T)\dot\theta+m^2(T)\sin\theta=0,$3

(Pérez-Poyatos, 30 Apr 2026). This suggests that while KM is dynamically consistent, prominent early KM phases are not generically preferred by present cosmological data.

KMM can also be embedded in nonstandard expansion histories. In a PBH-dominated era, modified Hubble evolution and entropy injection change the trapping history of both vacuum and kinetic misalignment. A recent treatment including both semiclassical PBH evaporation and the memory-burden backreaction finds that the axion and PBH parameter space consistent with the observed relic abundance changes significantly, and that PBH-induced gravitational waves and hot-axion dark radiation supply complementary probes (Bandyopadhyay et al., 7 Jan 2025).

The broader phenomenological picture is therefore mixed. Some realizations populate regions that are underabundant in standard misalignment and place QCD axion or generic ALP dark matter in reach of IAXO, MADMAX, ALPS II, KLEVER, and future CMB experiments (Eröncel et al., 2024, Co et al., 2020, Lee et al., 2023). At the same time, UV completion is nontrivial because the origin of the kick, the damping of the radial mode, the preservation of the strong-CP solution, and the management of fragmentation all become part of the definition of the mechanism. This suggests that kinetic misalignment is best understood not as a single model but as a family of charge-driven axion production scenarios whose viability is set by how the initial angular momentum is generated, conserved, diluted, and eventually trapped.