Magnetically Arrested Disk (MAD) State

• Magnetically Arrested Disk (MAD) state is an accretion configuration where saturated, large-scale poloidal magnetic flux near a black hole balances ram pressure to regulate flow.
• Efficient jet launching via the Blandford–Znajek process is achieved as high magnetic pressure and turbulent reconnection mediate angular momentum transport.
• GRMHD simulations and multiwavelength observations—from X-ray binaries to AGN—validate MAD behavior, including cyclic flux eruptions and sub-Keplerian disk rotation.

A magnetically arrested disk (MAD) is an accretion flow configuration in which the accumulation of large-scale poloidal magnetic flux near a black hole saturates to a dynamically dominant level, such that magnetic pressure balances or exceeds the ram pressure of the inflowing gas. This saturation suppresses further advection of net flux, fundamentally regulates angular momentum transport, and acts as the key mechanism by which powerful, relativistic jets are launched via the Blandford–Znajek process. The MAD paradigm is supported by extensive general relativistic magnetohydrodynamic (GRMHD) simulations and multi-wavelength observations spanning black hole X-ray binaries to active galactic nuclei. MADs exhibit cyclic behavior, efficient jet launching, flux eruptions, and a global disk structure controlled by both poloidal flux advection and toroidally-amplified magnetic pressure.

1. Physical Criteria and Dynamical Definition

A MAD is characterized by saturation of net poloidal magnetic flux at the horizon, quantified by the dimensionless flux parameter: $$\phi_{\rm BH} = \frac{\Phi_{\rm BH}}{\sqrt{\dot{M}\,r_g^2\,c}}$$ where $\Phi_{\rm BH}$ is the absolute hemispheric poloidal flux threading the black hole, $\dot{M}$ is the mass accretion rate, and $r_g = GM/c^2$ the gravitational radius. Empirical and theoretical analysis finds that the onset of the MAD regime corresponds to $\phi_{\rm BH} \gtrsim 50$, leading to magnetic pressure $p_{\rm mag} \sim B^2/8\pi$ balancing the ram pressure $\rho v_r^2$ near the horizon (Mocz et al., 2014, Zhang et al., 2023, Lalakos et al., 29 May 2025, Xie et al., 2019, Salas et al., 2024, Suková et al., 2023).

The MAD constraint is achieved via inward advection of large-scale poloidal field lines from the environment (e.g., ISM or companion envelope) by flux-freezing in a high magnetic Reynolds number flow. When the field becomes dynamically important, further flux accumulation is arrested, and accretion is intermittently mediated by non-axisymmetric interchange instabilities or reconnection-driven eruptions (Mondal et al., 2018, Begelman et al., 2021, Li et al., 2024).

The spatially averaged plasma beta parameter, $$\beta_{\rm ave} = \langle p_{\rm gas} \rangle / \langle p_{\rm mag} \rangle$$, also serves as a global diagnostic. Flows with $\beta_{\rm ave} \lesssim 1$ robustly indicate entry into the MAD state across both ideal and resistive regimes (Aktar et al., 29 May 2025, Suková et al., 2023).

2. Magnetic Flux Saturation, Transport, and Eruptions

The saturated MAD state sustains a quasi-steady balance between advective and diffusive transport of magnetic flux. In GRMHD, the net flux transport velocity $v_\Phi = v_{\rm adv} + v_{\rm diff}$, where $v_{\rm adv}$ arises from large-scale inflow and $v_{\rm diff}$ from turbulent (Reynolds-averaged) EMFs. In simulations, $|v_\Phi| \ll |v_{\rm adv}|, |v_{\rm diff}|$, so net flux is nearly stationary over orbit-averaged timescales, with $\phi_{\rm BH}$ remaining close to its threshold value (Jacquemin-Ide et al., 29 Oct 2025).

Flux eruptions represent stochastic deviations from equilibrium, as local interchange and reconnection events transiently expel flux outward. The recurrence timescale for flux eruptions scales as $t_{\rm rec} \sim 1500\, r_g / c$, orders of magnitude longer than the orbital time, consistent with Sgr A* and M87-like AGN variability (Jacquemin-Ide et al., 29 Oct 2025, Salas et al., 2024). These eruptions drive large-scale outflow variability, modulate horizon-scale accretion rates and underpin the observed stochastic jet duty cycles.

3. Disk Structure, Angular Momentum Transport, and Instabilities

MAD disks are generically geometrically thick ($H/R \gtrsim 0.1$), highly magnetized ($\beta \ll 1$ in the inner regions), and display strongly sub-Keplerian rotation ($\Omega / \Omega_K \sim 0.4-0.6$). The condition for sub-Keplerian rotation is robust, with analytic models and simulations showing $\ell_0 \simeq 0.5-0.6$ for the midplane specific angular momentum (Begelman et al., 2021, Li et al., 2024). Saturated poloidal flux inflates a vertical "magnetosphere," while differential rotation and MRI-generated dynamo action amplify the toroidal field $B_\phi$, which dominates the pressure and angular momentum transport at all radii beyond the ISCO (Begelman et al., 2021, Zhang et al., 2023).

While axisymmetric MRI is often suppressed due to high vertical field strength, non-axisymmetric MRI modes, magnetic Rayleigh–Taylor (RT) interchange instabilities, and convective instability structures, as formalized by generalized Høiland criteria, remain active and mediate turbulent transport. In thin MADs ($H/R \lesssim 0.1$), RT instability and interchange-driven bubble formation regulate the episodic release and redistribution of magnetic flux, maintaining efficient angular momentum removal and controlling $\alpha$ viscosity (Marshall et al., 2017, Salas et al., 2024).

4. Jet Launching, Power Scaling, and Spin Evolution

MAD disks enable highly efficient jet production via the Blandford–Znajek (BZ) process, which extracts rotational energy from the spinning black hole. Jet power scales quadratically with both net poloidal flux and BH spin: $$P_{\rm jet} \sim \kappa \Phi_{\rm BH}^2 \Omega_H^2 / c$$ with $\kappa$ a geometric factor and $\Omega_H$ the horizon angular frequency (Mondal et al., 2018, Narayan et al., 2021, Zhang et al., 2023, Aktar et al., 29 May 2025).

Jet and outflow efficiencies in the MAD state routinely reach $\eta = L_{\rm jet}/(\dot{M} c^2) \gtrsim 100\%$ for high spin ($a \gtrsim 0.5$), demonstrating that jets can extract more energy than is supplied by rest-mass accretion alone (Narayan et al., 2021, Lalakos et al., 29 May 2025). MAD jets are typically parabolic or monopolar in structure, and their geometry, width, and collimation depend on both spin and saturated flux, with prograde spins producing thinner disks and wider jets compared to retrograde cases (Narayan et al., 2021, Zhang et al., 2023).

Persistent MAD jets also drive black hole spindown. GRMHD and analytic models confirm that luminous MADs converge to a universal low equilibrium spin $a_{\rm eq} \approx 0.3$ for $0.03 \leq h/r \leq 0.1$, independent of further thinning (Lowell et al., 24 Feb 2025). The efficiency of angular momentum extraction is set by the saturated flux and toroidal field winding, with prograde disks losing spin more rapidly than retrograde or non-rotating ones (Zhang et al., 2023, Narayan et al., 2021, Lowell et al., 24 Feb 2025).

5. Observational Signatures and Multi-scale Applications

MAD formation is supported by multi-wavelength observations of delayed radio and optical flares following X-ray peaks, characteristic of jet activation after magnetic flux build-up. In MAXI J1820+070, the measured time lags ($\tau_{\rm RX} \sim 8$ days, $\tau_{\rm OX} \sim 17$ days) directly map the transition in corona morphology and disk thermal instability corresponding to MAD onset (You et al., 2023).

MAD states naturally explain episodic jet outbursts, ultra-fast outflows (UFOs), and variable broadband emission from AGN and X-ray binaries. In particular, the persistent, stochastic broad-velocity outflows ($v > 0.5c$) and duty cycle timescales ($t_{\rm MAD} \sim$ days–Myears) match observed features in M87*, Sgr A*, and low-luminosity AGN (Suková et al., 2023, Salas et al., 2024, Lalakos et al., 29 May 2025).

In quiescent BH binaries, the MAD paradigm accounts for hard X-ray spectral tails, optical–X-ray broadband correlations, and PeV cosmic-ray production by enabling efficient nonthermal particle acceleration in magnetically dominated reconnection zones (Kimura et al., 2021).

Resolution studies show that turbulent convection, RT mixing at flux-tube boundaries, and plasmoid-mediated reconnection require high-resolution grids to capture accurately. These phenomena affect jet–disk mixing, sheath thickness and temperature, yield enhanced variability, and may be essential for predicting polarimetric and spectral observables targeted by VLBI and X-ray telescopes (Salas et al., 2024).

6. Theoretical Variants, Accretion Regimes, and Distinguishing Criteria

MADs are observed and modeled within a broad range of accretion regimes. The canonical flux scaling $\Phi_{\rm BH} \propto (\dot{M} r_g^2 c)^{1/2}$ is found to hold for both radiatively inefficient (ADAF/RIAF) and radiatively efficient (thin/super-Eddington) disks, with appropriate corrections for radiative cooling and disk aspect ratio (Mocz et al., 2014, Avara et al., 2015). MAD onset depends on the supply of external flux; if the outer disk field strength is too weak ($\beta_{\rm out} \gtrsim 100$), MAD formation via flux advection is suppressed (Li et al., 2024).

Distinguishing MAD from SANE (standard and normal evolution) disks relies on a combination of criteria: saturated dimensionless flux ($\phi_{\rm BH} \gtrsim 30-50$ for MAD, $\lesssim 10$ for SANE), sub-Keplerian rotation ($\Omega/\Omega_K \sim 0.5-0.6$), constancy of $\beta$ with vertical height, dominance of $B_\phi$ over poloidal components, and positive modified Høiland criteria near the disk midplane (Begelman et al., 2021).

MADs also impose global constraints on timing features: for example, the strong magnetic "welding" of the inner disk to the hole suppresses Lense–Thirring precession and the associated type-C QPOs in BHXRBs. Observation of strong QPOs and jet precession precludes a MAD geometry in luminous hard states (Fragile et al., 2023).

7. Astrophysical Implications

The universality and self-regulation of the MAD state underlie several key phenomena:

• Jet production and feedback in AGN is naturally limited by MAD spindown; luminous MADs drive BH spins toward a universal equilibrium $a_{\rm eq} \sim 0.3$, with significant consequences for feedback energy budgets over cosmic time and consistency with LIGO/Virgo/KAGRA spin measurements (Lowell et al., 24 Feb 2025).
• AGN jet duty cycles are set by cyclic MAD–RAD transitions, with characteristic timescales of tens of Bondi accretion periods (Lalakos et al., 29 May 2025).
• MAD-induced variability and NFT flares in Sgr A* can be directly modeled as flux eruptions (Jacquemin-Ide et al., 29 Oct 2025, Salas et al., 2024).
• In gamma-ray bursts, MAD variability explains the characteristic second-scale pulse structure and correlations between quiescent intervals and subsequent active phases (Lloyd-Ronning et al., 2016).
• Radiative properties of MADs are systematically brighter than SANEs at fixed accretion rate, with the SED shaped by turbulent magnetic heating, modified cooling rates, and reconnection-driven acceleration (Xie et al., 2019).

MAD physics is robust across scales (from stellar-mass XRBs to $10^9 M_\odot$ AGN) and accretion states, enabling unified modeling of horizon-scale processes, jet feedback, and multi-epoch variability across astrophysical populations.

---

MADs represent the self-sustaining, magnetically-dominated endpoint of accretion flow evolution when sufficient net poloidal flux is supplied, with consequences extending from horizon-scale disk physics to galaxy-scale feedback and cosmic ray populations. The synergy of analytic modeling, large-scale GRMHD simulation, and direct observational diagnostics continues to refine the theoretical landscape and its astrophysical applications.