Version françaiseFR

Magnetic field in astrophysical and geophysical structures 

Magnetic field in astrophysical and geophysical structures

There is little doubt that the magnetic field plays a major role in the formation, evolution and structure of many astrophysical and geophysical objects like galaxies, molecular clouds, accretion disks, stars and planets. As such, magnetic field has received considerable attention during the last decades. However in spite of these efforts, many questions remain largely open from astrophysical and geophysical points of view, on theoretical and numerical grounds. This is largely due to the inherent complexity of the non-linear magnetic processes as magnetohydrodynamic (MHD) turbulence and dynamo, to the uncertainties regarding the validity of the ideal MHD approximation and to the difficulty in computing MHD flows. The present project has two main goals. First we want to investigate the role of the magnetic field in various astrophysical and geophysical applications. In particular we want to study its origin, its effects and its structure. Second, we want to develop numerical tools suitable for such investigations, and make these codes freely available to the research community. This MHD code development activity and the service provided to the community is, to our knowledge, not covered by any existing project in France. An important aspect of our project is to gather researchers with different backgrounds in MHD with a common objective: the development and validation of French open source codes with a variety of applications. This project gathers experts in galaxy and star formation, dynamics of the interstellar medium, accretion disks, dynamo theory, magnetosphere and plasma physics. This complementary knowledge is, in our opinion, a great strength of the project considering the complexity and the generic nature of many MHD effects. Such interdisciplinary collaborations are already active within our group.

In this four year project, we want to focus on a subset of astrophysical and geophysical problems in which MHD plays a crucial role, namely formation and evolution of stars and galaxies, with special emphasis on the generation of the magnetic field in these objects, interstellar turbulence, planetary magnetopause and dynamo action. There is a strong synergy with S. Balbus's Chaire d'Excellence CEMAG project at ENS, and we plan to carry on extensive simulations on the CEMAG cluster during this new project. But the MAGNET project, which overlaps only slightly the Chaire d'Excellence time range, covers a much broader set of astrophysical and geophysical topics. Moreover it involves a much larger team, and aims at the delivery of advanced MHD codes, which is outside of the scope of the Chaire d'Excellence. MAGNET will also complement the HORIZON project, by addressing the possibly crucial role of magnetic fields in the formation of galaxies, an objective which is clearly beyond the work plan of HORIZON.


Galaxies are believed to be formed by continuous accretion of dilute gas and small mass progenitors. Differences in the mass assembly history result in galaxies with different shapes and colours, from large spirals to red ellipticals. This promising picture is however complicated by the small scale physics of the interstellar medium, which controls the star formation history in galaxies (Rasera & Teyssier 2006, A&A 445, 1), as well as feedback processes driven by supernovae explosions or AGN-driven outflows. These small scales processes (e.g. de Avillez & Breichtwerdt 2005, A&A, 436, 585) are known to produce dramatic effects on larger scales and to influence the galaxy formation history itself. This interaction between cosmological scales and interstellar scales is the current frontier in studies of galaxy formation and cosmology (Tacker & Bryan 2006, ApJ in press, astro-ph/0512027). The role and the generation of the magnetic field in galaxies will be investigated by performing both large scale numerical simulations aiming to describe the galactic scales as well as simulations of intermediate scales describing the dynamics of the turbulent and magnetised interstellar medium and their influence on the galaxy evolution.


Magnetic field has a deep imprint on the star formation process (Shu et al. 1987, ARA&A 25, 23). First magnetic field provides a significant support to the gas against gravitational collapse, second it allows the transport of angular momentum in the prestellar phase through magnetic braking, through the launching of MHD outflows and jets and through MHD turbulence triggered by the magneto-rotational instability. Different groups have performed detailed numerical studies at various scales. On the molecular clouds scale, 3D numerical simulations were performed using ideal MHD on a fix grid (Gammie et al. 2003, ApJ, 589, 728). At the scale of a collapsing dense core, Machida et al. (2005, ApJ, 362, 369) and Banarjee & Pudritz (2006, ApJ in press, astro-ph) have done 3D MHD calculations using a nested grid. In their simulations, a disk forms and outflows and jets are launched. Our objective will be to pursue along the line of these works using the versatility of our 3D RAMSES-MHD adaptive mesh refinement (AMR) code, including various non ideal MHD effects like ambipolar diffusion, which plays a crucial role in molecular clouds. Special attention will be given to the physical MHD processes taking place in the disk and in the jets. MHD effects remain most important on stellar dynamics once the star is formed. Most stars, and certainly the Sun, maintain their own magnetic field through self excited dynamo action. This mechanism is intimately related to the cyclic behaviour of stellar activity. Self-excited dynamo theory still resists analytical as well as numerical works in a parameter regime relevant to natural objects (galaxies, stars and planets). Cyclic behaviours for solar models are only captured so far through simplified 2D models (known as "alpha-omega"). The understanding and description of the cyclic magnetic behaviour of the Sun in 3D dynamo models for stars remains a challenging issue that would require a proper modelling of the tachocline. Besides, magnetic boundary conditions, and the matching with the force-free solar corona, are usually circumvented through the assumption of a potential outside field. The coupling between coronal currents and dynamo action is still to be understood.

Planets, Magnetosphere.

A striking property of dynamo action seems to be its ability to work on a large variety of natural bodies. Most planets in the solar system (Venus and Mars excepted) exhibit a magnetic field. About 1/100th of the size of the Sun, motions in the Earth’s liquid core are still vigorous enough to regenerate the magnetic field against diffusion. Although these are small objects compared to the Galaxy, the parameter regime relevant to the Earth is still out of reach of present numerical simulations. Yet some models are able to reproduce most of the observed features of the geomagnetic field. A detailed comparison of how these models evolve as their parameters and geometries are varied so as to model the other planets in our system is still missing and would provide important insight on the validity of these models and on planetary magnetic field generation. The Earth magnetosphere is doubtlessly the most explored astrophysical plasma. Thanks to different space missions, it is indeed possible to get in-situ and non disruptive measurements, allowing to make real progress in understanding challenging phenomena such as turbulence and magnetic reconnection. This latter is believed to be the mechanism leading to the penetration of the solar wind particles into the magnetospheres of the planets. However all the existing theoretical and numerical models of reconnection do not take into account the presence of large scale turbulence and its possible influence on the small scales where reconnection occurs. We aim here to improve these models by running numerical simulations in which large scale turbulence will be included as a boundary condition to the reconnection problem. The last experimental results on turbulence obtained at CETP (Sahraoui et al., PRL, 2006) will be considered as inputs into these new models.

State of the art and developments in numerical modelling.

The continuously growing power of massively parallel supercomputers opened new and exciting perspectives to numerical MHD. Nevertheless, in spite of numerous international efforts and recent progress, the computation of astrophysical and geophysical MHD flows remains a challenging issue. This is largely due to the complexity of the MHD equations, which entails seven waves instead of three in the hydrodynamics case and to the divergence free constraint on the magnetic field, which imposes the development of special numerical techniques. Although a few public tools for MHD are available in the rest of the world (FLASH, NIRVANA, ATHENA, ZEUS), none of them has been developed in France. The use of such tools requires a high level of expertise making the proximity of local experts a decisive advantage for users. The use of existing tools as a “black box” often induces the risk of interpreting numerical artefacts as physical effects. Besides, MHD codes being presently less versatile than hydro codes, they must often be significantly modified before they can be applied to a different problem than the one for which they were originally developed.

We propose to pursue the development and delivery of three different types of MHD codes which can be used in various contexts, namely a grid based MHD code based on conservative Godunov scheme, a spectral code suitable for the study of incompressible (or weakly compressible) MHD in spherical objects like planets and stars, a particle-in-cell code (PIC) which allows the investigation of non ideal MHD effects.