Equilibrium heat relaxation experiments provide evidence that the ground state of the commensurate spin density wave (SDW) compound (TMTTF)2Br after the application of a sufficient magnetic field is different from the conventional ground state. The experiments are interpreted on the basis of the local model of strong pinning as the deconfinement of soliton-antisoliton pairs triggered by the Zeeman coupling to spin degrees of freedom, resulting in a magnetic field induced density wave glass for the spin carrying phase configuration.PACS numbers: 75.30.Fv,75.40.Cx,75.10.Nr Metastability results from energy minima in a system phase space, separated from each other by energy barriers. In some complex systems such as spin glasses, the number of metastable states and the scaling of the energy barriers with the system size are such that ergodicity is broken, meaning that time averaging is not equivalent to ensemble averaging because the system is "trapped" in a valley of the energy landscape. Metastability is also found in model systems such as molecular magnets or Josephson junctions, described by an energy potential V (ϕ) as a function of a single degree of freedom ϕ. We investigate below experimentally and theoretically the residual degrees of freedom of a spin density wave (SDW) at very low temperature in a magnetic field, interpreted as the properties of a classical potential V (ϕ(y i )) for the SDW phase ϕ(y i ) at the coordinate y i along the chain of a strong pinning impurity.Below the Peierls transition temperature, charge density waves (CDWs) and SDWs in quasi-one-dimensional (quasi-1D) compounds are characterized by a spatial modulation of the electronic density (or spin) along the chains. The phase profile is the result of a compromise between the elastic energy that penalizes large phase gradients, and the pinning energy that tends to fix the phase at the strong pinning centers. These two ingredients of the Fukuyama-Lee-Rice model [1] lead to metastability as the result of collective pinning. Collective pinning is however frozen below a glass transition of order ∼ 50 K, as shown by dielectric susceptibility experiments [2]. At very low temperature, the residual degrees of freedom in a zero magnetic field correspond to the local defects of the local model of strong pinning [3,4,5,6,7,8]. Larkin [4] and Ovchinnikov [5] have shown that a single strong pinning impurity leads to a bound state of an electronlike soliton and a hole-like antisoliton. This results in a potential V (ϕ(y i )) with multiple minima [6], leading to slow relaxation in agreement with the very low temperature heat relaxation experiments [5].In order to explore the role of a magnetic field, we choose the compound (TMTTF) 2 Br with a sufficiently narrow spectrum of relaxation times because of its commensurate (antiferromagnetic) ground state [9,10]. This allows a systematic study of the equilibrium energy relaxation over almost one decade in temperature and a direct comparison to the local model of strong pinning without introducing an...