In this paper Caldeira gives a model for the Stern Gerlach device in the vacuum. the incoming particle is described by the tensor product of a space term and a spin term a |u> + b |d> the SG is in the vacuum (no air around). Under the effect of the spatial variation of the magnetic field, there is entanglement of these two degrees of freedom. the spatial part splits into two Gaussians that deviate and then hardly overlap. By a partial trace on the external degree of freedom (the spatial position of the Gaussian), one gets a decohered reduced density matrix. Calfeira writes at the end that if we recombine the paths of these two Gaussians we retrieve the starting state and thus the lost coherences. Suppose that we have three regions: In the first Gaussian separate. in the second a device blocks their remoteness and the paths are parallel. In the third region a device recombines yhe two possible paths. If in region 2 we have the vacuum we get what Caldeira described with a final recoherence . Now suppose now that in the second region there are several air molecules (say 1 on the way up and 1 on the way down),I guess that in the Hamiltonian an interaction term must be added The device in region 3 remaining unchanged, how does QM predict different behaviors for the particle out of region 3 (with or without gas in region 2)? I recall that in both cases the output particle of region 1 has null non-diagonal terms in the density matrix.