On the cosmology of scalar-tensor-vector gravity theory

We consider the cosmological consequences of a special scalar-tensor-vector theory of gravity, known as MOG (for MOdified Gravity), proposed to address the dark matter problem. This theory introduces two scalar fields $G(x)$ and $\mu(x)$, and one vector field $\phi_{\alpha}(x)$, in addition to the metric tensor. We set the corresponding self-interaction potentials to zero, as in the standard form of MOG. Then using the phase space analysis in the flat Friedmann-Robertson-Walker background, we show that the theory possesses a viable sequence of cosmological epochs with acceptable time dependency for the cosmic scale factor. We also investigate MOG's potential as a dark energy model and show that extra fields in MOG cannot provide a late time accelerated expansion. Furthermore, using a dynamical system approach to solve the non-linear field equations numerically, we calculate the angular size of the sound horizon, i.e. $\theta_{\text{s}}$, in MOG. We find that $8\times 10^{-3}\text{rad}<\theta_{\text{s}}<8.2\times 10^{-3} \text{rad}$ which is way outside the current observational bounds. Finally, we generalize MOG to a modified form called mMOG, and we find that mMOG passes the sound-horizon constraint. However, mMOG also cannot be considered as a dark energy model un;ess one adds a cosmological constant, and more importantly, the matter dominated era is still slightly different from the standard case.