The possibility of using a dynamic environment to achieve and optimize phase synchronization in a network of self-excited cells with free-end boundary conditions is addressed in this paper. The dynamic environment is an oscillatory bath coupled linearly to a network of four cells. The boundaries of the stable solutions of the dynamical states as well as the ranges of coupling parameters leading to stability and instability of synchronization are determined. Numerical simulations are used to check the accuracy and to complement the result obtained from analytical treatment. The robustness of synchronization strategy is tested using a local and global injection of Gaussian white noise in the network. The control gain parameter of the bath coupling can modulate the occurrence of synchronization in the network without prior requirement of direct coupling among all the cells. The process of synchronization obtained through local injection is independent of the node at which noise is injected into the system. As compared to local injection, the global injection scheme increases the range of noise amplitude for which synchronization occurs in the network.