Identification of the optimum time-delay for chaos synchronization regimes of semiconductor lasers

It is shown that the optimum correlation for chaos synchronization of master-slave semiconductor lasers occurs at a delay time that is sensitive to both the injection strength and the frequency detuning of the driving field. This effect may be overlooked if the correlation function is not evaluated globally. The correlation function has been used as the de facto method for quantitatively determining the degree of synchronization achieved between unidirectional coupled chaotic semiconductor lasers, see for example . The correlation function is a continuous function of the delay time but is normally evaluated at only two delay times. It is shown that this approach can cause a misidentification of the dominant synchronization process and can also mask important temporal fluctuations in the nature and quality of the chaos synchronization. In essence, two factors contribute to the potential for misidentification, the first is the inherent small time shift required to obtain the optimum correlation and the second is the quasi-periodicity that is present in certain chaotic regimes. This paper shows that a reevaluation of some of the published numerical studies of chaos synchronization is necessary.