Nitrogen states in Ga(As,P) and the Koster-Slater model

Recent photoluminescence data in Ga(As,P): N as a function of P fraction, $x$, indicates the existence of three states associated with isolated nitrogen impurities for $x\lesssim 0.47$. For $x\gtrsim 0.47$, however, only one state has been measured with photoluminescence. This state, ${\mathrm{N}}_X$, has been described by a Koster-Slater model in recent work and the numerical agreement between the predictions of these calculations and the data was thought to support the model. We show that the model is extremely sensitive to parameters used in calculations and that such numerical agreement does not justify the model. The only support for the Koster-Slater model is the observation of a single state. Recent electromodulation data indicate the existence of two states.