You can make your own estimate; assume a certain fluence, power per square meter, f, required to activate the detector; start with power P, an initial beam diameter, D, and a beam divergence $\theta $. As the beam travels a given distance z, you can now calculate the divergence, and the increase in the size of the wave front, and the fluence at that distance.
You will now be able to determine the available fluence at every distance, given the assumption that there are no losses, or you can include a loss parameter, such as a certain proportion per meter due to scattering.
The most powerful lasers today can produce brief pulses, under 100 femtoseconds in duration, with perhaps 10 joules per pulse, and an initial beam diameter of one millimeter. Of course you can invent more powerful lasers for your project, but this gives you a start. This system will also require some fantastic degree of pointing stability, for not only will you have to point at a distant star, but you will also have to point at the actual detector in orbit about that star, and actually hit it.
We can actually detect single photons, but for communications you will need to provide a communications protocol which will permit the detector to know that it has received a signal, as opposed to a random photon.