L1 is ok and is used for matching purposes. You have to add another inductor from the anode of the diode to ground so that the DC current can flows through this new inductor, the diode and the through the load.
The value of this added inductor have to be high enough not to cause significant RF losses. I think a value of about 270 nH. In the simulation you can use a very large value (1 mH) to see if all is ok, but in real world you cannot, due to the self resonace frequency of the inductor.
Two other things: 50 ohms is a quite heavy load. Are you sure about it ? The other one is: I suggest you not to connect directly the cathode of the diode to the load+output cap, but putting in series an inductor to filter the residual otuput RF. In this case you have to account for its slew rate, that means it needs a while to reach the DC final value