For a 5-15 watt transmitter, the IRF510 is a pretty good choice.
Its Drain-source breakdown voltage is 100V which gives a reasonable amount of headroom.
(on mod peaks the supply voltage will be 24V , assuming a 12V powersource.
The RF voltage on the FET will be something more than 2 times that, say peak voltage of 60-70 Volts.
The input capcitance of the IRF510 is quite low (Ciss ~= 150PF) which makes it
easy(er) to drive than larger mosfets with higher gate capacitance.
Why not try to get more out of the IRF510? The problem is Rds - the on Resistance,
which is 0.4 ohms for the IRF510. The output impedance seen by the FET is roughly V*V/2P
where V is DC supply voltage and P = power out.
If P = 10 and V=12 then the output impedance is about 12*12/2*10 = 7.2 Ohms
If P = 20 and V=12 then the output impedance is about 12*12/2*20 = 3.6 Ohms
As the value of Rds gets larger relative to the output impedance the efficciency drops
and you stat producing more heat and less RF.
At the small end of the scale, you could try MTP3055Es. They are tested at 7 Mhz,
but they are gate capacitance is about 3 times that of the IRF510, and thier breakdown
voltage is rated at 60 Watts. If you can drive them you may be able to get more out
than an IRF510, assuming you adjust the output network ...