NeX:
i had the same idea about a DC side MOSFET, i had already bought the components to replace the AC side like the original design but a quick experiment on the DC side left me with a blown transistor, i think maybe the capacitor acts like an infinite load? and it sucked too many amps through the transistor. but i am not sure, i didn't pursue it further but maybe i should. i will post a schematic up in a sec
My best guess is the voltage that's responsible for this. The transistor must be able to handle PEAK voltage (that 230V is RMS, for the motor that will be fine) - so over 320V peaks. You'll need one that can handle 350V at least, preferably 400V or more. Current won't be that bad - 4A rated, so stall current may be 10-15A or so, that'd be peak. 4A is not much for a power MOSFET.
Then I don't see a flyback diode in your diagram; around every zero crossing of the AC supply the current in the motor will drop drastically, and that may cause reverse voltage spikes which may kill the transistor. Makes me also wonder what type/capacitance that capacitor is.