Comment by jarvist

The speed of sound does not depend on pressure, only temperature.

I suspect this is occurring because the (sonic) flow through the nozzle cools as it expands, therefore the speed of sound drops, making the same flow now supersonic in the cooled gas.