On mixing enhancement by secondary baroclinic vorticity in a shock–bubble interaction