Thermodynamic argument about SnO2 nanoribbon growth