418  PREPARATION  OF  ALUMINIUM,  SODIUM,  ETC. 
While  the  operation  is  in  progress,  the  tube  C  is  kept  at  a 
temperature  of  400°  to  572°  F.,  and  the  cylinder  D  is  heated  so 
as  to  be  barely  red  at  the  under  surface ;  the  reaction  between 
the  chloride  of  aluminium  and  sodium  is  attended  with  so  great 
an  evolution  of  heat  that  it  is  sometimes  necessary  to  remove  the 
fire  entirely. 
Yery  soon  after  the  reaction  has  commenced,  the  chloride  of 
sodium  combines  with  the  excess  of  chloride  of  aluminium,  form- 
ing a  double  salt,  which  is  sufficiently  volatile  to  be  carried  to 
the  adjoining  tray,  where  the  aluminium  is  eliminated  by  means 
of  the  sodium.  The  reaction  does  not  commence  in  one  tray, 
until  it  has  ceased  in  that  which  precedes  it,  and  it  is  at  an  end 
when  on  removing  the  cover  (W)  of  the  cylinder,  the  last  tray 
is  found  to  contain,  in  the  place  of  the  sodium,  a  black  mass, 
covered  with  a  colorless  liquid  which  is  the  double  chloride  of 
aluminium  and  sodium.  The  trays  are  then  withdrawn,  and  re- 
placed by  a  fresh  charge  of  sodium. 
When  the  contents  of  the  trays  are  cold  they  are  transferred 
to  crucibles  of  cast  iron  or  clay,  and  heated  until  completely 
melted,  and  the  double  chloride  begins  to  volatilize.  Most  fre- 
quently the  reaction  between  chloride  of  aluminium  and  the  sodium 
is  not  completed  in  the  cylinder,  the  sodium  being  protected  by 
a  crust  of  chloride  of  sodium  ;  but  a  double  chloride  of  aluminium 
and  sodium  at  the  top  of  the  trays  is  always  sufficient  to  ensure 
the  perfect  absorption  of  the  sodium  in  the  crucible,  and  the 
aluminium  remains  finally  in  contact  with  a  large  excess  of  chlo- 
ride of  aluminium,  which  is  indispensable  for  the  success  of  the 
operation. 
When  the  crucibles  are  cold,  the  layer  of  chloride  of  sodium  is 
is  removed  from  the  top,  and  from  the  lower  part  the  globules  of 
metal  are  separated  by  washing  with  water;  but,  unfortunately, 
the  chloride  of  aluminium  is  dissolved,  and  exercises  a  very  de- 
structive action  upon  the  metal,  so  that  the  globules  obtained  are 
not  larger  than  a  pin's  head.  These  are  collected,  dried,  put 
into  a  crucible,  melted,  and  run  into  an  ingot. 
It  is  necessary  to  take  especial  care  in  removing  from  the 
sodium  every  particle  of  carbon  which  is  sometimes  mixed  with 
it  when  badly  prepared  or  insufficiently  purified,  for  otherwise  con- 
siderable quantities  of  metallic  cyanides  or  cyanates  are  formed, 
