Aluminium and Silver metals are characterised by a relatively short supercool (supercool is
the characteristic of a freezing pure metal to remain liquid at a temperature below that at
which the solid melts). The supercool of these metals can be expected to be less than
0.5°C. The typical supercool for copper is 1.5°C t o 2.5°C.
The cell is placed in the furnace, suitable insulation and cover added and a monitoring
thermometer inserted, (for Al + Ag a HT SPRT (High Temperature Standard Platinum
Resistance Thermometer) and for Cu a standard thermocouple of suitable type and
construction). Initially the furnace controller should be set 5-10°C below the nominal fixed
point temperature and left to stabilise. At this point the monitor thermometer is used to
calibrate the furnace controller which is then set to be +1 to +2°C above the fixed point
temperature allowing for any controller offset. The temperature rise is monitored with a
bridge and/or recorder connected to the thermometer.
Following the melt arrest, the temperature of the cell will rise to the controlled temperature.
The metal in the cell is now entirely in the liquid phase and may be maintained in this
condition for any desired period of time, for example, to accommodate to a calibration
schedule. For optimal realisation of the freeze plateau the molten cell should be left at +5°C
overnight prior to freezing the cell the next day.
To freeze, the furnace controller is set below the actual freeze temperature (for pure metals,
melt and freeze temperatures are theoretically identical). The suggested setting is 1°C
below the freeze temperature; this is, assuredly, below the bottom of the supercool. The
furnace is allowed to cool to this new setpoint temperature, taking typically 30 to 45 minutes
to do so.
When the monitor indicates that the cell is at, or below, its freeze temperature, the monitor is
removed to a rack or the pre-warming tube of the furnace and replaced by a cold rod of
quartz. This initiates nucleation. After 2 minutes the rod can be removed and replaced by
the monitor again.
This procedure creates a radial freeze from the inside and outside walls of the cell towards
the centre.
A typical melt/freeze sequence is shown in Figure 4.
If the cell is left too long in the furnace without initiating the freeze as described above,
nucleation will occur and the cell will begin to freeze from the bottom of the cell upwards.
This will result in a short, imperfect, plateau and, moreover, give an incorrect value of freeze
point (typically 10mK below that expected).