SF6 contamination by air in the Pelletron

1
SF6 contamination by air in the Pelletron

• Recycler Departmental Meeting
• November 5, 2008
• A. Shemyakin

2
Results of the gas analyses performed in MI-31
assembly
July 17- repair of the storage tank leaks was
finished
Oct 8- repair of the Pelletron tank leaks was
finished
Jun 10, Jul 28, Aug 21 29- five bottles of SF6
were added

• Three consecutive tests of the gas in the
  Pelletron tank gave similar results SF6 is
  heavily contaminated by air
• The same company
• 2007 test was good

3
What might have contributed

• There were leaks in the Pelletron tank
• Between Jan and Oct 2008, the SF6 loss was 1.5
  2 psi/month (out of equivalent of 100 psia in
  both tanks)
• The leak added air in the time of gas transfers
  with a negative pressure in the Pelletron tank
• Decrease of the total SF6 amount forced to leave
  the very minimum relative pressure in the storage
  tank
• There was a large leak in the storage tank
• The most probable candidate because of scale of
  the leak
• The storage tank pressure gauge was found reading
  3 psig at atmospheric pressure instead of zero
• Temperature variations provided a good
  opportunity for sucking the air into the leaking
  storage tank
• Before transfers, the gauge was reading from 0
  (on 11-Apr-08) to 5 psig
• With a large amount of manually operated valves
  to set during transfers, its difficult to rule
  out completely a human error
• Until Jun-08, there were no readings storage tank
  P and T, so it is impossible to trace changes of
  PV/T during transfers

4
Effect of the Pelletron tank leak

• One of possibilities is an air contamination in
  the time when SF6 is pumped out of the Pelletron
  tank, and the tank is already at a negative (vs
  atmosphere) pressure
• Estimation from the following data
• min achievable pressure, 0.3 Torr in summer
  months
• Pipe between the gauge and the pump with 100 l/s
  conductivity in the molecular regime
• Pumping in this mode for 1 hour
• This effect adds only a small amount of air,
  equivalent of 0.5 psi per one transfer (out of
  85 psia)

5
Air intake through the storage tank

• Until 10-Jun-08, the storage tank pressure and
  temperature were not read by ACNET. Immediately
  after connection, a large leak was found in the
  pressure relief valve. The leak could have been
  there for a long time.
• It means a large loss of SF6 in the time of
  transfers and possibility to suck in air both in
  the pumping the gas into the Pelletron and in
  the time of temperature variations

Storage tank temperature and pressure after
connecting the gauges to ACNET. SF6 was added to
the storage tank two days before the connection.
Adding SF6 after repairing the storage tank leak.
The mass stayed constant after that.
6
Calculated mass during the transfers
Data (P, T) for transfers storage tank - from
technicians manual records in the log books,
Pelletron tank from data logger D44

• The table shows the sums of PV/T in both tanks
  before the transfer, with all gas in the storage
  tank, and after transferring gas back to the
  Pelletron
• Except the transfer on 11-Apr-08, numbers before
  the transfer and after the transfer back to the
  Pelletron are consistent within 1, and amount
  calculated for the gas in the storage tank is by
  7 lower (error in calibration of gauges).
• On 11-Apr-08, 6 was added in the time of
  pumping into the storage tank.
• Unclear how it might happen
• Storage tank pressure gauge read zero before the
  transfer (that corresponds to -3 psi)

7
Consequences

• Stability of HV and recirculation has not changed
• Most likely, due to the fact that we are running
  a nominally 6 MV machine at 4.3 MV
• Humidity is kept low by the alumina-filled
  dehumidifier
• Burst of the lost current are frequent only after
  tank openings
• Large absolute pressure of oxygen may result in a
  fire inside the tank
• On the other hand, amount of oil-like material is
  small, and recently there were no sparks
• Oxygen is said to deteriorate reliability
• Grease burning and contact oxidation
• Purification of so heavily contaminated SF6 is
  difficult
• According to companies Karl Williams have
  contacted, it can be only fully replaced, 100 k
  value
• A possibility to remove only oxygen are being
  looked at
• We may have to live with this gas and with a
  likely increased frequency of failures

8
Summary

• Tests showing a large air contamination in the
  system are most likely correct
• We are discussing sending a sample for testing in
  a different company
• Significant portion of air might have come
  through a large leak in the storage tank
• Only one instance, the transfer on April 11,
  2008, reliably shows a increase of PV/T (apart
  of adding SF6 from bottles)
• Still there is no complete explanation for almost
  40 of air in the gas
• Presently, there are no detectable leaks in the
  system
• There is no direct impact of the contamination on
  the cooler operation
• However, oxygen may deteriorate reliability of
  the Pelletron
• No cheap way to have the gas with no oxygen has
  been found yet