Gas Target

Revision history

Version	Author	Description	Date
1.0	D. Hutcheon	Initial Document	27 September 2002
1.1	D. Hutcheon	Hidden valves,shack	4 April 2007
1.2	D. Hutcheon	Leak checking	27 April 2007
1.3	D. Hutcheon	Decadal update	22 January 2019
1.4	A. Lennarz	Transferred to wiki	June 25th, 2025

Safety with hydrogen as the target gas

The following is a brief outline of safety considerations. For more details, consult the DRAGON Safety Report.

The strategy for safe operation with hydrogen gas is:

• avoid formation of an explosive mixture of air and hydrogen
• avoid ignition sources at pressures where combustion may take place

Normal operation

When the zeolite cleaning trap is in operation, there is a large inventory of "invisible" hydrogen adsorbed on the zeolite molecular sieve. For a trap inlet pressure (CMTRIN) of 45 Torr the adsorbed hydrogen amounts to 5 litres at STP, while the "free" recirculating hydrogen may be only 15% of this amount. The heat of combustion of this total amount of hydrogen is approximately 50 kJ - enough to raise 5 tonnes a height of 1 m.

IT IS ESSENTIAL THAT THE TARGET OPERATOR KNOWS WHETHER THE TRAP IS COLD AND WHETHER IT IS "LOADED" WITH HYDROGEN.

The correct response to the sonalert alarm from the LN2 automatic filler or from the EPICS system depends upon this knowledge. Therefore:

ANY SIGNIFICANT CHANGE IN THE INVENTORY OF HYDROGEN IN THE CLEANING TRAP MUST BE ACCOMPANIED BY AN ENTRY IN THE DRAGON “Equipment Status” elog (https://elog.triumf.ca/Dragon/}.

The entry should indicate trap status (warm/cold), estimated hydrogen inventory.. The 6-litre Buffer Tank should be used to control the amount of gas loaded into the trap. The number of 50-Torr "loads" from the Inlet Buffer Tank should be tallied during an initial fill of the trap. Any large deviation from the expected 15-20 "loads" to reach approx. 45 Torr pressure must be investigated and the problem fixed.

(During tuning for a new beam energy it is common for ISAC Operations to request that gas be removed from the DRAGON target. Each time that this is done, some of the "free" circulating hydrogen is lost. It is convenient to replenish the supply by adding 1 or 2 "loads" of gas through the Inlet Buffer Tank when the target cell is again filled with recirculating gas. This is considered a routine operation which need not be reported in the Equipment Status elog)

The purpose of the cleaning trap is to remove air and other impurities from the recirculating hydrogen gas. Thus, when the trap does its job it is almost impossible to detect any small leak of air into the target system. The implications for target operation with a trap are:

• A cold trap which has been loaded with hydrogen should be treated as though it contains air also, in a potentially explosive mixture. An exhaust fan must be running whenever a cold trap contains hydrogen, to allow safe dilution of hydrogen in the event of an unanticipated release from the trap through the pressure relief valve. Normal procedure when the trap is being warmed up is to pump on it with the roughing pump, which vents to the high-air-flow exhaust line.
• Before cooling the trap, it must be checked that any air leaks are less than could amount to 1.5 atm-litre over a 2-week period (assuming a trap will be kept cold no longer than 2 weeks at a time). Possible sources include leaks in the differential pumping stages, in the trap itself or in the buffer tank and gas supply lines. The leak rates should be measured by pumping down the system, then isolating it with the Roots blowers continuing to run; the gauge CMTRIN connects to an effective volume of 12 litres, from which the quantity of gas (leaking or outgassing) may be calculated using an observed rate of increase in CMTRIN. Other sources are HEBT or the DRAGON separator when isolation valves HEBT2:IV8 or DRA:IV11 are open. At an estimated 1000 litre/sec conductance in the beam pipes, the sum of partial pressures of air upstream and downstream of the target should be less than 1E-6 Torr when there is no gas added to the trap.
• Failure of the automatic LN2 filling system or failure of the primary exhaust fan must trigger immediate action, either to fix the fault or to begin controlled warmup of the trap.

Unattended gas target, trap containing absorbed hydrogen

At times it may be desirable for the DRAGON gas target operator to leave the target unattended for a period of many hours. This may be done, provided it is arranged that someone else (e.g. an ISAC Operator) checks the Trap status via Epics at least once every two hours. Of greatest concern is failure to keep the trap cold, for example if an LN2 dewar goes empty or the automatic filling system fails.

Protection of equipment

Equipment upstream of the gas target (HEBT) or downstream (Separator) is protected against excessive gas loads or shock waves from the DRAGON target via the interlock conditions on isolation valves HEBT2:IV8 and DRA:IV11. These require satisfactory readings (low pressure and no error condition) in ion gauges on either side of the valve. Equally, the interlocks protect the gas target against HEBT and the Separator. The interlocks on IV8 and IV11 should never be bypassed, nor should ion gauges IGU3 or IGD4 be operated in a mode where they give spurious indication of high vacuum, due to malfunction.

The Roots blower which operates at the highest pressure does not have any internal bypass. It should not be operated with the pressure at CMTRIN greater than 60 Torr.

Turbo-molecular pumps may be damaged if subjected to a burst of air (or other gas) when they are rotating at high speed. When venting the system after turbos have been on, follow the 'spin-down' operations given in the sections on "venting" below.

Gases of high molecular weight are less efficient than hydrogen or helium at cooling turbo pump rotors. Circulation of such gases at multi-Torr pressures has led to overheating and destruction of turbo pumps.

General description of the gas target system

Gas supply shack

Gas for the DRAGON windowless target, as well as the isobutane for its ion chamber detector, is supplied from a small building just outside the ISAC-1 hall to the east side ("the shack"). The general TRIUMF access key 2T85 will open the locked door. Gas is supplied from cylinders, passes through a flow gauge and then is conducted by stainless steel tubing along the inside north wall of ISAC-1 hall and finally south to DRAGON. Return lines provide for venting of gas at the shack: this happens for the isobutane of the ion chamber, but the target has its own separate venting line to roof fans.

A member of the Detector Facility Group is responsible for supplying gas cylinders and for general operations within the gas shack. One of the group should be informed when it is planned to start gas flow for an experiment or to shut off the gas at the end of an experiment.

Permanent metal lines connect hydrogen, helium and isobutane cylinders plus backup cylinders to a manifold located near the centre of the shack.

The manifold has clearly labelled flow meters and lines for the DRAGON target and isobutane detector. The experimenter should not disconnect or connect lines. The choice of hydrogen or helium gas is made by a valve below the target flow meter, together with opening the appropriate supply valve (between the manifold structure and the overhead cable tray) and closing the valve of the unwanted gas. Do not turn off gas at the cylinder.

Note that isobutane is stored in the cylinders as a liquid, not a high-pressure gas, so that the fullness of a cylinder is monitored by its weight. It is arranged that the backup isobutane supply will automatically come into service when the primary cylinder becomes empty.

Windowless gas target

A system of pipes and valves connects the roughing pump to any of several volumes: an inlet buffer tank; a cleaning trap; differential pumping stages; a Helium-3 handling system [currently (2019) not in use]. The valve interlock logic is designed to ensure that only one of these volumes is roughed down at a time.

Other connections permit: recirculation of gas through the cleaning trap or bypassing the trap; addition of gas from the buffer tank to the recirculation volume; controlled removal of gas from the recirculation system via an outlet buffer volume (3 litres). Pressure relief valves allow discharge from the cleaning trap or from the differential pumping volume into the roughing-pump exhaust line if pressures go above ambient air pressure.

The valves, pumps and gauges are controlled by EPICS interface to the PLC (programmable logic controller) of the DRAGON gas target and vacuum system. A right-button mouse click on the "Vacuum" button of the DRAGON EPICS menu will show the various pages associated with the gas target and separator vacuum. Figure 1 shows the page that gives overall control of the gas target Recirculation system.

EPICS page for the gas target Recirculation system. Control of the Roots blowers and ion gauges is on the page "Gas target".

In the most common mode of operation, recirculation mode, hydrogen gas is continuously injected into a windowless target cell. The gas which escapes through the cell apertures is recaptured, compressed, sent through a cleaning trap, and recirculated back into the target cell. Typical pressures are 4.5 Torr in the cell, 0.3 Torr in the manifold around the cell, 30-50 Torr after compression and in the trap, and 10-6 Torr at the HEBT and DRAGON Separator ends of the differential pumping stages.

Another operating mode also recirculates the gas, but sends it back directly in a line which bypasses the cleaning trap. This would be used if a special mixture of hydrogen "spiked" with heavy inert gas is to be used for normalization purposes.

Finally, in flow-through mode, gas might be pumped away by the roughing pump, with no recirculation. This last mode would likely be used only for limited periods, as it consumes gas at a high rate (approximately 250 Torr-litres/second for 4.5 Torr central cell pressure).

The cell pressure control is by a combination of the pressure of recirculating gas and the opening of a flow valve "GCV1" ( located under the "green table" surface of the gas target stand). Fine control of cell pressure may be done by opening/closing the fine flow control valve (GCV1), by pushbutton operation at a panel in the gas target rack. Operation at cell pressures above approx. 6 Torr requires opening of a manual valve which is in parallel with GCV1 (the "red-handled valve" located beside GCV1).

When the cleaning trap is to be used, addition of gas to the target system requires a considerable time: the trap material, X-13 zeolite molecular sieve, adsorbs hydrogen through mechanisms having rather different time constants. The first, rapid, adsorption takes place within seconds but the second mechanism has a time scale of 10-20 minutes. Loading of the trap is accomplished by isolating the Buffer Tank, filling it to 50 Torr from the supply cylinder, isolating it from the supply cylinder, opening it to the recirculation system, allowing gas to flow into the trap, and again isolating the Buffer Tank. By this procedure it is possible to keep track of the quantity of gas admitted to the system (and residing mainly in the trap).

During an experiment it may be necessary to reduce the inventory of gas in the trap/recirculation system, for example because the central cell pressure is to be reduced by a large factor. The inventory may be reduced in a controlled way by successively filling and emptying a known volume (3 litres) located between valves GOT1V and GOT2V. One "glug" through this volume should reduce the pressure at TRIN by about 10% when recirculating gas without the cleaning trap.

Operating the target

Pre-run checks

• Side plate secured in place on the target box. Gas supply and pressure gauge lines connected to the side plate. Check for obvious gaping holes in the pumping stages or turbo forelines.
• Check for readings in EPICS for the gas target pressure gauges (except ion gauges, which will be off at 1 atm.).
• Check in the gas shack that a cylinder with adequate H2 or He is hooked up to the DRAGON Gas Target manifold. The adjustable valve on the flowmeter should not be adjusted. The hand valve below the flowmeter should be open. Exception: if the line running from the gas shack to DRAGON is known to have been opened to air, the valve under the flowmeter should be closed and the line pumped out via the Inlet Buffer Tank before filling it with hydrogen.)

Pumping out the target

• Check that a fan is running on the hydrogen extraction line. Turn one on if necessary. Obtain the fan info page by selecting the option Hydrogen Extraction under the DRAGON menu item Vacuum.
• Select menu option Gas Recirculation, start roughing pump RP1. Open valve RV5, then valve RV2. Wait until the target cell pressure drops below 10 Torr. (Gauge CMGC reads 10 Torr full scale, CMTRIN 100 Torr full scale.)
• Select menu item Differential Pumping, start the Roots blowers: RB1B, RB1A, RB1, RB2A, RB2. Pressure in the gas cell and manifold should quickly drop to a small fraction of 1 Torr. If it does, pump for 10 minutes with the Roots blowers and roughing pump. Close RV2 and watch for pressure rise at CMTRIN. If the target had been open to moist air, expect pressure to rise at a rate of order 3 Torr/10 minutes due to water vapour. (100% relative humidity means 30 Torr partial pressure of water at 30C.) If the pressure rise is due to water vapour, the rate of pressure rise will decrease: open RV2 for 1 minute, then close it and repeat the observation of pressure rise at CMTRIN. If the problem was water vapour, this next measurement should show a substantial drop in rate of pressure rise: the expected behaviour from water vapour is a rate of order 0.1 Torr/10 minutes when the system has been pumped down 2-3 hours. The non-water leak rate should not be greater than 0.25 Torr/hour for the isolated system with Roots blowers on. Ideally, the target should be pumped out 1 day ahead of intended use, so that pressure rise overnight in the isolated system may be observed. If the time required to demonstrate that pressure rise is due to water and not an air leak is too long (e.g. beam delivery is imminent), the leak-tightness must be verified by means of a leak detector.
• Check that the flow of turbo-pump cooling water is On, by viewing Vacuum|TurboCooling. Start the turbo pumps TP1-TP9. They will take 15-20 minutes to come up to speed. Turn on the ion gauges IGU3 and IGD4 when turbos show "On" (light green icon). Pressures should be about 1E-6 Torr (or better) on U3 and D4. It is usual for the cell pressure manometer to be set with a small offset of order 20-30 mTorr. Note: If RP1 is being used as a backing pump for the turbo/Roots blower system, the "target empty" pressure may be not much better than the ultimate pressure achievable by RP1. This can happen if the system is in "recirculation" mode, with valves GCV1 and BYP1 open (or a return path through the Trap available). Additional note: if the zero offset of a capacitive manometer (CMTRIN, CMGC, CM1 or CMBT) is set too low, pump-down may cause one of them to give a negative reading. This will cause its controller to panic and indicate "noSens" on its front panel. EPICS will decide that the gauge isn't working properly and will trip devices whose interlock tests involve that gauge. Therefore, the capacitive manometer gauges should have their zero offsets adjusted to small positive values.

Filling the cleaning trap with hydrogen

• Pump out the cleaning trap (if it is at room temperature). Use roughing pump RP1, through valves RV5 and RVTR1. IVTR1 and OVTR1 should be closed. Close RVTR1 and verify that pressure CG4 remains steady.
• Raise the trap's dewar using the hand-cranked hoist, until the top lip of the dewar fits inside the sleeve of the trap-support lid.
• Connect a LN2 supply dewar to the filling line and pressurize to xxx psi using the dry nitrogen supply line connected to the nitrogen cylinder at MD2.   The LN2 Fill Valve controller is located in the gas target rack.  The EPICS page for the Recirculation system indicates 4 heights of the  LN2 level: "too low", "low", "high", and "too high". After 20-30 minutes the filling should stop, with 2 of the 4 LN2 level lights on.
• If possible, allow 6-8 hours for the zeolite in the trap to cool down. This will speed up the filling process, but is not essential.
• Pump out the Inlet Buffer Tank through roughing pump RP1 via valves RV5 and RV4. Close RV4 and let gas into the Inlet Buffer Tank through FILLV, up to a pressure of 300 Torr. Close FILLV, open RV4 to pump out the bufer tank. Repeat the previous two steps 2 more times to flush out the tank and the supply line from the gas shack.
• Close RV2, BYP1, RV4 and GOT1V or GOT2V. Open IVTR1, OVTR1. Flow control valve GCV1 should be 20% open. Roots blowers must be on.
• Open FILLV and let 50 Torr into the Inlet Buffer tank. Close FILLV. Open GINV to let gas flow from the buffer tank into the trap via the gas cell. Pressure should quickly settle at about 5 Torr on CMTRIN. Close GINV.
• Open FILLV and fill the buffer tank to 55 Torr, i.e. 50 Torr above the initial pressure. Close FILLV. Open GINV while watching GMTRIN to note how high the pressure "spikes" before settling. As the trap fills, the working pressure increases and the spikes reach higher pressures. It is important to avoid spikes above 60 Torr because that will cause the Roots blowers to trip off, to protect against possible over-heating.
• Continue filling the trap with 50 Torr  fills, until CMTRIN reaches about 40 Torr. This should take 15 fills ("glugs"), depending on how long the zeolite has been cooling. Record the number of units of hydrogen (1 unit = 50 Torr × 6 litres) when recording target status in the Equipment Status elog.
• Wait. The CMTRIN pressure should drop, with a time constant of about 20 minutes. When it drops below 30 Torr, top up to a pressure of order 45 Torr. Repeat the wait/fill until the pressure steadies in the range 40-45 Torr. At the last stages it may be prudent to go to smaller fills to make sure the transient pressure spike stays below 60 Torr.
• In total, the trap should hold 15 "glugs" of hydrogen at a CMTRIN pressure of 45 Torr. If it wants much less or much more than this, consult an expert.
• The Hydrogen Extraction fan must be left running whenever the trap is loaded with hydrogen.

Filling the cleaning trap with helium

The procedure for filling with hydrogen should be followed.   Repeat fills through the Inlet Buffer to reach required CMTRIN; adjust the flow valve (and possibly the "Red Handle" valve) to get the desired cell pressure CMGC.

The difference from the hydrogen procedure is that the zeolite in the trap does not adsorb helium the way it adsorbs hydrogen.

The ratios between the cell pressure CMGC and CMTRIN, IGU3 or IGD4 are different for helium and hydrogen, because the difference in molecular masses leads to different pumping speeds and tube conductances.

Circulating gas through the target, trap in use

• With trap isolated (IVTR1 and OVTR1 closed), rough out the differential pumping stages, cell, and blowers through RV2 and RV5. Close RV2.
• Open IVTR1.
• Open OVTR1.
• Adjust flow control valve GCV1 to get desired pressure in the cell (gauge CMGC). This is done by Open or Close push-buttons on the valve-motor control in the gas target electronics rack. If H₂ pressures above about 5 Torr are needed, the "red-handled" valve in parallel with GCV1 should be opened: this valve should be cracked open very cautiously, with a second person watching the CMGC pressure gauge.

Circulating gas through the target, trap not in use

• Pump out the target, if that has not already been done. Turn on roughing pump RP1, open valves RV5 and RV2 and rough out the system until the pressure CMTRIN or CMGC is less than 10 Torr. Turn on the Roots blowers (EPICS page Differential pumping) and pump until CMTRIN and CMGC reach a minimum reading. (Both gauges may have small offsets, to avoid Error message if the zero-point should drift into negative readings.)
• Close RV2 and observe the pressure rise, if any, on CMTRIN. It may be necessary to repeat several cycles of pumping and isolation to decide whether a pressure rise signals an air leak or outgassing of water vapour or something else.
• Open or close the flow control valve GCV1 to its 50% setting using the control buttons on the control panel in the Gas Target rack. Verify that the "red handle" valve GCV2 is closed.
• Isolate the recirculation loop from the roughing pump by closing valve RV2. (RV5 stays open.)
• Open valve BYP1, so recirculation bypasses the trap.
• Open RV4 and FILLV to pump out the Inlet Buffer Tank and supply line from the gas shack. Close RV4 and let the Buffer tank fill to 50 Torr. Close FILLV, open RV4 and pump out the buffer tank. Close RV4.
• Open FILLV and fill Buffer tank to about 50 Torr. Close FILLV.
• Open GINV long enough for the Buffer tank pressure to stabilize (expect about 12 Torr at CMTRIN), then close GINV.
• From the ratio of pressures at CMTRIN and in the target cell (CMGC), compute how much more gas is required to raise CMGC to the maximum required during the experiment, assuming a linear relationship between the amount of gas, the pressure CMTRIN and the pressure CMGC. Provided the projected CMTRIN is < 50 Torr, continue adding gas "quanta" via the 6-litre Buffer tank; add no more than 300 Torr-litre of gas at a time to the buffer tank (50 Torr × 6 litres).
• Record in the Equipment Status elog the amount and type of gas loaded into the recirculation system.
• Do not fill to higher than CMTRIN=50 Torr. (Trip limit of Rootsblower RB1B is 60 Torr.) If desired cell pressure cannot be attained with CMTRIN=50 Torr and GCV1 75% open, carefully open the "red-handled" valve which is in parallel with CGV1.
• If too much H2 gas was let into recirculation from the Buffer tank, a controlled amount may be pumped away through the roughing pump by using valves G0T1V/G0T2V and the 3-litre volume between them. Normally, G0T2V should be closed and G0T1V open to rough out the volume. Close 1V, open 2V, close 2V and open 1V to release a known amount of gas. Repeat if necessary. Record the changed hydrogen inventory.