The potential hazards associated with the operations of cryogenic targets in Hall C are reviewed and the efforts to ameliorate these hazards are described.
The Hall C cryogenic target system typically
consists of two completely instrumented targets, normally
containing cryogenic H
and D. In addition
a third target loop is installed in the scattering chamber
but these cells are usually not connected to a full gas handling system.
This third loop is used as a spare or as a Helium target.
The hydrogen and deuterium targets present a number of potential hazards. The most notable of these are associated with the fire/explosion hazard of the flammable gas, the hazards connected with the vacuum vessel and those of handling cryogenic liquids (ODH and high pressure). In this document the hydrogen target will be referred to but the deuterium target is essentially identical and almost all comments apply to both targets.
Hydrogen and deuterium are
colorless, odorless gases and hence not easily detected by human senses.
Hydrogen air mixtures are flammable over a large range of relative
concentrations: from 4 
to 75 H by volume. Detonation
of explosions can occur with very low energy input, less than
that required by mixtures of air and gasoline. At temperatures above
-250
C hydrogen gas is lighter than (STP) air and hence will rise.
At atmospheric pressure the ignition temperature is approximately
1000 degrees F but mixtures at pressures of 0.2 to 0.5 Atm can be
ignited at temperatures as low as 650 degrees F. Hydrogen mixtures
burn with a colorless flame [6].
The total volume of liquid hydrogen in the target depends on which type of target cell is installed. For the typical machined cell design with a 4 cm and a 15 cm cell in each loop, the total liquid volume is about 7.5 liters per loop. The volume changes between the liquid state and gas at STP by a factor of about 800. Thus filling the target requires about 6,000 STP liters of hydrogen (roughly the contents of a standard high pressure gas cylinder).
Each target is attached to a large, 1000 gallon (3800 liter) recovery tank. These tanks are charged to 50 PSIA (3.4 ATM) when the targets are warm. Each target is thus connected to about 13000 STP-liters of inventory. In addition, there is usually a partial high pressure cylinder of target gas in the racks behind the gas panels. The total explosive gas inventory associated with the targets is thus quite substantial.
The basic idea behind safe handling of any flammable or explosive gas is to eliminate oxygen (required for burning) and to prevent exposure to any energy source that could cause ignition. In the Hall C environment the most likely source of oxygen is of course the atmosphere and the most likely ignition sources are from electrical equipment.
Hall C contains a lot of electrical equipment and almost all of it could serve as an ignition source in the presence of an explosive oxygen and hydrogen mixture. We have made an effort to minimize the dangers from the equipment that is most likely to come into contact with hydrogen gas.
There are a number of electrically powered devices associated with the target gas handling system. The solenoid valves on the gas panels are approved for use in a hydrogen atmosphere as are all the pressure transducers in the system. The AC power for the solenoids is carried by wires which are contained in either hard or flexible conduit. There are also LED's on the gas panels that provide an indication as to the status of the valve solenoids. These are powered by a 24 Volt DC supply (of which the power leads are not shielded). The readouts for the pressure transducers are mounted on the gas panels and the AC power for these readout units is in conduit. All the pressure transducers have 4-20 mA outputs.
In addition to the electrical devices in the gas handling system there are a number of devices inside of or mounted on the scattering chamber.
All the devices which are in the scattering chamber must have their power delivered to them by wires in vacuum. The insulation of these wires should be radiation resistant, so Kapton has been used where available.
The following electrical items are in close proximity to or are actually in the hydrogen system.
s per foot
wrapped on a G10 or an anodized aluminum carrier board.
The maximum power available to this
heater is of order 800 Watts. The heater has a DC resistance of
2 s and is driven by a 40 Volt, 25 Amp power supply.
The current and voltage supplied to this heater are monitored by the control
system and there is a software power maximum enforced on the power setting
of this heater. The heater is connected to
the outside world by 8 gauge stranded wire with Teflon insulation.
The lower portion of this wire is encased in a fiberglass sheath which
will serve as additional insulation should the Teflon be damaged by
exposure to radiation. No noticeable deterioration of the
insulation occurred in the first two years of running.
A to excite them (into up
to 7500 ).
The Cernox resistors are connected to the outside world
with quad strand 36 gauge phosphor bronze wire with Formvar insulation.
The Carbon resistors are wired with 30 gauge Kapton insulated
copper stranded wire.
A. The pressure at the entrance to the roughing pump
is measured by a pair of Hastings thermocouple gauges (models DV-6 and DV-4D).
The DV-4D requires 0.009 Watts (0.029 AC Amps and 0.32 AC Volts). The
DV-6 requires 0.008 Watts (0.021 AC Amps and 0.38 AC Volts).
A vacuum interlock is now part of the target system. This interlock is triggered whenever the vacuum in the scattering chamber crosses above 5-10 Torr. It kills the power to the target high power heater, to the cold cathode gauge, and it closes the gate valve to the turbopump.
There are four flammable gas detectors installed to provide early detection
of hydrogen/deuterium leaks. These detectors are sensitive (and calibrated)
over the range from 0 to 50 Lower Explosive Limit (LEL) of hydrogen.
The electro-chemical sensors were manufactured by Crowcon Detection
Instruments LTD and
the readout (four channels) was purchased from CEA Instruments, Inc
(The Gas Master Four System). The readout unit provides two alarm
levels per channel. The low level alarm is tripped at 20 LEL
while 40 LEL activates the high level alarm.
These detectors require periodic calibration. This
calibration is checked by the Polarized Target Group.
Each channel has a relay output for both low and high level alarm
states and there is also a set of common relays for both alarm
levels (these common relays respond to the ``logical or" of the sensor inputs).
The common relays are connected to the Fast Shut Down System, FSD,
which removes the beam from the hall.
The most important aspect of hydrogen safety is to minimize the possibility of explosive mixtures of hydrogen and oxygen occurring. To this end the gas handling system has been made of stainless steel components (wherever possible) and as many junctions as possible have been welded.
The pressure in the gas handling system is monitored in numerous
places. Most importantly the absolute pressure of the target is
viewed by two pressure transducers, one on the fill line, PT9013
for H and PT9083 for D, and one on the
relief line, PT9023 for H and PT9093 for D.
These pressures are also measured by manual gauges. The fill line gauges
are PI9031 for H and PI9101 for D. The relief line
gauges are designated PI9013, H and PI9083, D.
The target cells themselves represent the most likely failure point
in the hydrogen system. The outer wall and downstream window
of each cell is typically made of
0.004 to 0.005 inch thick 3004 aluminum (Coors beer cans in a former
incarnation). There are usually two cells connected with metal gaskets
to each cell block, one 15 cm long
and one 4 cm long. Both cells have an outer diameter of
approximately 2.5 inches.
The upstream windows of the cells are made from 0.0028 inch thick 5052
aluminum.
These windows are machined out of 1.75 inch diameter (0.065 inch wall)
upstream window tubes which are in turn connected to the cell block
with metal gaskets.
The cell block components have been pressure tested hydrostatically at CEBAF.
Several beer cans have been pressure tested to over 100 PSIG.
Upstream windows of both 5052 and 6061 aluminum have been tested
to similar pressures. Finally, the entire completed cell block
assemblies were pressurized to 75 psig with helium gas, 
1.5 times the
maximum pressure expected (50 psia) during a sudden loss of vacuum incident.
The gas handling and controls system have been designed to prevent excessive pressure build up in the system in order to protect the target cells from rupture. The relief line of the target leads directly to the large recovery tanks. These tanks act as huge ballast and tend to damp out pressure excursions. In the event of a target disruption, such as loss of coolant, the gas evolved during boiloff goes to the recovery tanks. There are two sets of manual valves between the target and the recovery tanks which are used to isolate the tanks when the target is not in use. These valves must be locked in the open position before the targets are filled.
In addition to the tank there are a number of other pressure reliefs in the system. Under normal conditions (and even pretty abnormal conditions) these will never be needed but they are provided for extra safety.
All target pressure reliefs are connected to a dedicated hydrogen relief line 2 inches in diameter. The relief line is inerted with helium provided by the ESR 300K supply. The relief line leads to an elevated stack outside, on top of the Hall C dome behind the recovery tanks. On top of the stack is a parallel plate relief set at about 2 psi. Thus any vented target gas is placed in an inert environment until it is released outside of Hall C.
In addition to the reliefs on the gas handling system described above, the scattering chamber itself has a one PSIG relief, RV9064. This is the path that the hydrogen will take in the event of a cell failure.
The scattering chamber is always leak checked before service but obviously the possibility of vacuum loss can not be eliminated. The most likely sources of vacuum failure are:
The two spectrometer windows are both made from aluminum. The SOS window
is fifteen inches high and subtends 77 degrees
(24 to 101).
The window diameter on its frame is 47.5 inches.
The SOS window is made of
0.020 inch thick 2024 T3 Aluminum clad foil.
The HMS window is eight inches high and subtends 97.5 degrees
(5.5 to 103). The window diameter on its frame is 45.0
inches.
This window is made of 0.016 inch
thick 2024 T3 Aluminum clad foil.
These scattering chamber windows have been hydrostatically tested to 28 psig (1.9 atm),
well above the required safety factor of 1.25.
The small beam exit window has an even larger safety margin according to stress analysis. This window was purchased from a commercial vendor (Brush-Wellman).
In the unlikely event of a catastrophic vacuum failure it is important that the relief line of the targets be sized such that it can handle the mass flow caused by the sudden expansion of its cryogenic contents due to exposure to the heat load. A calculation has been performed which models the response of the system to sudden vacuum failure. That calculation indicates that the relief plumbing is sized such that the flow remains sub-sonic at all times and that the maximum pressure in the cells remains well below their bursting point. The calculation is presented in detail at https://polweb/hallc/Cryotargethttps://polweb/hallc/Cryotarget. However, that document was created in 2004, before the upgrade to the Hall C scattering chamber and target loops took place. The subsequent upgrade essentially made the new Hall C scattering chamber and target loops identical to those described in the document for Hall A. Therefore, the calculations appropriate for the larger Hall A volumes should be used, not those appropriate for the older Hall C system which is no longer in use as of spring, 2007.
The calculation was performed by following methods in an internal report from the MIT Bates laboratory [7]. The formulas and algorithm in the report were incorporated in two computer codes and those codes were able to reproduce results in the report (hence they represent an accurate implementation of the Bates calculation).
The calculation can be logically broken into two parts. First,
the mass evolution rate is calculated from geometric information
and the properties of both the target material and vacuum spoiling gas.
The principal results of this first stage are the heat transferred
per unit area, q, the boiloff time, t
, and the mass evolution rate, w.
Second, the capability of the plumbing to handle the mass flow
is checked. The principle result of this second step is the
maximum pressure in the target cell during the discharge, P
.
The calculations indicate that the reliefs are adequate.
The reference document referred to above also describes the safety
calculation for potential over-pressurization of the Hall C scattering
chamber. The scattering chamber can vent through relief valves and a
burst disk into the dedicated hydrogen vent in the even of a cell
rupture. The maximum pressure inside the 
1900 liter scattering chamber in an
event of this type, as described in the
https://polweb/hallc/CryotargetHydrogen Target Safety
Assessment Document is expected to be only about 4 psig.
In the unlikely event that a line which carries helium coolant were to rupture the large chamber relief valve is capable of handling the full coolant flow rate.
This is really more an issue of target stability than one of safety. However, a target with a carefully regulated temperature will presumably not undergo worrisome pressure changes.
The targets are normally temperature regulated using a software PID loop that takes as input the temperature as given by one of the Cernox resistors and outputs a control voltage to the power supply of the high power heaters.
Solid hydrogen is more dense than the liquid phase so freezing does not endanger the mechanical integrity of a closed system. The chief hazard is that relief routes out of the system will become clogged with hydrogen ice making the behavior of the system during a warmup unpredictable.
The freezing point of deuterium is higher than that of hydrogen and higher than the temperature of the gas used for cooling.
Experience has shown that it is difficult to freeze the targets as the efficiency of the heat exchangers is ruined by even a small amount of ice. Still care needs to be taken to keep the targets free from air or water contamination and a trained operator should always be present.
The total volume of the targets is relatively small with the entire
scattering chamber containing only 6500 STP liters of target
fluid when both targets are full. As the scattering chamber
is located in the middle of Hall C
(ie. not in a confined area) which has a volume of approximately
26 x 10
leters, the ODH hazard is minimal.
The target controls have been implemented with the EPICS control system and with hardware very similar to that employed by the accelerator. The basic control functions reside on a VME based single board computer. The graphical interfaces to the control system require one of the Hall C computers to be present as well.
All of the instrumentation for the target is downstairs in Hall C. Most of the equipment (in fact all of the 120 V AC equipment) is on an Uninterruptable Power Supply (UPS). The equipment whose power is not on UPS is
The principal functions that the control system performs are:
Whenever hydrogen (or deuterium) is condensed in the system, a responsible person must be on duty in the counting house. This individual is designated the target operator. He/she must be authorized to operate the target and the local target expert must keep a list of all the authorized individuals. In order to become eligible to act as a target operator a person must first be trained by one of the target experts, and then sit at least a few hours with an already trained target operator as a practical. The training from one of the experts forms the ``theory'' part of the training and consists of a prepared talk given by the expert during which questions are strongly encouraged. The practical training typically takes place in the Hall C counting house and consists of a guided walk through of the controls system.
Target Operators must be familiar with the documents
http://www.jlab.org/~smithg/target/target_operator.pswhich describes the things a target operator should know how to do,
and
http://www.jlab.org/~smithg/target/tgt_howtos.pswhich describes how to do those things. An extensive source of
information of assistance to the target operator is provided at
http://www.jlab.org/~smithg/target/Hall_C_Cryotarget.html#Standard-Pivot-TargetsPump-purge procedures, cooldown procedures, and emptying procedures
are the responsibility of the JLab Target Group. In case of
problems the Target Operator's first point of contact is one of the
Target Experts, typically the Hall C Contact Person first, then the
target experts not belonging to the JLab Target Group, then the JLab
Target Group members. Names and pager/phone numbers are included in
the above links.
The principal problem in hydrogen safety is the prevention of explosive hydrogen and air mixtures. The mechanical aspects of this system have been built to minimize the possibility of a hydrogen release. The components that are most susceptible to failure have been tested and the results of these tests indicate that the system should stay intact under the most extreme conditions that are likely to be encountered.
In the unlikely event that hydrogen is released, a mechanism is in place to detect its presence quickly and remove at least the beam as a possible ignition source.
A control system has been developed to allow the careful monitoring of the target systems behavior and it will respond to any aberrant behavior in order to minimize the consequences. Essentially the entire control system has been placed on uninterruptable power so that short power outages can be waited out.