A portable cryostat for the cold transfer of polarized solid HD targets:HDice-I NIM

A portable cryostat for the cold transfer of polarized solid HD targets:HDice-I NIM
Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
Contents lists available at ScienceDirect
Nuclear Instruments and Methods in
Physics Research A
journal homepage: www.elsevier.com/locate/nima
A portable cryostat for the cold transfer of polarized solid HD
targets: HDice-I
C.D. Bass a,1, C. Bade g, M. Blecher n, A. Caracappa b, A. D'Angelo k,a, A. Deur a, G. Dezern a,
H. Glueckler c, C. Hanretty m, D. Ho d, A. Honig h,2, T. Kageya a,n, M. Khandaker f, V. Laine a,i,
F. Lincoln b, M.M. Lowry a,b, J.C. Mahon g, T. O'Connell j, M. Pap c, P. Peng m, B. Preedom l,
A.M. Sandorfi a,b,n, H. Seyfarth c, H. Stroeher c, C.E. Thorn b, X. Wei a,b, C.S. Whisnant e
Thomas Jefferson National Accelerator Facility, Newport News, VA 23606, USA
Brookhaven National Laboratory, Upton, NY 11973, USA
Forschungszentrum Jülich GmbH, Jülich, Germany
Carnegie-Mellon University, Pittsburgh, PA 15213, USA
James Madison University, Harrisonburg, VA 22807, USA
Norfolk State University, Norfolk, VA 23504, USA
Ohio University, Athens, OH 45701, USA
Syracuse University, Syracuse, NY 13210, USA
Université Blaise Pascal, Clermont-Ferrand, 63177 Aubiere, France
University of Connecticut, Storrs-Mansfield, CT 06269, USA
Universita’ di Roma “Tor Vergata” and INFN Sezione di Roma2, 00133 Roma, Italy
University of South Carolina, Columbia, SC 29208, USA
University of Virginia, Charlottesville, VA 22903, USA
Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, USA
art ic l e i nf o
a b s t r a c t
Article history:
Received 17 September 2013
Received in revised form
23 October 2013
Accepted 24 October 2013
Available online 31 October 2013
A device has been developed with moveable liquid nitrogen and liquid helium volumes that is capable of
reaching over 2 m into the coldest regions of a cryostat or dilution refrigerator and reliably extracting
or installing a target of solid, polarized hydrogen deuteride (HD). This Transfer Cryostat incorporates
a cylindrical neodymium rare-earth magnet that is configured as a Halbach dipole, which is maintained
at 77 K and produces a 0.1 T field around the HD target. Multiple layers provide a hermetic 77 K-shield as
the device is used to maintain a target at 2 K during a transfer between cryostats. Tests with frozen-spin
HD show very little polarization loss for either H ( 1 7 2%, relative) or D (0 7 3%, relative) over typical
transfer periods. Multiple target transfers with this apparatus have shown an overall reliability of about
95% per transfer, which is a significant improvement over earlier versions of the device.
& 2013 Elsevier B.V. All rights reserved.
Solid hydrogen deuteride
Frozen-spin target
Spin relaxation
Halbach rare-earth magnet
1. Introduction
A successful description of the excitation and decay modes of a
composite system is a basic test of how well the underlying forces
are understood. Quantum-Chromodynamics (QCD) has been able
to account for many of the properties of hadrons, although
the excited states of the nucleon continue to pose significant
challenges. These excited states are short-lived, and only the
lowest-energy nucleon excitation – the Δð1232Þ resonance – is
Corresponding author. Tel.: þ 1 7572695457; fax: þ1 7572696418.
E-mail addresses: [email protected] (C.D. Bass),
sandorfi@JLab.org (A.M. Sandorfi).
Current address: Le Moyne College, Syracuse, NY 13214, USA.
0168-9002/$ - see front matter & 2013 Elsevier B.V. All rights reserved.
moderately well-isolated. Higher-energy excitations appear as
broad and overlapping resonances in pion scattering and meson
photoproduction, which can only be disentangled through the
application of detailed partial wave analysis (PWA) techniques.
Furthermore, many more excited states have been predicted [1,2]
than have been extracted using PWA.
A number of factors complicate the study of excited nucleon
states [3]. Most important is the lack of experimental data on the
majority of the 16 possible spin matrix elements in pseudoscalar meson photoproduction, which has created ambiguities in
PWA [4]. To address this shortcoming, experiments have been
conducted with frozen-spin targets of solid hydrogen deuteride
(HD) at the Brookhaven National Laboratory (BNL) in the Δð1232Þ
resonance region [5], and at the Thomas Jefferson National Accelerator Facility (JLab) throughout the higher-energy resonance
C.D. Bass et al. / Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
excitation region [6]. The goal of the latter has been a complete
experiment, in which most (if not all) possible reaction observables involving explicit spin orientations of the beam, target, and
recoiling baryon are measured. More than half of the spin observables cannot be determined without a polarized target, and the use of
a polarized target in conjunction with measurements of double- and
triple-polarization asymmetries can determine all 16 of them with a
single orientation of target polarization [4].
As a polarized target for nuclear and particle physics, frozenspin HD (known locally as HDice) has many attractive features.
Apart from minimal contributions from the target cell, the HD
material is low-Z, pure, and free of background sources. In
addition, HD contains a single polarizable neutron and this freedom from dilution gives HD experiments focusing on reactions
with polarized neutrons, such as JLab E06-101, a particularly high
figure of merit [6]. However, the production of frozen-spin HD is
considerably more complex than that of conventional polarized
nuclear targets, requiring several specially designed and constructed cryostats and target manipulation tools, as well as a
number of commercial dewars and dilution refrigerator and
magnet systems.
This is the first of a series of papers that describe the preparation, polarization, and handling of frozen-spin polarized HD
targets. The ability to move these targets between cryostats
without significant polarization loss is crucial because the targets
are produced and polarized in a location separated from the
experimental hall and are transported to a detector once they
have reached a frozen-spin state. This paper describes the design
and operation of a Transfer Cryostat (TC) that has been used at BNL
and JLab for moving HD targets between cryostats. These operations are complex and typically require several hours. The TC
discussed here is the current and most successful design of a series
of devices that evolved during the HD target program. (Condensed
discussions of earlier designs can be found in Refs. [7–9]).
2. The HD target production and run sequence
We begin with a brief description of the typical sequence used
to create a frozen-spin polarized HD target and transport it to an
In-Beam Cryostat located in an experimental hall for use in nuclear
physics experiments. The HD target production requirements and
cold transport steps have determined the TC design characteristics.
2.1. Production of a HD target
An empty target cell (Fig. 1) is mounted on the end of a
capillary tool and lowered into the cold-bore of a variabletemperature cryostat referred to as the Production Dewar (PD).
The target cell is composed of a shell of polychlorotrifluoroethylene3 (pCTFE) that is fixed to a copper target ring possessing an
interior left-handed (LH) M26 1 thread and exterior righthanded (RH) M35 1 thread. Thin aluminum cooling wires span
the interior of the pCTFE shell and are soldered into the target
ring (800–2100 in number, depending on the diameter of the
target cell).
The target cell is cooled to just above the triple-point temperature of HD (16.6 K), and distilled HD gas is condensed through the
capillary and into the target cell. The HD gas contains trace
amounts of H2 and D2 that are necessary for the polarization
process [9]. Once the target cell has been filled with liquid HD, the
temperature is lowered, creating a solid HD crystal. The cell is then
The chemical composition for polychlorotrifluoroethylene is C2ClF3.
Fig. 1. Cross-section of a typical HD target that was used during the E06-101
experiment at JLab. A shell of pCTFE is attached to a copper target ring that has an
interior left-handed thread (engaged by tools to move the target cell) and exterior
right-handed thread (used to mount the target in a dewar). Several hundred thin
aluminum wires span the length of the target cell and are soldered at one end into
the target ring. A 50 mm long cylindrical crystal of HD is grown in the target cell by
condensing and then freezing distilled HD gas. The embedded aluminum cooling
wires allow heat, which is generated in the HD crystal during the polarization
process and in-beam experiments, to escape from the target and into the mixing
chamber of a dilution fridge.
lowered to the 4 K copper threads of the PD where the capillary
tool is removed.
The HD target is then positioned within a set of crossed coils
housed inside the PD that are used for NMR studies [10]. The
polarizations P(t) for H and for D in the target increase to their
thermal equilibrium values (as determined by the PD field and
temperature) with a time dependence parameterized as
PðtÞ ¼ P TE ð1 e t=T 1 Þ
with time t, equilibrium polarization PTE, and relaxation time T1.
At this stage the relaxation time is of the order of seconds for H
and 10 s to 100 s of seconds for D, depending on the concentrations of the H2 and D2 impurities as well as the temperature and
field strength in the PD; consequently, the H and D quickly reach
their respective thermal equilibrium polarizations. These polarization levels are small (less than 0.001 under PD conditions) but are
more than adequate to calibrate the NMR electronic circuit.
2.2. Polarizing the HD target
The HD target is extracted from the PD with the use of the TC,
which has a liquid helium (LHe) volume that can rotate and
translate within the TC (Section 3.1) and extend into the cold bore
of another cryostat. The LHe volume incorporates a M26 1 LH
thread on its lower end that matches the interior thread of the
target ring. The LHe volume is extended into the PD until it
contacts the target, after which counterclockwise rotation of the
LHe volume screws the TC thread into the HD target cell until it is
fully engaged. Continued counterclockwise rotation of the LHe
volume causes the exterior M35 1 RH thread on the HD target
ring to unscrew from the cold-bore of the PD. The LHe volume and
attached target are withdrawn into the TC, which is then moved
from the PD to an Oxford-1000 dilution refrigerator and superconducting magnet system (DF). The transfer operation is shown
schematically in Fig. 2. The PD and DF are arranged in pits so that
their attachment points are at a common level.
Once attached to the DF, the TC LHe volume and HD target are
extended and rotated clockwise to engage the exterior thread of
the target ring into the matching thread in the cold-bore of the DF.
Once the HD target is fully engaged in the cold-bore, continued
rotation unscrews the thread on the end of the TC LHe volume
from the target ring, and the LHe volume is then withdrawn into
the TC. The dimensions of the TC are constrained by the narrow
cross-section and depth of the cold-bore and polarizing magnets
within the DF.
C.D. Bass et al. / Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
After the HD target is secured within the DF, the field is raised
to 15 T and the temperature lowered to the 15 mK region. Because
of wavefunction symmetry requirements, the first rotational states
of the H2 and D2 impurities present in the HD are polarizable and
metastable. These polarize rapidly and transfer their polarization
to HD, leading to typical values in the range P(H) 6075% and
P(D) 18 75%. As the metastable impurities decay away (with
relaxation times of 6 and 18 days for H2 and D2, respectively), the
solid HD reaches a frozen-spin state [9]. A target is nominally held
in the DF at high field and low temperature for approximately
2–3 months to allow the spins of the H and the D to decouple from
the HD lattice.
While relaxation times of months to years are easily reached at
high field and very low temperatures, the size constraints of the TC
limit the strength of the transport field and the operational
temperature. This results in target transfers where the HD experiences significantly shorter relaxation times that vary approximately with B/T (on the order of a day for PH and a week for PD
under TC conditions).
The HD target is transferred to a Storage Dewar (SD) and held at
1.6 K and 5 T until it is needed for a nuclear physics experiment.
The SD and TC are subsequently moved to the experimental area.
From a staging area within the experimental hall, the HD target is
extracted from the SD and loaded into the In-Beam Cryostat (IBC)
for the experiment.
2.3. Transporting the HD target to the experimental hall
2.4. Post-experiment calibrations
The HD target is transferred from the DF and back into the PD
for evaluation of its polarization using NMR. For this transfer, the
TC must provide both a holding field as well as low temperature.
The IBC contains an NMR polarimeter that provides frequent
measurements during an experiment. This is cross-calibrated with
the polarimeter of the PD where the initial thermal equilibrium
polarization measurements have been performed. Transfer losses
have been estimated by carrying out multiple transfers between
the PD and the IBC. As a final cross-calibration after the experimental run, the HD target is transferred from the IBC to the PD
(within the experimental hall) for a final polarization measurement and then evaporated for gas analysis.
3. Design and construction features
Fig. 2. Schematic of a typical cryostat arrangement during the stages of an HD
transfer. Cryostats used in the polarization cycle are positioned in pits so that their
TC attachment points are at a common level.
The cold transfer of a target between cryostats with the TC is
accomplished by engaging and rotating the threads of the target
ring (Fig. 1), as discussed above. A target can be unscrewed from
one cryostat, withdrawn into the TC, and then inserted and
screwed into another cryostat. During transport between
Fig. 3. Schematic diagrams of the Transfer Cryostat (TC). The upper figure shows the TC in its retracted configuration, which can be used as a self-contained dewar to carry a
polarized HD target. The lower figure shows the extended configuration, which is used to reach into the cold-bore of another cryostat when coupled through a vacuum gate
valve. During transfer operations, the TC is supported vertically with the top (containing the indicated Pulley) shown at the right of the figure.
C.D. Bass et al. / Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
cryostats, a permanent magnet array in the TC provides a holding
field for preserving target polarization.
The core design of the present TC was originally developed at
BNL for handling frozen-spin HD targets during experiments at the
Laser-Electron-Gamma-Source (LEGS) [5,9]. The overall dimensions of the TC were constrained by the geometry of the target
cell, the dimensions of the cryostats used for production and
polarization, and the requirement of maintaining target polarization during transport between cryostats.
Operational reliability is a crucial factor due to the number of
successful transfers needed to place a polarized target in beam. As
outlined in Section 2, the successful utilization of an HD target for an
experiment usually requires five transfers with the TC. Because of the
significantly shorter spin relaxation times under the TC conditions,
the time elapsed in each TC transfer must be minimized. Earlier
generations of TCs used with HD targets had transfer success rates
between 65% and 75% [7–9]. Such efficiencies result in an overall
success rate of less than 25% for five transfers, (0.75)5, which makes
the running of a major experiment difficult. Since its original
construction and operation at BNL, several important additions and
modifications have been incorporated into the design of the TC to
improve operational reliability. The resulting device has a transfer
success rate of about 95%. Thus, five transfers using the current
device succeed more than 75% of the time, which is an efficiency
factor that can be accommodated within a major experiment. The
current design described here was used during the E06-101 experiment [6] at JLab.
3.1. Moveable cryogen volumes
The liquid cryogen spaces of the TC are fabricated as a pair of
concentric cylindrical and toroidal volumes surrounded by an
insulating vacuum space formed from a series of highly compressible bellows, as shown in Fig. 3. The 3.3 m long central cylindrical
volume forms a 1.1 liter LHe volume. The lower end of the LHe
volume terminates in a mechanical connection to the target.
The top half of the LHe volume is jacketed by a second wall that
is vacuum tight at the upper end but is open to vacuum at the
lower end. A port at the top of the volume allows the insertion of a
liquid helium transfer lance for cryogen filling. During TC cooling,
the transfer lance mates to a 1.3 m long permanent thin-walled
tube that effectively extends the lance and directs the LHe to the
bottom of the volume. This prevents vapor blockages and enables
a rapid cool-down. Additional ports at the top of the volume
contain an electrical feedthrough for temperature and liquid-level
sensor leads and a vent for helium boil-off.
Surrounding the LHe volume and separated by vacuum is a
320 mm long, 1.8 l toroidal liquid nitrogen (LN2) volume that is
mechanically attached to the top of the bellows assembly.
A 1380 mm long, double-walled tube extends down from the
bottom of this toroidal volume providing a 77 K-shield with a total
length of 1700 mm for the LN2 volume. It is this extended section
that enters a mating cryostat during target insertion or extraction
operations. Two 3 mm thick insulating spacers made of G10 epoxy
fiberglass are mounted on the inside surface of the LN2 extension.
The spacers are separated by 50 cm and align the central LHe tube
by forming a slightly larger aperture at four points on the
circumference. A set of three access tubes extends from the LN2
volume through the top of the vacuum bellows assembly. Two of
the tubes serve as fill and vent lines for liquid nitrogen, and the
third tube carries the leads for a capacitive liquid level sensor for
the toroidal volume.
The LHe and LN2 volumes are insulated from room temperature
by a 15 l vacuum space formed by a 1.9 m long series of highly
compressible welded bellows, a sliding seal assembly at the top
of the bellows, and a gate valve at the bottom of the TC. This
surrounds the jacketed length of the LHe volume. A winch at the
bottom of the TC supports the vacuum bellows assembly against
vacuum forces through a cable via a pulley located at the top of the
TC. This winch and cable system controls the motion of the LHe
volume and the bellows. Compression of the bellows permits
vertical motion of the top of the bellows assembly, the connected
LN2 volume, and the central LHe volume, which penetrates
through the sliding seal assembly at the top of the bellows. The
configurations of the TC in its fully retracted and fully extended
positions are shown in Fig. 3. In the latter configuration, the TC can
extend a 2 K connection deep into the cold regions of all cryostats
used with polarized HD. During TC extraction, the mechanical
advantage of the winch is used to pull the LHe tube up against the
vacuum force. When a stainless steel ring, clamped around the
outside of the LHe volume at a point 2 m above the target, meets
the top plate of the bellows assembly, the LHe volume is then
entirely within the LN2 volume. Further action of the winch then
expands the bellows as it raises both LHe and LN2 volumes.
The weight and motions of the TC are supported and guided by
a 100 mm 50 mm by 3.3 m long aluminum strong-back tube that
spans the length of the cryostat. On top of the strong-back is a
linear track. Five sliding bearing blocks ride the track and guide
and support the vacuum bellows assembly. Additionally, the top of
the LHe volume is guided and supported by a sliding bearing block
riding on the track.
The outer jacket wall of the LHe volume makes contact with a
pair of sliding O-ring seals at the top of the bellows assembly. This
set of seals allows translation and rotation of the LHe volume and
the target attachment screw thread at its lower end. A pump-out
port is located between the pair of sliding seals and creates an
evacuated region that is separated from the main vacuum space.
Fig. 4. An exploded-view and photo of the low-temperature Wobbler connection.
A copper disk with left-handed threads pivots against a copper disk that is
connected to the bottom end of the LHe volume with an indium joint. The two
disks are thermally coupled by a set of four copper braids (visible in the photo) that
are electron-beam welded to each disk.
C.D. Bass et al. / Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
The vacuum space can be evacuated through a gate valve located
at the bottom of the TC using an onboard turbo-molecular pump
(Agilent/Varian M81). The vacuum space is connected to the interior
of the strong-back tube, which adds 16 l of expansion volume.
A relief valve prevents over-pressurization in the event that the
target is vaporized due to loss of cooling.
3.2. TC connection to the target – the Wobbler
Engaging the inner LH threads of a target cell with the
matching threads on the end of the LHe volume is accomplished
by rotating a section of the vacuum-jacketed LHe tube that is
exposed above the TC bellows assembly. A flexible connection or
Wobbler is mounted on the bottom of the LHe volume. The
Wobbler (Fig. 4) is made from two copper disks whose relative
angular alignment is maintained by a pair of axial screws.
A co-axial set-screw provides a 1 mm stand-off between the disks
and forms a central pivot point. Springs mounted on each of the
axial screws provide tension between the disks. The lower disk has
a set of left-handed M26 1 threads that match the internal
threads of a target cell. The two disks are thermally connected
by four copper braids that have been electron-beam welded to
both disks. The Wobbler assembly is bolted to a copper termination on the end of the LHe volume, and a 50 μm thick indium foil
ensures good thermal contact.
The Wobbler compensates for slight misalignments between
the TC and a cryostat during target transfer operations and
prevents cross-threading. The pivoting action allows the TC to
“find” the correct starting point within a target thread. In addition,
the target threads are also machined with a blunt start to help
define the proper engagement. During the tightening of the
Wobbler connection to the LHe volume, the joint is nominally
heated to approximately 140 1C, slightly below the indium melting
point of 156 1C, to ensure that the indium can flow and fill the gap
between surfaces. Because the critical field of indium [11] is well
below the fields in the TC, the indium does not become superconducting, which would generate a thermal break.
3.3. 77 K-Shield
Threaded into the base of the LN2 volume is a cylindrical rareearth permanent dipole magnet, which is discussed in Section 3.5.
Connected to the bottom of the magnet cylinder is a circular springloaded shutter fashioned from 12 overlapping leaves, as shown in
Fig. 5. The shutter leaves can be rotated inward against the spring
action; when pressed against a thin ring (a shutter-opener) during
target transfer operations, the shutter is mechanically opened.
Shutter-openers are installed within each cryostat at locations that
are 40 K or colder. This design is similar to an 11-leaf assembly
described in [7]. When closed, the shutter functions as an extension
of the 77 K-shield and prevents 300 K thermal radiation from being
incident on a target held within the TC.
Operational experience has revealed that about 1/3 of the time,
one or more of the spring-loaded leaves fail to completely close
when the cooled shutter is lifted off a shutter-opener ring within a
cryostat, which could expose the target to 300 K radiation. An
auxiliary LN2 baffle has been incorporated above the TC gate valve
(Fig. 6) in order to minimize the impact of shutter closure failures.
This auxiliary LN2 baffle is formed from a cylindrical brass tube
ending in a sliding copper plate which are both conductively
cooled by contact with an external LN2 volume mounted on the
side of the TC. Once the TC has been completely retracted (see
Fig. 3), the auxiliary LN2 baffle plate is pushed into position,
sealing off the bottom of the LN2-cooled brass tube and effectively
capturing the shutter within a LN2-cooled cup. In this way, the
exposure to 300 K radiation from any gaps between shutter leaves
is limited in duration to about 30 s, which is the time needed to
raise the shutter off from its cold “opener”-ring and fully retract it
behind the auxiliary LN2 baffle.
3.4. Vacuum lock connection
A full view of the lower section of the TC is shown in Fig. 7,
with a cut-away model on the left and a photo of the assembled
unit on the right. The bottom ISO-63 flange attaches to a matching
flange on a receiver cryostat. During the cycle of an HD target, the
polarization relaxation times are the shortest during the periods
spend in the low-field and comparatively high temperatures of the
TC. Minimizing the elapsed time in the TC reduces the potential
polarization loss during target transfers. For this reason it is important
Fig. 5. Spring-loaded inward-opening 77 K radiation shutter; side view (top-left)
and an end view, looking in from the back (top-right). The shutter has eleven
identical overlapping leaves and a longer twelfth one that covers the central gap. A
photo of the shutter viewed from the back (the target side) in its closed position is
shown in the lower-left panel. The lower right panel shows the shutter with its 12
spring-loaded leaves forced partially inward when pressed against a shutteropener.
Fig. 6. Full 77 K shield. In the retracted position, the target is centered in the 77 K
dipole magnet. The leaf shutter of Fig. 5 is threaded onto the dipole, which is
directly connected to the TC tube containing LN2. The shutter is captured in an
auxiliary cup, which is conduction cooled from a separate LN2 volume.
C.D. Bass et al. / Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
Fig. 7. Cutaway (left) and photo (right) of the engagement end of the TC. The solid HD target is connected to the LHe tube through the “Wobbler” and is centered within the
cylindrical dipole magnet, which is directly coupled to the LN2 tube and serves as a 77 K thermal shield. The forward end of the magnet is capped by a leaf-spring shutter and
an auxiliary LN2 cup, which complete the 77 K shield. During a transfer, the short spool-piece after the gate valve attaches to a matching valve on another cryostat. An onboard turbo-molecular pump rapidly establishes a vacuum within this interconnection volume. A vacuum gauge located on the inlet of the turbopump (shown to the right
and behind the turbopump) provides a monitor of the pressure from atmosphere into the 10 7 mbar range.
to minimize the time devoted to simply re-establishing a good
vacuum in the connection between the TC and its mating cryostat.
For rapid pump down, an on-board turbo-molecular pump (Agilent/
Varian 81-M) is directly mounted on the short spool piece connecting
the TC gate valve and the matching gate valve on a receiver dewar.
This brings the interlock pressure into the 10 6 mbar range within a
few minutes, as measured by a cold-cathode gage mounted directly on
the spool piece extension.
3.5. Halbach magnet
For frozen-spin HD, magnetic fields greater than about 100
Gauss are needed to prevent rapid HD polarization loss. Within
the TC and attached to the bottom of the LN2 heat shield is a
cylindrical permanent magnet array, arranged in one of the
Halbach configurations designed to produce an internal dipole
field [12] of approximately 0.1 T. The array consists of 16 sectors of
neodymium magnets that are mounted in a copper shell. The
magnetization direction for each of the neodymium alloy sectors is
shown schematically in Fig. 8. Relative to the top sector, each other
sector located at an azimuthal angle ϕ has a field direction that is
rotated by 2ϕ. This produces a net dipole field in the interior of the
cylinder that is transverse to the axis.
The magnetic material is an alloy of neodymium, iron, and boron
(NdFeB), obtained from Tridelta Magnetsysteme [13]. A roomtemperature solenoid, top panel of Fig. 9, has been used to magnetize
the alloy. Four bars of the NdFeB alloy are held in the desired
orientation transverse to a plastic sleeve, shown here partly inserted
into the bore of the air-core solenoid. A current pulse from the
Fig. 8. Cross-section of the TC magnet made from 16 sectors of neodymium
magnets. The directions of magnetization in each sector are indicated by the
arrows. This Halbach configuration produces a net transverse dipole field within
the array.
discharge of a capacitor [14] generates the field pulse shown
in the middle panel of Fig. 9. The 3.3 T peak field exceeds the
saturation level of the material and results in a permanent
C.D. Bass et al. / Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
Fig. 10. Cutaway schematic of the Halbach dipole and its internal room-temperature fields. In the retracted position, a solid HD target would be centered at about
z¼ 80 mm.
10 mm from the central axis varies by about 10% due to variations
in magnetization of the neodymium sectors.
The fields produced by neodymium magnets are temperature
dependent and have a maximum field strength at about 135 K,
where the alloy undergoes a spin-reorientation [15]. This phase
transition is reversible as the magnet is thermally cycled.
Measurements of the magnets used in this Halbach dipole are
summarized in Fig. 11. At temperatures below the phase transition,
the field strength decreases monotonically. At 4 K, the field
strength of the magnets is approximately 90% of the value at
300 K; at 77 K (the nominal temperature of the Halbach dipole in
Fig. 9. The solenoid used to magnetize sections of neodymium alloy (top panel),
shown with a fixture that holds up to four sections in the desired orientation.
A typical magnetizing pulse is plotted in the center panel. The magnetized sections
are captured within a copper sleeve (bottom panel) to create a Halbach dipole with
a field transverse to the axis of the assembly – see text.
magnetization. The neodymium alloy bars are machined with a
trapezoidal cross-section and, once magnetized, are captured within
channels that have been wire-cut from a thick-walled copper tube, as
shown in the lower panel of Fig. 9. Each of the 16 channels contain
three bars end to end that run the length of the magnet assembly.
The room-temperature fields generated by the Halbach array
are plotted in Fig. 10. On axis, the field is flat over the central
80 mm length at about 1.1 kG. When the TC is fully retracted,
targets are centered near the z ¼80 mm position. HD target
diameters used at BNL and JLab have varied in size with radii
between 7.5 mm and 12.5 mm. The field at an average radius of
Fig. 11. The temperature dependence of the field strength for neodymium
magnets. Different magnets were made from samples of neodymium alloy as
measured with different Hall-probes. Black circles were measured with a probe in a
flowing bath of room temperature gas. Data shown as squares and diamonds were
measured with a temperature-compensated cryogenic Hall probe.
C.D. Bass et al. / Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
the TC during transfer operations), the field strength is approximately 95% of that at 300 K.
The TC extracts polarized HD targets from cryostats with
longitudinal magnetic fields. Because the motion of the target cell
during extraction is negligible compared to the Larmor frequencies
for H and D, the spins adiabatically rotate as they enter the TC
Halbach transverse dipole. Thus, the Halbach dipole brings two
important characteristics to the TC: it provides an adequate
holding field in a limited radial space, and its interior transverse
orientation avoids zeros in the magnetic field that could occur
between two solenoids with unintentionally opposed directions.
3.6. Counterbalanced lifting arm
During transfer operations a retractable arm is used to position
the TC off-center from a hoist to eliminate mechanical interference
from crane components (see Fig. 12). The lifting point slides on
rails and is positioned using a motor and lead screw. Lead
counterweights are fixed on the free end of the lifting arm. When
the TC is supported by a hoist and the lifting arm is extended,
adjustments in the position of the lifting point on the arm produce
vertical movements of the TC. The large mechanical advantage
provides a delicate control when docking the TC on the top of
another cryostat.
For service operations, the lower end of the TC is connected
to the free end of the lifting arm by an approximately 1 m long
stainless-steel cable. With the cable in place, movement of the
lifting point along the arm causes the TC to tilt. In this way the TC
can be maneuvered to a 301 position for service, cooling, or storage
on a cart. Movement of the lifting point to its farthest position on
the arm allows a horizontal orientation of the TC, which permits a
rapid removal of cryogens, if required.
3.7. Instrumentation
The cryogen levels in the volumes of the TC are continuously
monitored during operations. A capacitance level meter [16]
measures the LN2 level, and a superconducting wire [16] measures
the LHe level. The temperature of the LHe tube is monitored with a
Cernox resistor [17]. During transfer operations, the partial pressures of mass-3 (HD), mass-4 (4He), mass-28 (N2) and mass-32
(O2) are continuously monitored with a residual gas analyzer
(Stanford Research Systems RGA-100 [18]) that samples the outlet
of the turbopump that is pumping the space containing the target
(Fig. 7).
4. Operation and performance
4.1. Preparations and cooling
After each thermal cycle of the TC, the Wobbler connection
to the LHe tube is tightened to 3.4 N m at 140 1C, and then again
at room temperature to ensure good contact across the indium
interface as discussed in Section 3.2. Then, with the spool piece at
the bottom of the TC blanked off and the vacuum-gate valve open
(Fig. 7), the central space is evacuated using the on-board
turbopump to less than 5 10 7 mbar. The cable and winch
mechanism indicated in Fig. 3 holds the moveable volumes in
place against the vacuum force.
The LN2 and LHe volumes are next purged of condensed water
from previous use by repeated cycles of pumping to 0.05 mbar and
backfilling with 1 atm of helium gas. TC cooling takes place with
the TC supported on a cart at an angle of 30○ from horizontal. The
liquid nitrogen volume is first filled with LN2. The LHe volume is
next pre-cooled with cold helium gas and then filled with LHe. The
average time to cool to 4 K is approximately 45 min.
Once the TC has reached a stable temperature of 4 K, the auxiliary
LN2 volume is filled. Just prior to the start of a transfer operation, a
roughing pump is connected to the back of the central LHe tube and
the TC temperature is lowered to 2 K, which takes about 20 min to
stabilize. The TC can then be lifted from its cart with the fixture
discussed in Section 3.6. Once in a vertical position, the TC is ready to
be positioned over a cryostat for a target transfer operation.
4.2. Cryogenic hold times
The internal TC LN2 volume has an average hold time of 1.3 h and
is typically topped off once during a transfer. The auxiliary LN2
volume will remain cold for over 6 h. With the LN2 volumes cold, the
LHe volume has a hold time of about 18 h at 4 K. When the central
LHe volume is pumped to lower the temperature to 2 K there is an
initial 40% reduction in the LHe level. From that point on, the 2 K LHe
hold time is about 10 h. These times are sufficient for the transfer of a
solid polarized target between two cryostats, which takes about 2 h.
4.3. TC insertion into a cryostat
Fig. 12. Photo of the counter-balanced lifting arm used to position the Transfer
Cryostat off-center from a hoist. The adjustable lifting point with its counter weight
provides the delicate control needed to connect the TC to another cryostat.
Once the spool piece under the TC gate valve (Fig. 7) is mated to a
matching vacuum gate valve on another cryostat, the space between
the two gate valves is evacuated, by first roughing through the onboard turbopump and then to high vacuum with the turbopump
spinning. The gas exiting the turbopump is monitored with an RGA
(Section 3.7). Three stabilizing chains anchored to fixed points near
C.D. Bass et al. / Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
the cryostat are attached to the top section of the TC and their
tensions adjusted to level the TC so that its axis is held rigidly parallel
to that of the cryostat. Aligning and fixing the TC position is essential
in order to constrain the motion of the TC LHe tube when extended,
as in Fig. 3. (A poor alignment of the TC with the mating cryostat can
lead to excessive friction when rotating the target threads. This can
generate thermal spikes, which can raise the HD into the temperature region where the polarization can decay more rapidly due to a
shortened T1 or even material can be lost from the target cell due to
high vapor pressure. This is partly mitigated by operating at 2 K, as
discussed in the next section.)
Once the pressure between the gate valves has been reduced
below 5 10 6 mbar (within a few minutes, or in practice by the
time the stabilizing chains have been adjusted) the lower gate
valve of the connecting cryostat is first opened and the residual
gas is sampled to verify a similar or lower pressure, particularly
for He. (Variable temperature cryostats such as the PD, discussed
in Section 2.1, are typically operated with a few mbar of He that
provides a thermal exchange between the HD and a separate LHe
reservoir. This exchange gas must be thoroughly pumped away
before beginning transfer operations, since any residual He entering the TC during the transfer process could thermally short its
LHe tube and result in a subsequent loss of cryogens. Depending
upon conditions, this preparatory exchange gas pumping can take
many hours since the vapor pressure of cold He is quite low.) The
TC gate valve is then opened and the shutter plate of the auxiliary
LN2 shield is retracted. The TC bellows assembly is then allowed to
compress under the vacuum force by unspooling the support cable
using the winch shown in Fig. 3. As it does so, the LN2 and LHe
volumes extend into the cryostat. This continues until the shutter
(Fig. 5) seats in the shutter-opener ring of the cryostat, which
mechanically opens the 12 spring-loaded leaves (Fig. 5). The
shutter-opener now supports the TC LN2 volume and no further
compression of the TC bellows is possible. Each HDice cryostat has
a shutter-opener ring anchored to a point with a temperature of
40 K or lower, and coils that guarantee a continuous coverage of
magnetic field below the TC Halbach dipole.
Further extension into the connecting cryostat is accomplished
by sliding the LHe tube down through the double O-ring seal.
During this procedure, the space between the two O-rings is
evacuated and the pressure at the inlet to the roughing pump is
continuously monitored. A counter-weight is attached to the cable
above the LHe tube to offset the vacuum force and provide better
control. The LHe tube is lowered until the Wobbler (Fig. 4) contacts
the target. The LHe tube is then rotated clockwise (against the LH
threads) until the starting thread on the Wobbler falls into the
blunt start of the LH target thread. Subsequent counter-clockwise
rotation of the LHe tube engages the Wobbler threads in the inner
threads of the target ring (Fig. 1).
4.4. Thermal spikes from thread engagement at 2 K
The temperature of the connecting cryostat is continuously
monitored at a point close to the target threads. Turning the
threads of the Wobbler against the copper target ring generates
heat. It is important that this motion be smooth and moderately
slow to allow the heat to dissipate. When the TC is used at 4 K
(without LHe pumping), it is difficult to avoid spikes in target
temperature that reach into the 6–7 K region where the vapor
pressure of HD becomes appreciable (10 3 mbar at 7 K). In
contrast, when the TC LHe volume is pumped to reach 2 K, the
added cooling from the TC keeps the thermal spikes generated by
inter-thread friction below 4 K where the HD vapor pressure is
negligible (5 10 9 mbar). With repeated counter-clockwise rotation of the TC LHe tube, the left-hand (LH) Wobbler threads
bottom in the inner LH threads of a target ring (Fig. 1). Additional
torque is then needed to break the right-hand (RH) target threads
free from the cold pedestal of the connecting cryostat, which
results in another temperature spike. As counter-clockwise rotation of the TC LHe tube continues, the target threads out of the
cryostat pedestal. The associated temperature spikes tend to be
larger at this stage because the larger radius RH target threads
have a greater contact surface area. Once the target is completely
disengaged from the cryostat pedestal, its temperature is then
entirely determined by the 2 K TC LHe tube. It is then pulled up,
first into the TC Halbach dipole by raising the LHe tube with the TC
winch system. (In this process, the outside of the TC LHe tube that
has been cooled by exposure to the environment of the connecting
cryostat has to be pulled through the TC double O-ring sliding seal,
indicated at the top of the bellows assembly in Fig. 3. To prevent
the O-rings from freezing, heated air is used to warm the exterior
region of the sliding seal prior to and during extraction of the TC
LHe tube. The rate of extraction must be adjusted by monitoring
the pressure in the line pumping the space between the O-rings
of the sliding seal.) Continued use of the winch raises the LN2
volume, allowing to spring close the 12-leaf shutter and bringing
the target completely up into the TC, after which the auxiliary LN2
shutter plate is closed to complete the 77 K thermal shield around
the target. Finally, the vacuum gate valves on the TC and the
cryostat are closed.
At this point the TC with its polarized target can be moved to
another cryostat. Inserting the target into another receiver cryostat
follows the same sequence as outlined above, except that the TC
LHe tube is rotated clockwise to deposit the target and disengage
the TC Wobbler.
4.5. Performance with frozen-spin HD
Of the approximately 2 h transfer process, the average time
during which the temperature and magnetic field seen by an HD
target is determined by the TC is about 50 min. To study the
polarization loss due to spin-lattice relaxation (T1) while the target
is in the TC, several test lifts were performed with frozen-spin HD
targets. Typical results are shown in Fig. 13, with H- and Dpolarizations monitored by field-sweep NMR. The area of the
resulting NMR signal for H is proportional to the H-polarization
and is plotted in the upper panel, and the corresponding Dpolarization signal is shown in the lower panel. The NMR data
collection required a few minutes and in between these periods
the target was held in a cryostat at 2 T and 3 K.
After approximately 1 day (the 25.3 h point in Fig. 13), the
target was extracted from the cryostat, lifted up into the TC and
held there for 50 min before being returned to the cryostat for
further NMR measurements. The data points shown as open circles
in the figure were collected before this test lift. The curves were
fitted to the pre-lift data, parameterized as PðtÞ ¼ P o e t=T 1 and
extrapolated out into the post-lift period. While the spinrelaxation times were very long, and as a result could only be
roughly determined in a day (3277 34 h for H and 203571283 h
for D), it was clear that the polarization signals measured after the
test lift (the solid diamonds in Fig. 13) were completely consistent
with the expected dependence. The fractional polarization losses
for both H ( 1 72%, relative) and D (0 73%, relative) during the
transfer operation were insignificant.
The transfer operations with the TC are the most delicate stages
in the HD target cycle and require the monitoring of many
separate parameters. As such, depending upon the proximity of
the source and the receiver cryostats, a reliable transfer typically
requires between 5 people, when the cryostats are next to each
other as shown schematically in Fig. 2, and up to 9 people when
loading a target into a large experimental hall during which
separate operations must be performed at multiple elevations.
C.D. Bass et al. / Nuclear Instruments and Methods in Physics Research A 737 (2014) 107–116
0.1 T field generated by a cylindrical neodymium magnet array that
is assembled in a Halbach configuration and maintained at 77 K.
Targets are engaged and disengaged using a combination of lefthanded and right-handed threads, and proper thread engagement
over 2 m is ensured by a pivoting 2 K flexible connection. A springloaded 12-leaf shutter and an auxiliary shutter provide a hermetic
77 K-shield during transport of a target between cryostats. Direct
tests with frozen-spin HD show negligible loss in either H- or Dpolarization during typical 50 min transfer periods. Multiple uses of
this device indicate failure rates less than 5% per transfer. Because the
process of polarizing an HD target and bringing it to an In-Beam
Cryostat requires multiple transfers, the overall success rate for
mounting an experiment with this device exceeds 75%. This has
allowed the successful execution of experiments at Brookhaven
National Lab and at Jefferson Lab.
The authors are indebted to Richard Ruggiero for his help with
the TC design at BNL and to the staff of the Central Institute for
Engineering, Electronics and Analysis (ZEA) at Forschungszentrum
Jülich for the careful construction of much of the Transfer Cryostat.
We are grateful to Bill Clemens for his help e-beam welding critical
components at JLab. We thank Doug Tilles and the JLab Hall-B
technical crew for their dedicated assistance during the polarized
target operations that utilized the Transfer Cryostat at JLab.
This work has been supported by the United States Department
of Energy, Office of Nuclear Physics Division, under contract
DE-AC02-98-CH10886 supporting Brookhaven National Laboratory, and under contract DE-AC05-06OR23177 under which Jefferson Science Associates operates Jefferson Laboratory, the US
National Science Foundation, the German Ministries for Science
and Education supporting Forschungszentrum Jülich, and the
Istituto Nazionale di Fisica Nucleare of Italy.
Fig. 13. Time dependence of the H (top panel) and D (bottom panel) polarizations
of a frozen-spin HD target in a cryostat at 3 K and 2 T. At the 25.3 h point, the target
was extracted from the cryostat, held in the TC for 50 min and then returned to the
cryostat. The data shown as open circles represent the areas of NMR signals before
this test lift and the solid lines were fits to the expected relaxation, PðtÞ ¼ P o e t=T 1 ,
extrapolated here into the post lift period. The solid diamonds were measured after
the target was returned to the cryostat.
The potentially vulnerable steps in the transfer process have been
discussed above, along with our mitigating procedures. Nonetheless, since many of these involve a level of qualitative human
judgment, the success rate, while now quite high, is invariably less
than 100%.
5. Summary
We have described a device with moveable liquid nitrogen and
liquid helium volumes that is capable of reaching over 2 m into a
cryostat or dilution refrigerator in order to extract a polarized target
of solid hydrogen deuteride (HD). This Transfer Cryostat maintains
HD temperatures at 2 K and surrounds a target with a transverse
[1] S. Capstick, W. Roberts, Phys. Rev. D 58 (1998) 074011.
[2] R.G. Edwards, J.J. Dudek, D.G. Richards, S.J. Wallace, Phys. Rev. D 84 (2011)
[3] A.M. Sandorfi, J. Phys. Conf. 424 (2013) 012001.
[4] A.M. Sandorfi, S. Hoblit, H. Kamano, T.-S.H. Lee, J. Phys. G 38 (2011) 053001.
[5] S. Hoblit, et al., LEGS-Spin Collaboration, Phys. Rev. Lett. 102 (2009) 172002.
[6] 〈http://www.jlab.org/exp_prog/proposals/06/PR06-101.pdf〉.
[7] N. Alexander, J. Barden, Q. Fan, A. Honig, Rev. Sci. Instrum. 62 (1991) 2729.
[8] M. Rigney, et al., Nucl. Instrum. Methods A 356 (1995) 116.
[9] X. Wei, et al., Nucl. Instrum. Methods A 526 (2004) 157.
[10] C. Thorn, A. Caracappa, in: AIP Conference Proceedings, vol. 980, 2008, p. 397.
[11] R.W. Shaw, D.E. Mapother, D.C. Hopkins, Phys. Rev. 120 (1960) 88.
[12] K. Halbach, Nucl. Instrum. Methods A 169 (1980) 1.
[13] Tridelta Magnetsysteme GmbH, D-44287 Dortmund, Germany 〈http://www.
[14] Hirst Magnetic Instruments Ltd., Falmouth, Cornwall TR11-4SN, UK 〈http://
[15] K.J. Strnat, D. Li, H. Mildrum, in: Proceedings of the 8th International Workshop on Rare Earth Magnets and their Applications, Dayton, OH, 6–8 May
1985, Paper no. VIII-8.
[16] American Magnetics, 112 Flint Rd, Oak Ridge, TN 37830, USA.
[17] LakeShore Cryotronics, 575 McCorkle Blvd, Westerville, OH 43082 USA.
[18] Stanford Research Systems, 1290-D Reamwood Av, Sunnyvale, CA 94089, USA.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF