The strong demand for clean energy encourages scientists to contribute to these social efforts [1]. Since the 1980s, there have been suggestions of possible advantages coming from the use of nuclear-polarized reactants for fusion [2–4].

The well-known results on spin 1/2 (³He) – spin 1 (D) fusion reactions, shown on cross sections and differential-cross sections on collision experiments [5], open a variety of proposals and dreams on the field of nuclear fusion, but some questions require answers:

• Can we have fuel with sufficient density and polarization satisfying fusion requirements?

Magnetic fields are required to maintain the polarization of polarized nuclear substances in various activities and during transportation. Specifically, the main purpose of this contribution is to provide holding fields for polarized fuel in fusion tests [6]. Ideas and support on this topic come from studies and developments on spin-dependent studies in nuclear and subnuclear physics [7] in which nuclear-polarized targets are used.

A hollow bulk superconductor can hold polarization substances inside itself, provide the required holding fields, and shield external fields [8,9]. These features are important improvements with respect to conventional coil-based solutions [10] and can be hosted in the production system of nuclear-polarized targets. Finally, they become useful for the preparation of nuclear-polarized fuels.

Additional advantages include minimal space needed to fit in the experimental environments, maximum field compactness, absence of heat loads from current leads, and the capability to operate without stabilizers. Furthermore, because MgB₂ has low mass density and is made of low atomic number elements, the reaction products experience less energy losses in the path of the material.

MgB₂, as a superconductor material, has adequate critical current, critical field, critical temperature (39 K), and machinability [11]. Concerning the temperature, polarized hydrogen or deuterium targets require cryogenic temperatures well below the critical temperature of the MgB₂, which, therefore, can be easily matched or incorporated within their cooling systems. Finally, the whole system can be easily moved from the preparation site and transported to the experimental and test facilities.

Our starting point has been the interest in this material as a magnetic field generator for high-energy physics nuclear targets [7]. In recent years, the nuclear spin physics community has shown increasing interest in exploiting the experience and knowledge acquired in the fields of nuclear and subnuclear fundamental physics to nuclear fusion programs [1].

Direct contact with one of the inventors [12] of a technique for the MgB₂ synthesis gives us the chance to undertake feasibility studies to tailor the work to specific requirements.

In this article, we will start with our preliminary feasibility study and apparatus [13], which brought us to the present upgrading. This information is useful for deep understanding and monitoring of the behavior of bulk MgB₂ hollow cylinders that can address the choice of the proper production procedure to be implemented in fusion programs with polarized fuel and polarized nuclear targets.

The material chosen for our investigation is MgB_2, obtained by the reactive liquid infiltration (RLI) method. This technique, proposed and invented by an Italian group [12], allows the production of massive samples of different geometries [14]. It is promising with respect to other techniques: the hot isostatic pressing technique proposed by a Japanese group [15] and the Mg vapor sintering on B fibers technique proposed by an American group [16].

We have implemented our first prototype for feasibility studies using an MgB₂ cylinder, machined from materials obtained starting with boron grains with a large size variability (from ∼1 μm to ∼200 μm), which has provided promising results [13]. We are encouraged to upgrade the preliminary system for more detailed investigations of different production procedures while also implementing the field mapping inside the cylinder. In the following, we will link the description of the previous installation, giving evidence on the upgrading and the possibility of new monitoring of the behavior of various samples under study.

After upgrading and commissioning the new system, we plan to test additional cylinders produced using boron grains of different sizes. In particular, we will focus on three MgB₂ batches produced with a maximum boron grain size of 40 µm, 100 µm, and 160 μm, respectively [17].

The prototype apparatus described in a previous article [13] has been upgraded. Figure 1 shows the present configuration on which we will show the modifications applied for better operational use.

For the feasibility studies, a hollow cylinder sample of MgB₂, having a length of 86 mm and an outer diameter of 39 mm, was installed. It has a nominal thickness of 2 mm, tapering from 2 mm in the middle to 1 mm at the edges. We obtained very promising results concerning the trapped magnetic field (942 mT) and the shielded magnetic field (900 mT) with an applied external field of 980 mT at the temperature of 13 K. The trapped field lasts longer than 6 days. In the following, we will present the preliminary system, which is our starting point for the final upgraded system, and the methods used in the upgrade.

The prototype is shown in Figure 2. The subsequent modifications are shown in Figure 3. The items are evidenced in emphasized characters, which will be described for the prototype and subsequent modifications.

Cryogenic Unit: The superconducting cylinder was cooled by a coldhead (Edwards 6/30), driven by its helium compressor (Cryodrive 3.0), having a nominal temperature and a cooling power of 77 K and 30 W, respectively, on the first stage and 10 K and 6 W on the second stage. Our preliminary system has allowed us to reach a minimum temperature ∽ of 12.5 K.

For the upgrade, we have borrowed a new coldhead from our colleagues [18] (SHI-Sumitomo RDE–418D4), driven by its helium compressor (F–50H): this new cooling system allows us to work in more stable conditions and reach a still lower temperature (∽ 8 K). The nominal temperatures and cooling powers are, respectively, 50 K and 42 W on the first stage and 4.2 K and 1.8 W on the second stage. With this new coldhead, we have solved the stability problems that have limited our investigations in the feasibility studies.

In the prototype system, the sample temperature was controlled by a home-made resistive heater made by welding small cartridge heaters in cylindrical pits in a copper collar, cut into two halves that were clamped on the second stage of the coldhead (L in Figure 2), taking care to make good thermal contacts. A cryogenic linear temperature sensor (CLTS) was glued on half of the collar for temperature control and monitoring by an ITC-503S multipurpose controller (Oxford Instruments).

In the upgraded system, similar heater cartridges were inserted and welded in pits, symmetrically and peripherally drilled directly on the bottom part (d₂ in Figure 3) of the copper rod (d₁), as close as possible to the copper sample holder, named simply “sample can” (F) in Figure 2 for the prototype and (e) in Figure 3 for the upgraded system. In this way, we gained better control of the temperature of the cylinder and reduced the time required for its heating procedures.

The sample can hosts the cylinder, (G) in Figure 2 for the prototype and (f) in Figure 3 for the upgraded system. Inside the cylinder was a sensor holder (H) in (Figure 2) in the prototype. For the upgraded system, a new holder with more sensors has been designed and assembled. The new sensor holder will be described in detail in this report in connection with the results obtained for the commissioning of the present system.

External magnet: The magnetic field is provided by a VARIAN electromagnet (model V3603) with a nominal maximum current of 180 A. An available power supply (Agilent 6692A), remotely controlled via a GPIB interface, allows a 110 A feed, providing a magnetic field of 980 mT, measured in the middle of the poles by a Hall sensor (Arepoc HHP-NU).

In the upgrade, we use a new 200 A/50 V power supply (CAEN NGPS 200–50–Enhanced), controlled via an ethernet communication port. Powering the magnet at its maximum allowed current (180 A), we measured 1.2 T in the middle of the poles. The required magnetic field for fusion application is in the range of 1 T, which matches exactly with the system used to produce hyper-polarized molecules as fusion fuel from the recombination of a polarized atomic beam [6].

Vacuum system and thermal insulation: Figure 1 shows the present upgraded system.

We use the two existing cylindrical chambers from the preliminary vacuum system: the bottom one (5A) in Figure 1 is in aluminum because it is placed between the magnetic field poles (3), and the upper one is in stainless steel (5B), on which the Edwards 6/30 coldhead can be connected with a custom 0-length flange. The two chambers are connected by two DN100LF flanges.

For the upgraded system, a new stainless-steel chamber (5C in Figure 1) was required to fit the new coldhead. The design of the upgrade preserves the opportunity to return to the prototype configuration. The stainless-steel chamber (5C) is equipped on the top with a fixed DN150CF flange, on which the new coldhead is connected. A different custom 0-length flange is required for this new coldhead. On the bottom of the chamber (5C), a rotatable DN150CF flange is available for the connection to the similar one on the top of the chamber (5B). On the side, the chamber (5C) is equipped with two DN40CF flanges, facing each other, and with the other two DN63CF flanges at 90° with respect to the previous two, also facing each other. All these flanges are available for service feed-throughs, and three of them are used for the Sub-D-15 feed-throughs on the DN40CF 0-length flanges, which are also visible in Figure 1.

The stainless-steel chamber (5B) is equipped with a DN150CF flange on the top for its connection with the aforementioned chamber (5C) and on its side with a DN63CF flange for a turbo pump (Turbo-V 81-M from Agilent), backed by a scroll pump (Varian SH110) and with other two DN40CF flanges: one for a dual Penning/Pirani gauge (compact full range PKR 251 from Pfeiffer Balzers), and the other for service feed-throughs. We connected a DN40CF T-piece to have two feed-throughs available. In the upgraded configuration, the service feed-throughs were installed on the side flanges of the chamber (5C), which are visible in Figure 1. A Pirani gauge (Pfeiffer PTR81) is connected on the inlet of the scroll pump, backing the turbo pump.

In the upgraded system, we exchanged the aforementioned turbo pump with a 70 L s⁻¹ (TV70LP Macro–Torr from Varian/Agilent), and we added a similar one in series to compensate for the possible degradation of the scroll pump or for periodically required maintenance (we exchange the scroll with an equivalent system from Edwards or Agilent).

The MgB₂ cylinder in the prototype system was mounted in the sample can, (F) in Figure 2. The sample can was fixed on the bottom of the copper rod, (E) in Figure 2 for the prototype and as (d₁) and (d₂) in Figure 3 for the upgraded system. The copper rod from the top was connected directly to the second stage of the coldhead (Edwards 6/30) for the prototype. For the new system, we have a proper copper transition (c) to connect the copper rod to the second stage of the new coldhead.

The thermal shield surrounding the copper rod and the sample can has a 3-mm-thick wall and is in the shape of a cylindrical copper tube tapered at the top, with a conical geometry, to a higher diameter to fit the connection on the first stage, (a) in Figure 3. To have access to the sample can after removing the aforementioned parts, we remove only the bottom part (h) of the thermal shield. The next steps to have access to the cylinder for its exchange are as follows: remove the RhFe temperature sensor, which is glued on a copper slice, screwed on the bottom cover of the sample can with two M5 brass bolts, and then remove the copper cover by simply unscrewing the remaining four M5 brass bolts. Thin indium strips were used in all thermal joints and surfaces that must be in good thermal contact.

The sample can was modified to allow the exchange of cylinders from the bottom by simply removing the coldhead with a crane kept connected to the chamber (5C), as is shown in the right side of Figure 3. All the service feed-throughs, all the vacuum wires, all the sensors, and all shielding, (g) and (h) in Figure 3, are also kept connected on the vacuum side. The connectors from the air side can be easily removed, as can the coldhead air-side connections, helium pipes, and control cable. To have access to the sample can, only the bottom shield, (h) in Figure 3, must be removed. Details on how to exchange the cylinder are shown in Figure 4: the top drawing (A) shows a transparent view of the sample can inside the cylinder (also in transparency), and inside, the sensors’ holder is visible. After removing the RhFe sensor, fixed with one of the six M5 bolts of the bottom cover (γ) in the middle drawing (B), the aluminum locking ring (δ) is accessible. It has four pins that fit in the four drilled holes on the copper bottom cover of the sample can. The locking ring can be pulled out, and the cylinder, (ϵ), fixed with epoxy glue on the locking ring, slides out from the sample can and the sensors’ holder, as shown in Figure 4C. Both these sample cans and the sensors’ holders will remain fixed to the copper rod (d₂) in Figure 3, which hosts the heater sticks. A new cylinder already glued onto another aluminum locking ring can be installed inside. The exchange can be performed without removing the delicate cabling and wiring of the sensors. The ribbon wires cable (α) goes between the copper rod and the thermal shield, coming out from one of the ports (i) of the top shield (g) in (Figure 3), and is connected to service ports of the vacuum chamber (5C). The locking ring just removed can be unglued and used for another cylinder. The upgrade allows us to slide cylinders with an inner diameter greater than 30.0 mm and an outer diameter less than 38.7 mm into the sample can, with the sensors’ holder fixed inside it. This, in turn, allows us to characterize cylinders with wall thickness from 2.3 mm, considering 0.1 mm of tolerance in both diameters, to a thinner one that can run sufficient supercurrents.

The aluminum vacuum chamber (5A) in Figure 1 has an outer diameter of 70 mm and a wall thickness of 3 mm. Two sets of strips, made with three layers of Myoflex (Oxford Instruments), were wound around the 62-mm-diameter thermal shield for good thermal insulation from the aluminum chamber. A set of strips or spacers, about 2 cm wide, are used to ensure thermal insulation between the sample can, (e) in Figure 3, the copper rod (d₁/d₂), and the copper transition (c) and the top (g) and bottom (h), forming the copper thermal shield. In various parts of the system, we have used Kapton adhesive tapes and Kapton foils to fix wires and protect the wires during the insertion of the whole system shown in the right picture of Figure 3 inside the vacuum chambers (5B) and (5C) in Figure 1.

In case of pumping after 1 day from when the system is exposed to the air for servicing, installation, or exchange of a cylinder, the pressure readout of the dual Penning/Pirani gauge (8A) gives values below 10⁻⁶ mbar at room temperature and 10⁻⁸ mbar, when the coldhead reaches its minimum temperature. The pressure values are comparable for both the prototype and the upgraded system.

Temperature measurements: The temperature of the cylinder is monitored by a calibrated rhodium–iron (RhFe) sensor (Oxford Instruments), glued on a thin copper slice fixed on the bottom of the sample can with two M5 brass bolts, which is fixed with four similar bolts the cover of the sample can.

For the upgrade, another calibrated temperature sensor from Cernox (Oxford Instruments) is installed inside the cylinder on the holder of the sensors. We observed heating power released on the holder, induced by the powering of the Hall sensors. Therefore, we changed its material from Teflon, (H) in Figure 2 for the prototype, to aluminum to have better thermal conductivity between the holder of the sensors and the sample can. The renewed holder of sensors is shown in Figure 7A.

Control and data acquisition: LabView C routines and Bash scripts are used to pilot the Oxford ITC-503S heater controller, to control and record the power supply of the external magnet, to read and record pressure values via a multigauge controller (TPG256A from Pfeiffer–Balzers), to measure temperature sensor resistances via the well-known technique of four wires through a multimeter scanner (Keithley 199 System DMM/Scanner), and to acquire magnetic fields via a module (Arepoc USB2AD controller), capable of powering and controlling six Hall probes.

Thermal cycle: In the preliminary system, about 7.5 h were required to cool the sample can from room temperature to the minimum temperature, as shown in Figure 5A. Using the coldhead heater at a power of 65 W (the ITC-503S controller can supply a maximum 40 V on a resistance greater than 20 Ω, and we arranged the heater cartridges to exceed 20 Ω to fulfill the protection limit of the system), it takes about 1 hour to heat the sample can to 60 K, which is done after each test to ensure a complete transition before the next cooling for a new measurement: the trapped field should disappear at the MgB₂ critical temperature (39 K). One and a half hours are needed to cool to nearly 13 K, the minimum working temperature usually before ramping the magnet down or up. We wait at least 1 hour at the desired temperature.

With the new system, we have decreased the mass of the copper rod connected to the second stage by reducing its diameter from 50 mm to 25 mm (d₁) in Figure 3, from the flange required for the connection to the coldhead to the bottom part (d₂), which remains at 50 mm in diameter, to drill pits and host the cartridge heaters. As a result, we have reduced the time to cool the system from room temperature to the minimum temperature (∼8 K), recorded by the RhFe sensor, to a little more than 3 h (see Figure 7B), compared to the 7.5 h needed for the prototype (see Figure 5A). In the upgraded system, with respect to the prototype, we also have better control over the thermalization of the whole system, thanks to the Cernox temperature sensor mounted inside the cylinder on the sensors’ holder. We must wait 1 h more to reach the lower temperature (13 K, see Figure 7B) recorded inside the cylinder, and then we can start ramping the magnet for magnetization or shielding measurements. In the prototype, we have no measurements of the temperature inside the cylinder; therefore, we wait an arbitrary 5 h longer to be certain, as shown in Figure 5A, when the ramp down of the current supplied to the magnet starts. This upgrade also reduces the time for heating to exceed the critical temperature and return to the minimum temperature. It requires between 3.5 h in the prototype, without control of the temperature inside the cylinder as aforementioned, against less than 1 h in the upgraded system with the measurements of the thermalization under the sample can, which faces the outer surface of the cylinder, thanks to the RhFe sensor, and on the sensors’ holder inside the cylinder, which faces the inner surface of the cylinder, thanks to the Cernox sensor.

This better control and monitoring was gained both on overcoming the critical temperature and on reaching the minimum, or desired, temperature.

To test the behaviors and the characteristics of an MgB₂ cylinder, it must be inserted, sliding from the bottom between the sample can and the holder of the sensors, and connected at the bottom of the copper rod.

In the prototype system, only one Hall probe was installed in the center of the cylinder, and it was not possible to exchange the cylinder without disconnecting the wires and cables.

The upgraded system has six Hall probes, which allow mapping of the magnetic field inside the cylinder, radially and longitudinally. A temperature sensor (Cernox) is mounted to the holder. Each of these sensors requires four wires. We have used a braid of four very thin twisted wires (AWG36 size) with Teflon insulation; therefore, we organized the system to keep the cabling and wiring connected to the feed-through for the cylinder exchange operation.

The Cernox sensor is useful for monitoring the temperature seen from the inner surface of the cylinder and also for considering the sensitivity dependence on the temperature of each Hall probe, as we will show in the result of the upgraded system.

The MgB₂ cylinder is axially centered with aluminum rings, and the bottom of the sample can is closed with a copper cover. The RhFe temperature sensor, which provides the temperature of the sample can, is fixed on the bottom of this copper cover. The bottom part, (h) in Figure 3, is then fixed on the top part (g) of the thermal shield, taking care to insert small tapes of indium for good thermal contacts between the connection flanges. The coldhead as a whole, shown in Figure 3 for the upgraded system, can be moved by a crane and slowly inserted in the chambers (5B) and (5A), which stay on the magnet support.

In the prototype system, the exchange of the cylinder was not possible without disconnecting all the wires, which must pass through the port (i) of the bottom part (h) of the thermal shield, then remove the RhFe sensor and finally, the bottom copper shield of the sample can. In the preliminary system, only chamber (5A) remained on the magnet support, and the coldhead (Edwards 6/30) was moved together with chamber (5B), disconnecting the bellow on the outlet of the turbo 7B. Every time we needed to work on the cylinder or inside the copper shield, we had to disconnect the sensors inside and manage the wires, which must slide through the port (i). For the upgraded system, instead, the vacuum pump and the gauge remain on the magnet because they are connected on the chamber (5B).

We used two different procedures to cool the samples to T _F (the final lowest temperature, or the one set fixing proper heating of the coldhead heater), below T _C , the critical temperature, and to evaluate both the trapping and the shielding capabilities of the MgB₂ cylinder.

The first procedure consists of cooling the cylinder in the presence of an applied magnetic field, B _C . After reaching T _F , the field is slowly reduced to 0. This procedure triggers the appearance of supercurrents, trying to develop a magnetic field comparable to B _C . This procedure, known as field cooling (FC), in view of our applications and interests, was named simply magnetization.

The second procedure consists of cooling the cylinder with B _C equal to 0. After reaching T _F , the field is slowly increased to a maximum value. In this case, the supercurrents will try to generate a magnetic field, as opposed to the external one, thus shielding its presence. This procedure, known as zero-field-cooling (ZFC), was called simply shielding.

We had evidence of both behaviors already in the preliminary studies, which we recall here as a starting point to have clearer evidence and a better interpretation of the observed behavior in the upgraded system with more sensors installed inside the cylinder.

We report the results from the commissioning of the upgraded system, satisfying the following requirements: working in a 1 T magnetic field range of the recombination of an atomic polarized beam in hyper-polarized molecules [6], controlling the temperature to find the best working conditions, and being stable for 6 or more days. The autogenerated field is sufficient to keep the polarization; if it is prepared in the recombination system at 1 T, it maintains a field higher than 300/400 mT. The mapping of the field inside the cylinder allows us to choose where to properly locate the condensing surface for the polarized substances or, in the case of its use of a polarized nuclear target, how to scale the target dimension or the cylinder geometry to fulfill the requirements still under design [7].

One of the first proofs is the dependence on temperature of the sensitivity of the Hall probes, as shown in Figure 7B.

The sensors we have installed are calibrated at a temperature of 297 K. The dependency of Hall probes on the temperature is well known, as reported in the literature [19] for the same types of installations. We must correct the raw data acquired by the Arepoc controller: a sixth-order polynomial fit is applied to the corresponding data acquired using the calibration at 297 K in the function of the temperature T _Cernox . We then must be sure that the cooling starts from a temperature higher than 297 K because around this temperature, the variation is bigger (see Figure 7B). As a result, the installation of the Cernox sensor on the holder of the sensors has been mandatory for the temperature corrections and is not only useful for monitoring the temperature difference between the outer and inner surfaces of the cylinder.

The improvement in the cooling time is evident from the comparison of Figure 5A, in which ∼ 7.5 h are required to reach the minimum temperature of the RhFe readout, starting from a temperature higher than 297 K, instead of a little bit more than 3 hours in Figure 8A for the new system. In both cases, the RhFe sensor made good thermal contact with the bottom of the sample can.

One more advantage is that in the preliminary system, we were not able to monitor when the inner surface neared the minimum temperature. In the upgraded system, this is monitored by the Cernox sensor, and it takes a little more than 4 hours, 1 hour more than the time required to reach the minimum temperature on the bottom of the sample can. In the preliminary system, we had to wait an arbitrary time because we were not sure about the thermalization of the inner part inside the cylinder. Figure 5A shows that we waited more than 5 h after the RhFe sensor read the minimum temperature to be sure of the cooling of the whole MgB₂ cylinder, and only then did we start to ramp down the current supplied to the magnet for the FC procedure or ramp for the ZFC procedure.

To move to another measurement, we must heat the cylinder and overcome the critical temperature, observing the transition of it from the superconducting state to the normal one. The sample can warm earlier than the sensors’ holder, and in the magnetization case, we observed that the trapped field is completely lost only when the Cernox temperature is higher than 39 K. The parts of the cylinder that are still below the critical temperature can run supercurrents, but they are reduced as the temperature increases.

In summary, we report the main achievements of the upgraded system in comparison with the feasibility study system in Table 1, thanks to the experience gained with it.

For the commissioning of the system, we have inserted an MgB₂ cylinder prepared with boron grains having a maximum size of 40 μm, P40, with a length of 97.0 mm and an inner and outer diameter, respectively, of 32.5 mm and 38.5 mm in the sample can. The P40 is more unstable but can support higher super currents (Giunchi, 2016); therefore, it is a good test to monitor the sensitivity of our system to instabilities that can be overcome at higher temperatures. In Figure 8A, the values, calibrated and corrected with their respective temperature sensitivities, of the four sensors that monitor the transverse magnetic fields inside the cylinder, are reported for the magnetization procedure. The system is sensitive to flux jumps: three flux jumps are recorded at the lower temperature of the FC procedure (8.5 K on the bottom of the sample can), which are more evident on the Hall probes S _{4 ect} and S _{6 ert} located on the edge of the cylinder. The flux jumps are also recorded by the middle probes, S _{2 mct} and S _{3 mrt}, where the transverse field losses are much less.

As shown in Figures 8A, B the probes measure higher fields close to the wall of the cylinder, radial probes S _{3 mrt} in the middle and S _{6 ert} on the edge, with respect to the probes in the center of the symmetry axis, S _{2 mct} in the middle and S _{4 ect} on the edge. In addition, higher magnetic fields are observed in the middle than on the edge of the cylinder, along the central position S _{2 mct}, at the middle vs. S _{4 ect} at the edge, and along the radial position S _{3 mrt} at middle vs. S _{6 ert} at the edge.

The system allows us to monitor the behavior of the fields with respect to the temperature. We could control the flux jumps by increasing the temperature. For the sample P40 under study, we could follow the number of flux jumps from three to two, to one and finally to zero. In Figure 8B, at a temperature recorded by the RhFe on the bottom of the sample can of 12.5 K, the case without flux jumps is evident, in which the trapped magnetic field was higher than in the case of three flux jumps at the lowest temperature of Figure 8A.

The commissioning of the upgraded system fulfills the requirements for investigating the behavior of a transverse magnetic field along the longitudinal direction, which is the symmetry axis of the geometry under study, of a hollow cylinder, and also radially, giving the possibility to provide the mapping and also the homogeneity of the field for use as a holding and transportation field for nuclear-polarized substances.

Table 2 reports the quantities that can be recorded, allowing the characterization of the MgB₂, showing the capabilities of the system in investigating the behavior of the superconductor bulk material at a set temperature, acquiring the distribution of the magnetic field, and recording its values. The spatial variation of the field can be provided along the radial coordinate in the middle, for example, (ΔB/Δr) _m = ( S 3   mrt - S 2   mct )/(11 mm), where 11 mm is the distance between the Hall probe along the radius coordinate, and (ΔB/Δl) _c = (S _{4 ect}- S 2   mct )/(48 mm), where 48 mm is the distance between the Hall probe along the symmetry axis of the cylinder. These commissioning results are already useful for the fusion fuel application in the recombination system. The present P40 cylinder can be prepared in the 1 T magnetic field of the recombination system of polarized atoms in hyper-polarized molecules[6] because it maintains a sufficient holding field on the order of hundreds of mT. The mapping allows us to define where to locate the condensing surface for the polarized substance inside the cylinder. To maintain hyper-polarized molecules, it is conservative to overcome 300 mT, which can be reached by applying 1 T, which is the field used in this recombination apparatus.

Thanks to the experiences gained in preliminary feasibility studies on a hollow bulk superconducting MgB₂ cylinder, an upgraded system that allows testing different production procedures has been redesigned and commissioned to allow easy cylinder exchange and mapping of the magnetic field inside them.

This dedicated apparatus allows us to investigate the possible exploitation of bulk MgB₂ in the production and transport of polarized fuel in future nuclear fusion facilities.

After a series of modifications and improvements from the initial design, the apparatus has been successfully commissioned and is ready for proper characterization of high-temperature superconducting materials in a cylindrical form, which fits the geometry of an inner diameter greater than 30.0 mm, an outer diameter less than 38.7 mm, and a maximum length of 116 mm.

The requirements for nuclear fusion fuels have already been fulfilled in the commissioning work with a P40 cylinder, which is not, according to the literature, the best-performing one, at least for its stability. However, it has shown good behavior in the field of 1 T, perfectly tuned with the polarized atomic beam recombination apparatus [6] and provided autofields sufficient to maintain nuclear polarization in the molecules. In the future, different samples with different grain sizes will be systematically characterized, with different thicknesses and different geometry, tapered cylinders with a thicker wall at the edges, for instance, to find the best solution, which means higher fields, higher stability, good homogeneity, and less bulk material, which means thinner cylinder walls.