A Tabletop Persistent-Mode, Liquid Helium-Free 1.5-T MgB2 “Finger” MRI Magnet: Construction and Operation of a Prototype Magnet

Abstract

This paper presents results of construction and operation of a persistent-mode, liquid-helium-free, small-scale prototype magnet for the development of a tabletop 1.5-T “finger” MRI system for osteoporosis screening. The prototype magnet, composed of 2 MgB₂ coils, one superconducting joint, and a persistent-current switch (PCS) built from a portion of one coil, was wound with a one continuous ~80-m long unreacted and monofilament MgB₂ wire and then reacted. The test magnet was charged successfully and generated the estimated target field of 1.75 T at 5 K with the proposed PCS operation. During initial persistent-mode, the field was slightly decayed due to the index dissipation of the joint; thereafter it sustained the persistent field of 1.7 T for 35 h. The test results validated the joint resistance of < 1.2 × 10⁻¹¹ as well as the proposed approach involving the PCS coil circuit model.

I. Introduction

AN MgB₂ wire has been one of the most viable options to replace low temperature superconductors (LTSs) used in commercial MRI applications due to its higher critical temperature and greater thermal stability than those of LTS counterparts, as well as availability for persistent-mode operation of the magnet [1]–[5]. In addition, the MgB₂ MRI magnet proliquid helium (LHe)-free features. Therefore, we can expect vides the improved immunity to the premature quench and the that the MgB₂ will facilitate user access to MRI services at more affordable price by reducing cooling and operating expenses.

Commercial MRI magnets are operated in persistent-mode that inevitably requires a splicing technique such as ideally superconducting or very low resistive (typically, ≤ 10⁻¹¹ Ω). Since 2000, feasibility studies of MgB₂ wire, manufactured by Hyper Tech Research, for the use of persistent-mode MRI magnet had been conducted at the Magnet Technology Division of the MIT Francis Bitter Magnet Laboratory [6]–[13]. For further challenges of future MgB₂ MRI system, we recently proposed the development of a spatial high-resolution tabletop 1.5-T MgB₂ “finger” MRI system in particular for osteoporosis screening [14]. In our 4-year project, we may further develop first-ever features as well as verify our previous techniques.

Here, we present the results of design, construction, and testing of a persistent-mode, liquid-helium-free, small-scale prototype MgB₂ magnet. The prototype magnet, composed of 2 coils, one superconducting joint, and a persistent-current switch (PCS) built from a portion of one coil, was wound with one continuous unreacted/monofilament MgB₂ wire and then reacted, i.e., wind-and-react method. Prior to with solid nitrogen, we have operated the prototype magnet, cooled by conduction cooling system, in persistent mode in the temperature range of 5–10 K. Based on the test results, we have also demonstrated a novel concept for a persistent-mode magnet in which one coil not only contributes to the field but also acts as a PCS.

II. Design and Construction

A. Conductor

An MgB₂ wire used for this study was manufactured by Hitachi, Ltd.. The bare wire consists of a 0.37 mm diameter MgB₂ core, a 75 μm barrier layer of iron, and a 60 μm layer of copper, which corresponds to the contents of 34, 32, and 34 %, respectively. Provided critical current (I_c) value of the wire is ≥ 200 A at 15 K and 2 T, a typical operating condition of our proposed MgB₂ coil. Note that, φ0.37 mm MgB₂ core in such condition rarely incurs flux jumping phenomenon [7].

B. Electromagnetic Design

Prior to our final goal to develop a tabletop 1.5-T MgB₂ “finger” MRI magnet, the performance of a superconducting joint, persistent current switch (PCS), and small-scale prototype magnet should be sequentially verified. Based on our previous test results (using MgB₂ wires manufactured by Hyper Tech Research), the wind-and-react method was adopted for this initial step.

Our 1st small-scale prototype magnet composed of 2 MgB₂ coils, was designed to be wound with one continuous MgB₂ wire; table I lists its key parameters. Both coil1 and coil2 produce the target field of 1.8 T at axial center of coil2, and a peak field (B_max) of ~ 2 T on inner wall of coil2. The I_op/I_c (B_max) of 0.54 at 15 K has a sufficient operational margin during persistent-mode [14]. As show in Fig. 1, a superconducting joint is located at the fringe area with the field of < ~ 0.2 T which is about an order of magnitude lower than B_max.

TABLE I.

Key parameters of the prototype magnet

Parameter		Specification
		coil1 (&PCS)	coil2
Wire diameter (with S-glass)	[mm]	0.79
Content of MgB2; Fe; Cu	[%]	34; 32; 34
r1; r2	[mm]	12.80; 16.47	12.80; 25.96
z1; z2	[mm]	−43.22; −29.94	−14.94; 14.94
# of turns per layer		16	36
# of layers		5	18
Total wire length	[m]	86
Operating current (Iop)	[A]	~108
Operating temperature	[K]	4.2; 10; 15
Overall current density	[A/m2]	180.72
Hoop stress (energized only)	[MPa]	<1.5	<4.5
Peak field (Bmax) @ Iop	[T]	0.75	2.02
Self-inductance	[mH]	0.18	9.74
Mutual-inductance	[mH]	0.13
Total inductance	[mH]	10.17
Total magnetic energy @ Iop	[J]	59.21

Fig. 1.

A cross-section profile of the prototype magnet and contours of axial magnetic flux densities at operating current of 108 A.

C. Superconducting Joint

The approach for superconducting joint of Hitachi’s wire is preliminarily to follow the procedure which is verified by our first-hand experience from Hyper Tech’s one. To establish a wire-bulk-wire connection, resistive Cu stabilizers of two bare wires were eliminated by an etching method. Both exposed wire ends were sheared with a proper angle and aligned together. Then, both wires were inserted into a stainless steel billet with pre-mixed Mg+B powder. Thereafter, the open end of the billet was partially sealed with a copper plug, and ceramic adhesive was pasted onto the top of the billet to avoid the evaporation of Mg during heat treatment at 600 °C for 12 h. During I_c measurement testing, one of the best samples tested at 20 K showed the I_c value of 255 A at 15 K, 0.25 T, which is applicable to our proposed persistent-mode operation at 108 A, though further improvement is necessary and currently ongoing [15].

D. Persistent Current Switch

A novel PCS concept for a persistent-mode magnet, in which one coil not only contributes to the field but also acts as a PCS, was first-ever proposed in this study.

Fig. 2 shows circuit diagrams of the test magnet during charging and persistent-mode with the proposed PCS operation. The circuit consists of L₁ and L₂ (self-inductances for coil1 and coil2), M (mutual-inductance), R_joint (joint resistance), R_heater (resistance of a heater).

Fig. 2.

Circuit diagrams of the test magnet during charging and persistent-mode with the proposed PCS operation.

As shown in Fig. 2(a), when T_PCS > 39 K (i.e., critical temperature of MgB₂), coil1 starts to become resistive (i.e., PCS open), and the power supply current (I_PS) can flow into coil2. After completion of charging, coil1 slowly recovers its super-conductivity with decreasing T_PCS (PCS closing), implying that the I_PS can flows into coil1 as well as coil2 under time-varying condition (see Fig. 2(b)). During ramp-down (when dI_PS/dt ≠ 0), the circuit model can be formulated by Kirchoff’s laws:

$$I_{\text{2}}=I_{PS}-I_{\text{1}}$$ (1)

$$L_2\frac{d{I}_2}{dt}-M\frac{d{I}_1}{dt}=L_1\frac{d{I}_1}{dt}-M\frac{d{I}_2}{dt}$$ (2)

Substituting Eq.1 in Eq.2, we obtain

$$(\frac{L_1+M}{L_2+M})\frac{d{I}_1}{dt}=\frac{d{I}_{PS}}{dt}-\frac{d{I}_1}{dt}$$ (3)

and

$$\frac{d{I}_1}{dt}=(\frac{L_2+M}{L_1+L_2+2M})\frac{d{I}_{PS}}{dt}=I_{op}\approx 0.97I_{PS}$$ (4)

Therefore, with decreasing the I_PS to 0, coil1 can be charged up to ~ 0.97dI_PS/dt (I₁ = I₂ = I_op). Then, both coils sustain the persistent I_op in a closed loop, as shown in Fig. 2(c).

E. Magnet Construction and Cooling System

Fig. 3 shows a photo of the magnet tested and a schematic drawing of a cross section of the conduction cooling system. One continuous MgB₂ wire insulated with S-glass was wound onto stainless steel former for coil1, and then wound onto copper former for coil2 with a constant winding tension of 10 N. After winding, both ends of the magnet were connected by a superconducting joint technique to achieve a closed loop. For thermal/electrical isolations of coil1 to act as a PCS, flexible Macor sheets, which is available for the heat-treatment at 600 °C, were placed between current lead/former as well as wire/former. After the heat-treatment for 12 h, a polyimide film heater (78.5 Ω at 300 K) to trigger the PCS was attached onto the surface of coil1.

Fig. 3.

A photo of the MgB₂ test magnet and a schematic drawing of a cross section of the conduction cooling system.

For the test environment of <15 K, a typical conduction cooling method was used for this study. The cryostat was comprised of a 2nd stage Cu plate, a radiation shield, a feed-through for voltage taps, thermocouples, and two heaters for heating 2nd stage and PCS, a pair of HTS current leads, one vacuum pump-out port, a two-stage Sumitomo cryocooler (SRDK-08D2). The cooling capacities of the 1st and 2nd stages of the cryocooler were 45 W at 40 K and 1 W at 4.2 K, respectively.

Table II lists the designed parameters of the cooling system. The pressure inside the cryostat was maintained at 10⁻⁷ Torr using a turbo pump in order to minimize convection heat input. The radiation shield was wrapped with 21 layers of super-insulation to reduce the radiation heat input. Two pairs of 6-mm wide GdBCO tapes were used to fabricate each HTS current lead. To enhance the heat exchange, Cu and stainless steel formers were directly connected to the coldhead by using OFHC sheets.

TABLE II.

Designed parameters of the cooling system

Parameter		Specification
		1st stage	2nd stage
Temperature	[K]	40	15
Conduction heat input	[W]	35.90	0.024
Radiation heat input	[W]	1.23	0.003
Joule heating @ 108 A	[W]	5.24	0.0004
Total heat input	[W]	42.38	0.027

III. Results and Discussion

A. Charging with Persistent Current Switch Operation

To validate the proposed persistent current switch (PCS) concept, typical charging tests were carried out at operating temperature (T_op) of 5 K with a PCS heater energy (E_PCS) of 0.15 W.

Fig. 4 shows the current, field, voltage, and temperature vs. time curves of the magnet obtained during charging at operating current (I_op) of 108 A. The power supply current (I_PS) for the target I_op was estimated as ~112 A, which is numerically obtained by Eq. 2. As shown in Fig. 4(a), when the heater was on, the temperature of PCS (T_PCS) within coil1 started to increase, reaching ~52 K (i.e., PCS open), and then the I_PS was applied to the magnet. As expected, the axial field (B_z) at the center of coil2 linearly increased with increasing the I_PS, and finally reached 1.79 T, which agreed well with the target magnetic field corresponding to the value of B_z per ampere (16.10 mT/A, coil constant of coil2 alone). Meanwhile, when the I_PS was applied, the voltage of coil2 (V₂) abruptly increased, then gradually decreasing to ~1.4 mV (L₂dI_PS/dt = 1.45 mV, inductance voltage of coil2 alone), and finally it dropped to 0 at 112 A (i.e., when dI_PS/dt = 0). These results imply that coil2 was fully charged at 112 A with a successful PCS operation of coil1 enabling open circuit as shown in Fig. 2(a). After charged at 112 A, the heater was off (PCS close) and then the I_PS started to ramp down. The discrepancy between the measured and ideal B_z may be attributed to having ~0.5 less turn a layer of actual winding due to manufacturing error. While the I_PS was decreasing, the B_z also decreased slightly, and finally then reached 1.75 T which is ~97 % of the initially induced B_z(1.79 T) at 112 A. Based on simple analysis of the circuit model (see Fig. 2(b)), when dI_PS/dt ≠ 0, the I_PS of 112 A can flow not only coil2 but also coil1, and eventually at I_PS = 0 the I₁ and I₂ became identical as the calculated I_op of 108 A (~0.97dI_PS/dt). Meanwhile, during ramp-down, the V₁ and V₂ were 0.028 and −0.027 mV, respectively, which is close to the inductance voltage of ~0.027 mV for coils 1 and 2 mutually coupled. This showed good agreement with the estimated results, validating the proposed approaches involving the PCS circuit model.

Fig. 4.

Current, field, voltage, and temperature vs. time curves of the prototype magnet obtained from charging testing at 5 K: (a) P/S ramp-up (b) P/S ramp-down.

Fig. 5 shows the voltage, current, and field vs. time curves obtained after charging and persistent-mode operation at 5 K. As shown in the previous Fig. 4, during charging the magnet at I_op = 108 A, the voltage (V_joint) of the joint was maintained at 0 constantly. When the I_PS reached almost 0, the V_joint started to increase and reached 2.5 μV, which suggested that the index loss of the joint occurs. Over a period of index loss, the B_z also decreased gradually and eventually reached 1.7 T (converted as I_op = ~104 A) at 4.4 h when the V_joint dropped 0. Thereafter, the B_z kept nearly flat during the continuously monitored period of 35 h, suggesting that the magnet successfully sustained the persistent center field of 1.7 T. During 35 h, the maximum and minimum measured fields were 1.70290 and 1.70262 T, respectively, indicating the joint resistance of < 1.2 × 10⁻¹¹ Ω which will be improved by our ongoing studies.

Fig. 5.

Voltage, current, and field vs. time curves obtained during charging and persistent-mode operation at 5 K.

B. Persistent-mode Operation with respect to T_op

For a detailed examination of the index loss of the joint part observed during previous test at 5 K, the V_joint and B_z of the magnet operated in the persistent-mode at 104 A were continuously measured with respect to T_op.

Fig. 6 shows voltage, field, and temperature vs. time curves of the persistent-mode magnet operated at the ranges from 5 to 8 K. When the heater energy (E_h) was applied to the 2nd stage coldhead, the temperatures increased, and then saturated at each level. As the E_h increased up to 3.5 W, the temperature of the joint (T_joint) increased up to 9.6 K while the V_joint remained unchanged. However, further increasing it up to 4 W led to the increase of the V_joint, indicating that the index dissipation occurred at 10 K with I_op = 104 A. Therefore, due to index dissipation of the joint, the B_z was decayed from 1.703 to 1.695 T, and it finally then sustained the persistent field with the recovery of T_joint. This means that the T_op of the test magnet with the joint performance is currently limited at <10 K.

Fig. 6.

Voltage, field, and temperature vs. time curves of the persistent-mode magnet with respect to T_op.

IV. Conclusion

Here we report the test results of construction and operation of a persistent-mode, liquid-helium-free, small-scale prototype MgB₂ for the development of a tabletop 1.5-T “finger” MRI system for osteoporosis screening. The prototype magnet composed of 2 coils, one superconducting joint, and a persistent-current switch (PCS) built from a portion of one coil. The magnet was cooled by a crycooler to the target operating temperature of < 15 K. At 5 K test, the magnet was successfully charged up to the estimated initial field of 1.79 T at the power supply current of 112 A, and then, during ramp-down, it finally reached the intended field of 1.75 T, which showed good agreement with the results estimated by the proposed PCS circuit model. After ramp-down, the field slightly decreased due to the index dissipation of the joint, and then the magnet sustained the persistent field of 1.7 T during 35 h, which validated the joint resistance of < 1.2 × 10⁻¹¹. Further studies are currently being carried out for the following purpose: 1) improvement of superconducting joint performance at higher temperature and field and 2) design/construction of 40-mm RT bore MgB₂ finger MRI magnet cooled by solid nitrogen.