Field Mapping and Automated Shimming of an HTS Magnet by ...

IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 25, NO. 3, JUNE 2015

4300804

Field Mapping and Automated Shimming of an HTS Magnet by “Internal” Active Shim Coils Located in the Bore of the Magnet Min Cheol Ahn, Jeongwoo Jang, Seungyong Hahn, Young-Gyun Kim, and Haigun Lee

Abstract—In order to obtain the target field homogeneity of a nuclear magnetic resonance (NMR) or magnetic resonance imaging (MRI) magnet, commonly used are active shim coils, i.e., axial and radial, that typically locate radially outside of the magnet. In a high-temperature superconducting (HTS) magnet, however, we reported a “strong” hysteresis in the field-to-current ratio of an external (placed radially outside the HTS magnet) active shim coil by the screening-current-induced fields (SCFs). This nonlinear behavior of an external active shim due to the SCF is one of the major technical challenges to achieve the target field homogeneity of an HTS NMR or MRI magnet. In this paper, we constructed and operated active shim coils, i.e., two axial (Z1 and Z2) and two radial (X and Y ), that were installed inside of an HTS magnet to investigate effectiveness of internal shim coils. The HTS magnet, wound with GdBCO coated conductors, consists of ten double pancakes. With the active shim coils, a customized field mapping and shimming system was constructed, which consists of a 3-D field mapper, a field analysis software package, and a set of power supplies to control the shim coils. A pre-shim field error of 953 ppm at a 10-mm-diameter cylinder was improved to 464 ppm after iterative automated shimming using the internal shim coils. The results demonstrate that the internal shim coils are effective to eliminate the SCF-oriented field errors, both radial and axial, which was practically impossible for the external shim coils. Index Terms—Field homogeneity, field mapping, hightemperature superconducting (HTS) magnet, internal active shim.

I. I NTRODUCTION

H

IGH-TEMPERATURE superconducting (HTS) magnets have been attempted to high magnetic field applications such as insert coils for GHz-class (> 23 T) LTS/HTS NMR [1], [2] and all-HTS MRI, extremity [3] or brain [4]. In order to obtain a target field homogeneity of an NMR or MRI magnet, commonly used are active shim coils, axial and radial, that

Manuscript received August 12, 2014; accepted October 20, 2014. Date of publication October 24, 2014; date of current version January 30, 2015. This work was supported by National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (No. 2012-046999). M. C. Ahn and J. Jang are with the Department of Electrical Engineering, Kunsan National University, Gunsan 573-701, Korea (e-mail: mcahn@kunsan. ac.kr). S. Hahn is with the Francis Bitter Magnet Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139 USA. Y.-G. Kim and H. Lee are with the Department of Materials Science and Engineering, Korea University, Seoul 151-744, Korea (e-mail: haigunlee@korea. ac.kr). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TASC.2014.2364717

TABLE I HTS M AGNET S PECIFICATION

typically locate radially outside of the magnet. In an HTS magnet, however, we reported a “strong” hysteresis in the fieldto-current ratio of an “external (placed radially outside the HTS magnet)” active shim coil, mainly due to the screeningcurrent-induced fields (SCFs) [5]. This nonlinear behavior of an external active shim is one of the major technical challenges to achieve a target field homogeneity of an HTS NMR or MRI magnet. As an alternative approach, we propose “internal” active shim coils that locate radially inside the bore of a magnet to mitigate any detrimental impacts of SCF on the shim coil performance. In this paper, 4 active shim coils, 2 axial and 2 radial, have been constructed with copper wire. Installed inside of an HTS magnet, they were tested in a bath of liquid nitrogen at 77 K to investigate performance of the internal shim coils. The results demonstrate that the internal shim coils are effective to eliminate the SCF-oriented field errors, which has never been achieved by the external shim coils to date.

II. E XPERIMENTAL S ETUP A. HTS Magnet An HTS magnet, wound with SuNAM GdBCO coated conductor, was designed and manufactured. The magnet consists of 10 double pancakes (DPs). All DPs were wound without turn-to-turn insulation, so called as no insulation. Table I shows specifications of the magnet. The self-field critical current of conductor is about 200 A at 77 K. To improve a pre-shimmed field homogeneity near the magnet center, each half of the 10-DP stack is axially separated from each other by 5 mm,

1051-8223 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

4300804


TABLE II S PECIFICATIONS OF S HIM C OILS

Fig. 1. Photograph of the HTS magnet used in the experiment.

Fig. 2. Three-dimensional mapping and automated shimming system (3D-MASS).

though electrically connected in series. Fig. 1 shows a photograph of the HTS magnet used in this experiment. B. Three-Dimensional Mapping and Automated Shimming System (3D-MASS) In our previous work [6], the magnetic field distribution was measured by a 3-D field mapper and analyzed using spatial harmonic coefficients. For this research, a new feature on an automated active shimming has been added to the mapper. The operational procedure of our 3D-MASS (3-D-Mapping and Automated Shimming System) is as follows. Firstly, magnetic fields within a target cylindrical volume of Φ10 mm × 10 mm length were measured by the 3-D field mapper. Secondly, spatial field harmonics were obtained using homemade software based on the associated Legendre polynomials. Finally, from the obtained field gradients, a target current for each shim coil was calculated and automatically supplied by an individual power supply using an active feedback control. Fig. 2 shows a picture of the 3D-MASS. It consists of a magnet power supply, a 3-D mapper, 4 power supplies for shim coils, and a LabVIEW control part.

Fig. 3. Shim coils fabricated for experiments: (top left) axial shims, (bottom left) radial shims, and (right) assembled shim coil set.

C. Active Shim Coils Two axial shim coils, Z1 and Z2, and two radial shim coils, X and Y , were fabricated. All shim coils were wound with AWG24 insulated copper wire whose diameter is 0.5 mm. Axial shim coils were wound on a 40-mm-diameter cylinder, and detailed specifications of the coils including axial position are described in Table II. Two radial shim coils were fabricated on a flexible plastic sheet, and then the sheet was wrapped directly on the axial shim coils. Fig. 3 shows fabricated shim coils, axial and radial, and the total assembly. The shim coil set was placed radially inside the magnet bore as shown in Fig. 4. D. Mapping Parameter A Lakeshore cryogenic Hall sensor, mounted on the 3D-MASS, was used to measure magnetic fields. The total number of unknown coefficients for our 2nd order harmonic field analysis is 9. To minimize the measurement uncertainty, a total of 44 mapping points were set along a cylindrical

AHN et al.: FIELD MAPPING AND AUTOMATED SHIMMING OF AN HTS MAGNET BY “INTERNAL” SHIM COILS

4300804

Fig. 5. Mapping results on shimming procedure. TABLE IV F IELD G RADIENTS B EFORE AND A FTER S HIMMING Fig. 4.

Schematic view of shim coils and HTS magnet. TABLE III S HIM C OIL S TRENGTHS M EASURED

B. Pre-Shim Mapping Result helix with 4 revolutions on the surface of the target cylindrical volume of Φ10 mm × 10 mm. MATLAB was used to solve this overdetermined problem. The magnetic fields were measured 20 times at each mapping point and then averaged.

III. E XPERIMENTAL R ESULT AND D ISCUSSION

For pre-shim field mapping, the magnet was charged at a constant current of 30 A and the magnetic fields at the target 44 mapping points were measured. After the completion of charging, we waited for 1 hour until the initial SCF decay slowed down. The mapping result was shown in Fig. 5 (empty squares). Overall peak-to-peak homogeneity was 953 ppm. Field gradients of Z1, Z2, X, and Y are listed in Table IV.

A. Shim Coil Strengths and Linearity Prior to the main test of the shim coils with the HTS magnet, each shim coil strength was measured independently at room. Then, the shim coil set was installed inside the HTS magnet bore and the whole system with the HTS magnet and the shim coil set was immersed in a bath of liquid nitrogen. The measured field gradients are listed in Table III. Linearity R2 values of shim coils were 0.9994, 0.9999, 0.9997, and 0.9995 for the respective Z1, Z2, X, and Y . Unlike the nonlinear behavior of the external shim coil, located radially outside of a magnet [5], all shim coils show a good linearity, which demonstrates that the internal shim coil is less prone to the detrimental effects of SCF than the external shim coil. The measured shim strength of X (10.11 Gauss/m/A) was 9.3% smaller than that of Y (11.15 Gauss/m/A), chiefly due to the manufacturing error during winding and wrapping.

C. Shimming Test Based on the pre-shim mapping result, the 3D-MASS automatically calculated a target current for each shim coil. For example, since pre-shimmed Z1 gradient and Z1 shim strength were 50.47 Gauss/m and −47.50 Gauss/m/A respectively, the Z1 shim coil current was set to be 1.1 A. Target currents of X, Y , and Z2 were 1.5 A, −1.9 A, and −1.5 A, respectively. Each shim coil current was independently supplied by an individual power supply. The shimming was conducted in an order of Z2 → Z1 → X → Y . Magnetic fields were mapped and analyzed at each shimming step. Fig. 5 shows mapping results on the first 4 shimming steps. In addition to the designated field gradient, each shim coil generates other field gradients chiefly due to the manufacturing uncertainty of shim coils. Therefore, the

4300804


Fig. 6. Estimated NMR lineshapes before and after shimming.

3D-MASS updated the previous shim currents and remapped the fields until no further changes were required for the active shim coils at each shimming step. The final field gradients after shimming were also listed in Table IV. Actively shimmed 4 gradients—Z1, X, Y , and Z2—were drastically reduced. Specifically, the Z2 gradient, known as a major error by the SCF, was reduced to 11.72 Gauss/m2 , 0.34% of the pre-shim value, −3414.03 Gauss/m2 . D. Estimated NMR Lineshapes Before and After Shimming The post-shim field mapping result was shown in Fig. 5 (solid squares). The final overall homogeneity was 464 ppm, ∼50% of the pre-shim value, 953 ppm. Note that the other “unshimmed” field gradients (ZX, ZY , C2, and S2) are barely changed even after the shimming. It can be inferred that ZX (ZY ) gradient still remains in the after-shimming plot. For more homogenous result, ZX and ZY shim coils should be effective. The results demonstrate the superior performances of the internal active shim coils over those of their external counterparts to eliminate the spatial field errors of an HTS magnet, especially SCF-related. Finally, 1 H NMR lineshapes were estimated from the mapping results as shown in Fig. 6 [7]. The half-peak bandwidth was reduced from 5.9 kHz (pre-shim) to 2.0 kHz (after-shim) mainly owing to the field gradient reduction in Z1, Z2, X, and Y . IV. C ONCLUSION We proposed a set of “internal” active shim coils located radially inside the cold bore of an HTS magnet. Located radially outside an HTS magnet, typical active shim coils were known to be prone to the detrimental effects of the screening-current-

induced fields (SCFs) on the performance of the “external” shim coils. To verify the proposed internal shim coils, four shim coils, 2 axial and 2 radial, were designed, fabricated, and tested with an HTS magnet in a bath of liquid nitrogen at 77 K. The homemade HTS magnet consisted of a stack of 10 doublepancake (DP) coils. We also developed a 3-D mapping and automated shimming system (3D-MASS) for the tests of the shim coils with the HTS magnet. The peak-to-peak homogeneity of the magnet was reduced from 953 ppm to 464 ppm by use of the proposed internal shim coils and 3D-MASS. In accordance, the half-peak bandwidth of the estimated 1 H NMR lineshape was also reduced from 5.9 kHz to 2.0 kHz. For better field homogeneity, some more shim coils, such as ZX, ZY , C2, and S2, are needed. In this paper, we have firstly and experimentally demonstrated the “internal” active shim coils. It can be said that this paper is the meaningful first step for alternative method of an effective active shimming for HTS magnets. Although the copper active shim coils were used in this paper, superconducting wires such as NbTi or MgB2 can be used for the shim coils in NMR/MRI magnet. In addition, even copper shim coils can be located in room temperature bore for preventing evaporation of LHe due to a heat intrusion. The drift of the power supplies, for the HTS magnet and shim coils, could affect the homogeneity and temporal stability. Furthermore, the magnetic field data by the hall probe could not be very stable, even though we used some regression using overdetermined system. More intensive researches about this internal shimming, including NMR probe and field-locking technique, should be investigated in the near future. R EFERENCES [1] J. Bascuñán, S. Hahn, Y. Kim, J. Song, and Y. Iwasa, “90-mm/18.8-T allHTS insert magnet for 1.3 GHz LTS/HTS NMR application: Magnet design and double-pancake coil fabrication,” IEEE Trans. Appl. Supercond., vol. 24, no. 3, Jun. 2014, Art. ID. 4300904. [2] Y. Yanagisawa et al., “Operation of a 500 MHz high temperature superconducting NMR: Towards an NMR spectrometer operating beyond 1 GHz,” J. Mag. Res., vol. 203, no. 2, pp. 274–282, Apr. 2010. [3] R. A. Slade, B. J. Parkinson, and R. M. Walsh, “Test results for a 1.5 T MRI system utilizing a Cryogen-free YBCO magnet,” IEEE Trans. Appl. Supercond., vol. 24, no. 3, Jun. 2014, Art. ID. 4400705. [4] Y. Terao et al., “Newly designed 3 T MRI magnet wound with Bi-2223 tape conductors,” IEEE Trans. Appl. Supercond., vol. 23, no. 3, Jun. 2013, Art. ID. 4400904. [5] S. Hahn, M. C. Ahn, J. Bascuñán, W. Yao, and Y. Iwasa, “Nonlinear behavior of a Shim coil in an LTS/HTS NMR magnet with an HTS insert comprising double-pancake HTS-tape coils,” IEEE Trans. Appl. Supercond., vol. 19, no. 3, pp. 2285–2288, Jun. 2009. [6] M. C. Ahn, S. Hahn, and H. Lee, “3-D field mapping and active shimming of a screening-current-induced field in an HTS coil using harmonic analysis for high-resolution NMR magnets,” IEEE Trans. Appl. Supercond., vol. 23, no. 3, Jun. 2013, Art. ID. 4400804. [7] S. Hahn et al., “Field mapping, NMR lineshape, screening current induced field analyses for homogeneity improvement in LTS/HTS NMR magnets,” IEEE Trans. Appl. Supercond., vol. 18, no. 2, pp. 856–859, Jun. 2008.