Implementation of an 84 Channel Actively Shielded Matrix Gradient Coil

Sebastian Littin¹, Feng Jia¹, Kelvin J. Layton¹, Huijun Yu¹, Stefan Kroboth¹, and Maxim Zaitsev¹

¹Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany

Synopsis

In this abstract, we present an 84 channel actively shielded matrix gradient coil. This coil was built and integrated into our 3T MRI scanner. Functionality and characterization measurements such as high voltage tests, eddy current and field map measurements were successfully performed. This system allows for generating spatial encoding fields in a highly flexible fashion, which enables the development of novel imaging techniques.

Introduction

Increasing the degrees of freedom usually allows for novel, sometimes advantageous solutions to a problem. The introduction of multiple receiver coils, e.g., allowed for parallel image acquisition. Multiple imaging techniques relying on nonlinear spatial encoding magnetic fields (SEMs) have been presented recently. Examples are parallel image acquisition with nonlinear gradients [1-5], curved slice imaging [6,7] and phase preparation for reduced field-of-view imaging [8,9]. All these techniques may benefit from an encoding system capable of generating arbitrary SEMs. A non-shielded multi-coil array, originally intended for shimming has been introduced previously [10,11], which can in principle generate a wide variety of SEMs. The design of an actively shielded matrix gradient coil was presented recently [12]. In this abstract, we present a prototype matrix gradient coil that was built and integrated into our system. First results characterizing the performance of the coil are presented.

Materials and Methods

An 84 channel matrix gradient coil was realized by manufacturing 84 carrier elements using a powder bed and ink-jet head 3D printer (3D Systems ZPrinters, Rock Hill, SC, USA). Due to its mechanical flexibility and zero internal eddy currents, Litz wire was used for winding each element and as feeding wires. The elements were designed with no interconnections to be soldered inside the coil. Two different element types were arranged in 7 rings with 12 elements, each. All elements form a cylindrical structure with two main current carrying surfaces and a common secondary current-carrying surface used for shielding (Figs.1 & 2). 120 copper tubes were integrated for water cooling. The whole structure was enclosed air tight by an outer shell constructed from the following parts: top and bottom rings made from PVC, rails made from PA6 and plates made from GRP (Fig. 3). Impregnation with epoxy was done under vacuum.

The coil was integrated into a 3T system (Siemens Healthcare, Erlangen, Germany) equipped with 12 additional gradient power amplifiers (IECO, Helsinki, Finland) and custom built control electronics. An open source framework [13,14] was used for pulse sequence programming. This framework was adapted to control the additional hardware and the clinical system simultaneously. Field maps and eddy current maps were measured by mapping the phase evolution using the linear gradient system of the MR scanner similar to [15]. Groups of 3 to 8 elements were connected serially in 12 clusters for first imaging experiments. These clusters were optimized [16] for generating linear SEMs.

Results and Discussion

High voltage tests up to 500V between neighboring elements and water cooling were performed successfully. Measured field maps of a single element (Fig. 4) are as expected and field deviations are within the measurement accuracy. Eddy current measurements show eddy currents below 2% (Fig. 5) of the generated field strength and can be further reduced with amplifier pre-emphasis. However, since the matrix gradient coil was optimized as a scaled down whole body system, the optimization was done for an eddy current generating surface much closer to the coil where predicted eddy currents are around 5% [17]. Simulated field maps based on 12 measured cluster field maps demonstrate the ability to generate linear x- and y-gradients with an error of 5.5% and 3.2%, respectively (Fig.6). Results of first imaging experiments are presented at this conference by Layton et al.

Conclusions

A shielded matrix gradient coil was successfully built and integrated into a 3T MR-scanner. Characterization measurements were performed and demonstrate the good coil performance. The highly flexible generation of SEMs will allow for novel acquisition techniques based on nonlinear encoding fields.

Acknowledgements

This work is supported by the European Research Council Starting Grant 'RANGEmri' grant agreement 282345.

References

[1] J. Hennig, et al. Parallel imaging in non-bijective, curvilinear magnetic field gradients: a concept study. Magn Reson Mater Phys. 2008; 21:5-14.

[2] G. Schultz, et al. Reconstruction of MRI Data Encoded With Arbitrarily Shaped, Curvilinear, Nonbijective Magnetic Fields. Magn Reson Med. 2010; 64:1390-1404.

[3] D. Gallichan, et al. Practical considerations for in-vivo MRI with higher dimensional spatial encoding. Magn Reson Mater Phys. 2012; 25:419–431.

[4] J. Stockmann, et al. O-space imaging: highly efficient parallel imaging using second-order nonlinear fields as encoding gradients with no phase encoding. Magn Reson Med. 2010; 64:447–456.

[5] L. Tam, et al. Null space imaging: nonlinear magnetic encoding fields designed complementary to receiver coil sensitivities for improved acceleration in parallel imaging. Magn Reson Med. 2012; 68:1166–1175.

[6] H. Weber, et al. Excitation and geometrically matched local encoding of curved slices. Magn Reson Med. 2013; 69:1317–1325.

[7] H. Weber, et al. Local Shape Adaptation for Curved Slice Selection. Magn Reson Med. 2014; 72(1):112-123

[8] W. Witschey, et al. Localization by non-linear phase preparation and k-space trajectory design. Magn Reson Med. 2012; 67:1620–1632

[9] S. Littin, et al. Zoomed Cardiac CINE-MRI Using Nonlinear Phase Preparation. Proc 22th ISMRM, Milan. 2014; #0428.

[10] C. Juchem, et al. Multi-slice MRI with the dynamic multi-coil technique. J Magn Reson. 2010; 204(2):281-289.

[11] C. Juchem, et al. Multi-coil magnetic field modeling. J Magn Reson. 2013; 236:95-104.

[12] S. Littin, et al. Shielded Matrix Gradient Coil. Proc 23th ISMRM, Toronto. 2015; #1022.

[13] K. Layton, et al. A Hardware-Independent Environment for MR Acquisition and Simulation. Proc 23th ISMRM, Toronto. 2015; #2484.

[14] http://pulseq.github.io/

[15] A. Welz, et al. Development and Characterization of An Unshielded PatLoc Gradient Coil for Human Head Imaging. Concepts Part B. 2013; 43(4):111-125.

[16] S. Kroboth, et al. Optimization of Matrix Gradient Coil Switching for a Limited Number of Amplifiers. Proc 23th ISMRM, Toronto. 2015; #1836.

[17] F. Jia, et al. Design of a Shielded Coil Element of a Matrix Gradient Coil. Proc 23th ISMRM, Toronto. 2015; #3091.

Figures

Figure 1: Matrix gradient coil during the build process. Two different element types forming two layers of main current carrying surfaces are depicted. Each element consists of two curls and was optimized for high gradient field strength, low eddy currents and low mutual coupling to elements on the neighboring current carrying surface [17]. The basic design of one element without shield is intrinsically balanced for force and torque.

Figure 2: Elements are held in place by the water cooling rods and slide into each other during the assembly.

Figure 3: Completed matrix coil insert on handling unit at the backside of the scanner. Outer windings of the shielding layer are visible through the GRP shell. Electrical connections of each individual coil element are accessible on the side panels of the handling unit.

Figure 4: Representative measured field maps of a central slice in rad/s/A of a single element in (left) coronal (right) transversal orientations.

Figure 5: Time evolution of measured eddy currents of a single coil element of a single pixel in a coronal slice.

Figure 6: Simulated field maps with a diameter of 19cm of a linear x (left) and y (right) gradient field. These SEMs can be generated by 12 clusters of serially connected elements and have been simulated to demonstrate the feasibility to generate even linear SEMs.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)

3561