Introduction

Radiofrequency coils made from coaxial cable were recently introduced^1,2. The cable’s inner and outer conductors are cut at positions opposite from each other and allow, due to the skin-depth effect, three separate currents to flow (inner conductor, inner and outer surface of the shield). A self-resonant structure is obtained without additional lumped components when the reactance presented by the two coaxial stubs in series cancels the loop’s inductance. Coaxial coils are mechanically very flexible, but the coil diameter cannot be chosen independently from the desired resonance frequency when operated at self-resonance. This constraint can be overcome by increasing the number of turns and/or gaps^3-5. In this work, the influence of the choice of interfacing components on the achieved SNR is studied on a coaxial coil near its self-resonance. Also, as a safety measure in case of malfunctioning active detuning an effective position for placement of a fuse is investigated.

Methods

The coaxial coil is made from thin and flexible coaxial cable (Molex 047SC-2901, Lisle, IL, U.S.A.) with a coil diameter of 80 mm, which corresponds to a self-resonance near 123 MHz. All interfaced coils were tuned and matched (< -19dB) to 50 Ω at 123.2 MHz, measured with a network analyzer (Keysight Technologies E5071C, Santa Rosa, CA, U.S.A).

Miniaturization and choice of L_T:
To miniaturize the large tuning inductor introduced in¹ and allow for precise and easier tuning, a tunable capacitor C_T was placed in parallel to inductors of different (smaller) values (Fig. 1a) ranging from L_T = 27 nH (corresponding C_T=53.5 pF) to L_T=360 nH (no C_T). Q_unloaded and Q_loaded were measured with a double-loop probe on a phantom (25L container filled with deionized water, 1.6g NaCl/L, DC conductivity ≈0.2 S/m, 1mL/L Magnevist). To determine SNR, GRE MR images (T_E=10 ms, T_R=300 ms, α=60°, FoV=300x300 mm²) were acquired in a 3T MR scanner (Siemens Magnetom Prisma Fit). In a first measurement series varying L_T and C_T, the matching capacitors C_M1 and C_M2 were not symmetric, showing that SNR suffered considerably depending on the difference of the two. Another measurement series with equal C_M was therefore conducted. To further save space and reduce coupling between inductors, self-wound toroid inductors were built and miniature size preamplifiers with a size of 10.9x9.1x2.9 mm³ (MSM-123281, Microwave Technology, CA, U.S.A.) were used.

Fuse:
A fuse to spoil the resonance in case of too high current flow was placed in the middle of the outer conductor, since it is located at the position of the port and can therefore be included in an interface PCB within the same housing. To quantify the efficiency of the fuse, flip-angle maps employing the satTFL method⁶ (T_E=1.9 ms, T_R=6460 ms, α=80°, FoV= 399x399 mm²) were acquired with the body coil without and with the coaxial coil present. With the coaxial coil present, three configurations were studied: 1. with a normal working active detuning circuit, 2. with the active detuning circuit intentionally defective, and 3. with a blown fuse. The fuse was blown by running a sequence with high transmit power and the defective active detuning circuit.

Preamplifier decoupling:
A transformation of the preamp’s low input impedance to a short at the coil port suppresses current flow along the coil to minimize crosstalk in arrays. The necessary phase shift (>180°) was realized by combining a balun (Fig. 1b) and a lumped element high pass tee phase shifter (Fig. 1c) to avoid a long cable between the interface and the preamp. The balun shifts the phase by 180° and converts the balanced signal to an unbalanced signal, reducing common mode currents. After the balun, an unbalanced high pass tee is sufficient. The electrical circuit of the entire interface connected to the coaxial coil is shown in Fig. 2. Bench measurements and SNR scans as described above were performed to investigate the performance of the balun.

Results

The performance of the interfaces with varying L_T (and corresponding C_T) are summarized in Tab. 1. In the measurements with ΔC_M>0 (Tab. 1a) the Q-ratio (Q_U/Q_L) increases with increasing L_T, whereas SNR is related to ΔC_M. For the ΔC_M=0 (Tab. 1b), both Q-ratio and SNR are increasing with increasing L_T. The optimal single channel interface was found with L_T=360 nH, no C_T, and C_M1=C_M2=5.6 pF. The flip-angle ratio maps in Fig. 3 show (a) good decoupling from the body coil for the intact interface, (b) strong B₁ artifacts (up to ±50% change) in case of the intentionally malfunctioning active detuning and (c) again good decoupling after the fuse was intentionally blown. Connecting the balun to the single-channel interface increased both Q_U an Q_L by more than 20% percent, also SNR increased by 11%.

Conclusion

An optimized interface for a coaxial coil operated near its self-resonance was demonstrated: The importance of symmetric matching and the advantage of using a balun was shown, a fuse positioned at center of the outer conductor is efficient. Small toroidal inductors and a miniature preamplifier are proposed to save space and weight and thereby maintain flexibility when used in an array.

Acknowledgements

This project was funded by the Austrian/French FWF/ANR grant, No. I-3618, “BRACOIL“.