Purpose

Monolithic Transmission Line Resonators (TLR) consist of overlapping gapped conductor strips on the two sides of a low-loss dielectric substrate. By varying parameters such as trace width, substrate thickness and permittivity, number of gaps and number of turns per layer, it is possible to build a distributed capacitance resonant structure. The TLR design has been shown to be a good choice for the construction of RF coil arrays^1-3. In this work we propose and evaluate, for the first time, the use of a single-turn-single-gap TLR as a combined RF-transceive and B₀-shim element. We compare this design to a standard loop^4,5 in terms of both RF and B₀-shimming performance. The benefits of our combined TLR/B₀-shim design are improved RF performance due to the fixed (independent of loop size) and smaller number of lumped elements, and 2x increase in B₀-shim efficiency. Our results show the TLR element design to be an ideal building block for high-channel-count integrated Parallel Reception, Excitation and Shimming (iPRES) arrays⁵.

Methods

A single-element TLR was built on a 3.175mm thick Rogers RO5880 substrate. As shown in Figure 1, we chose a rectangular TLR shape with outer dimensions of 40mm x 75mm. For these dimensions, using TLR theory³ augmented by numerical simulations (FEKO, Altair), we found that the TLR self-resonance could be tuned to 298MHz using a strip width of 2.5mm. Three lumped-element capacitors (Johanson Technology EIA1111) were used for matching, fine tuning and DC block (C_M=20 pF, C_T=4.7pF, C_block=300pF). Three RF-chokes (Coilcraft 132-20SM, L=538 nH, Z=1kΩ@300MHz) were added to the circuit to allow shim current to flow. A single drive point was used for both RF and shim currents, reducing cable interactions compared to other designs⁴. Three more coils were built for experimental comparisons: a second TLR coil without chokes and DC block, and two simple loops with 4 capacitors per loop (3xC_T=6.8pF, C_M=27 pF), with and without chokes, as shown in Figure 2. All coils were matched and tuned to 298MHz with better than -15dB reflection coefficient. Unloaded and loaded Q factors (Q_U and Q_L, respectively) were measured in S₁₂ using a dual-decoupled probe and a Keysight ENA E-5080A VNA. Scanner measurements were performed on a GE Discovery MR950 7T scanner. In order to evaluate the efficiency of the four coils in generating B₁ and B₀ fields, Bloch-Siegert B₁⁺ and multi-TE B₀ mapping was performed on all coils with and without DC current. A 32-channel array with 4 rows of 8 elements each was simulated via Method of Moments (FEKO) to study B₁ performance and via Biot-Savart formula to study B₀ performance.

Results

Table 1 shows the Q factors measured for the four comparison coils. The Q ratio (Q_U/Q_L) was higher for the TLR-based coil compared to the standard loop, both in RF-only and RF-shim versions. This indicates stronger coupling to the phantom for TLR-based coils. However, the larger drop in Q ratio upon addition of chokes indicates greater sensitivity of TLR-based coils to the presence of chokes. The B₁ maps in Figure 3 confirm these trends. The B₁ efficiency of the TLR is higher than that of the standard loop (on average over the slice, +13% and +21% compared to the RF-shim and RF-only versions, respectively). The insertion of chokes caused a drop in B₁ efficiency, which was stronger for the TLR than the standard loop: -18% and -8%, respectively. The B₀ maps show that, as expected, the TLR was 2x more efficient than the standard loop for generating B₀ offsets. Figure 4 (center) shows that a 32-channel iPRES array will have similar B₁ efficiency as standard quadrature volume coils, and that using SAR-aware RF shimming, high B₁ homogeneity (<5%) can be achieved while maintaining similar efficiency (12μT for 1 kW input power) and lower SAR (54% average over the volume) compared to CP mode. Figure 4 (right) shows that the whole-brain B₀-shim performance of the 32-channel iPRES array is similar to 4th order spherical harmonic shimming.

Discussion and Conclusion

We have demonstrated that the TLR can be easily modified to act as a high performance coil element for combined Transmit, Receive and B₀-shimming. In this work, we tested a particular geometry in which the standard loop and the TLR needed the same number of chokes; despite this, the TLR was shown to be significantly more efficient at generating both B₁ and B₀ fields. Several inter-element decoupling designs are available, including capacitive, transformer or shielding ring decoupling^2,3. We therefore conclude that the TLR-shim design forms an excellent foundation for the first practical iPRES array.

Acknowledgements

The authors acknowledge research support from the NIH (P41 EB015891, 1 S10 RR026351-01A1), GE Healthcare, and Zurich MedTech AG (Sim4Science program).