It is well established that transmit arrays are essential to mitigate B₁⁺ inhomogeneities at ultra-high field (UHF, ≥7T) [1, 2]. Furthermore, dual-row transmit arrays provide better RF performance in terms of reduced local SAR compared to single row transmit arrays [3-5]. In this work, we develop a 2x8 transmit array for a 7T 16-channel parallel transmit (pTx) system. Compared to our previous inductively decoupled design [5], we evaluated here transmit array with geometrically overlapped larger transmit elements. For high sensitivity during reception, a 31-channel tight fitting receive array was combined with the transmit array.

Experiments were performed on a Siemens 7T (297.2MHz) whole body MR scanner.

Transmit array: Sixteen loops were arranged in two rows, covering 20cm along the Z-direction, on a 275mm tube with 2.5mm wall thickness. Increasing the sample loading and integrating the receive array without influencing transmit performance were the factors that determined the configuration of the transmit array (Fig.1). Adjacent elements (transverse and longitudinal) were geometrically decoupled and the diagonal elements were decoupled using inductors. The two rows were stacked along Z-direction, without any angular rotation between them. This provides a straight forward solution to route the receive array cables through the middle of the transmit elements. Each loop consisted of 12 capacitors, two decoupling inductors and a PIN diode for active detuning.

Receive array: 31-receive elements were arranged in four rows on a helmet shaped former (L/R – 186mm, A/P – 220mm, S/I – 220mm) [5]. Receive elements were arranged around the helmet, following the contours of the brain anatomy. The top two rows had 10-elements each, forming a complete ring while 3rd and 4th row formed partial rings with 7 and 4 elements, respectively. Adjacent elements within the row were inductively decoupled. Each element of the lower row geometrically overlapped with two elements of the upper row. A figure-8 coil, with eye cut-outs, is planned for the 3rd row. This might contribute to enhancing the central SNR. A protection fuse (315mA) was installed in each receive channel. The entire receive electronics with preamplifiers (WanTcom, MN, USA), cable traps and scanner interface were assembled on to a holder mounted on top of the helmet, above the FOV of the transmit array (Fig.2).

Scanner interface: The coil can be interfaced to either single channel transmit or 16-channel pTx system. The initial results presented here were acquired in the single channel mode by driving the coil through a 1x16 splitter (Werlatone, NY, USA). For CP mode, the two coils in the same column had same phase and adjacent elements within the row had 45° phase increments. For active detuning, 16 of the 32 PIN-bias lines were routed to the transmit coil and the remaining 16 controlled the 32-receive channels. SNR maps and B1+ maps were acquired using a head and shoulder phantom filled with tissue equivalent solution [5]. For performance evaluation, a commercial 8Tx/32Rx array was used as the baseline.

Highest coupling in the transmit array was between the adjacent elements within the row (-8 to -11dB). The average isolation between the overlapped elements in the same column and the diagonal elements was -20.6dB and -24.8dB, respectively. All receive elements were matched to better than -18dB and the average preamplifier decoupling was 21.9dB. The unloaded Q of a single isolated receive element was 194 and the Q-ratio varied from 10 to 6 for a distance of 10 to 20mm from the phantom. The average active detuning of the transmit and receive array was 32dB and 36dB, respectively.

The transmit efficiency of the dual-row coil is within 7% of the baseline coil (Fig 3). It is interesting to note that the B₁⁺ profile in CP mode drops abruptly towards the end of the coil, though outside the brain volume. However, the dual-row configuration provides additional degrees of freedom to extend the longitudinal coverage, if required.

The SNR of the 31-channel receive array was better in the periphery, especially in the dome of the helmet. Similar SNR was achieved in the center of the brain (Fig.4). This gain is most likely due to the continuous arrangement of coil elements around the helmet and additional receive elements in the dome (20 coils in top two rows). Whole brain coverage with sufficient uniformity was achieved as demonstrated in the spin-echo in-vivo images (Fig. 5).

Initial performance evaluation of the developed coil array in CP mode shows whole brain coverage and high SNR. The full RF capability of the dual-row coil setup will be evaluated in a 16-channel pTx system.