Radio frequency (RF) transmit (Tx) arrays with maximum degrees of freedom are a precondition to fully exploit parallel transmit (pTx) capabilities since they allow to trade of homogeneous excitation, specific absorption rate (SAR) and input power¹. Flöser et al.² evaluated different local and remote transmit arrays at 7T using microstrip lines (MSL) with meanders³ as transmit elements. These meander elements (ME) can be used to minimize inter element coupling depending on the application⁴. In this work we simulated and evaluated the performance of 16 remote transmit arrays with different RF coil elements and element combinations arranged in one ring.Finite difference time domain simulations were performed in CST Microwave Studio (CST AG, Darmstadt, Germany) on a high performance computing cluster employing four Nvidia Tesla M2090 graphic units for distributed computing. A cylindrical phantom (50 cm long, 25 cm diameter) filled with tissue simulating liquid (ԑ_r = 45.3, σ = 0.87 S/m) was positioned in the center of each pTx array (cyan in Figure 1). The gradient shield and the bore were also included in each simulation. Depending on the geometric complexity of the simulated pTx array 25.9 ± 9.0 million mesh cells were used. Capacitor losses were included by introducing the corresponding serial resistance in the co-simulation. Altogether 16 different arrays with different coil elements and coil element combinations were simulated (Figure 1A-P). This included loops (Figure 1F) with 20 cm diameter and chamfered edges, microstrip line elements (MSL) in Figure 1Q, ME with 32.5 mm meanders (arrow in Figure 1R), ME with 50 mm size (Figure 1S), electrically shortened dipoles (Figure 1T) all with a total length of 25 cm and self-resonant half-wavelength dipoles (Figure 1U) with a length of 46.6 cm. The distance to the phantom was 12 cm for the loops and 15.1 cm for all other element types. After simulation the B₁⁺ fields were exported to Matlab (The MathWorks, Inc., Natick, MA, USA) for further processing. There, a singular value decomposition was performed and the cumulated sum of singular values was calculated to characterize the degrees of freedom for RF shimming. For each array configuration magnitude least square shims were computed to homogenize the RF excitation. The standard deviation of the absolute value of the reconstructed B₁⁺ field is the basis of the comparison between the arrays.Regarding the S-parameters, a reflection of better than Sxx = -20 dB was achieved for all elements in all arrays. Average coupling values Sxy with the standard deviation as error bar (red in Figure 2) show optimum results for 4 and 8 short dipoles and the combination of 4 dipoles and 4 loops. The plot also depicts maximum Sxy values (blue) between identical element types and maximum coupling values Sxy between different element types if loops are present. As it can be seen, the maximum Sxy never occurred between different element types. To evaluate the degrees of freedom for each array, the cumulative sum of singular values¹ was calculated (Figure 3). Each 8 channel array outperforms every 4 channel design as it can be expected. If the shielded elements are compared, the ME with 50 mm meanders performs best, followed by the ME with 32.5 mm meanders and the MSL independent of the configuration (4, 8 elements or a combination with loops). The highest values are reached for the combination of 4 dipoles and 4 loops followed by the same setup but with shortened dipoles. This matches the observations of the accessible shimming homogeneity (Figure 4). Here, the lowest standard deviation of a magnitude least square shim is obtained for 4 dipoles and 4 loops followed by 4 shortened dipoles combined with 4 loops. If MSL, ME 32.5 and ME 50.0 are compared (Figure 4), in each case, 8 elements allow for a higher shim homogeneity than either 4 of those elements, 4 loops or the combination of 4 elements and 4 loops. This is not true for dipoles and electrically shortened dipoles. Consequently, a combination of dipoles and loops seems to outperform the other investigated arrays regarding pTx capabilities.Concerning the maximum shim homogeneity of the compared single row pTx arrays, a combination of dipoles and loops, followed by a combination of electrically shortened dipoles and loops, seems to be most promising. Further work will include investigations with an appropriate human voxel model and validation of SAR efficiency which becomes important en route to in vivo applications.