The ideal current patterns corresponding to the RF excitation and reception of the transverse magnetization at UHF demonstrate a transition away from the reactive near-field region, and include contributions from the transition and source-free radiative wave regimes (3). Visualization of these ideal current patterns has provoked work into combining dissimilar RF element designs such as loops, dipoles, monopoles, and other variants. However, dissimilar radiation patterns yield new forms of coupling between elements and provide implementation barriers. Therefore, this work demonstrates the application of a new method - general coupling matrix synthesis - for decoupling arbitrary RF arrays and apply it to a planar, 8-channel, nested loop/dipole array.

The electromagnetic interactions of any RF array can be modelled as a system of loop-voltage equations described by Fig. 1a. Following a series of filter transformations highlighted in Fig. 1b and Fig. 1c, it is possible to synthesize a simple 2-stage cascaded filter, placed only between adjacent elements, that can perform decoupling between nearest- and further-neighbours in densely populated arrays. The decoupling circuit in Fig. 1d is realized for a planar, 8-channel loop/dipole array (see Fig. 2). Decoupling circuits, and placement of λ/4-admittance inverting cables between elements are visible in Fig. 2b,c, and d. Where cabling was not appropriate, matching was obtained via series match capacitors, shown in Fig. 2d in balanced form at the dipole input.

The loop-dipole array was composed of four resonant loops (dimensions: 23 x 14 cm with 4-mm wide struts) and four inductively shortened dipoles (dimensions: 21.5 cm with 7-mm wide struts) were implemented with 2 oz. copper traces routed atop 0.79-mm-thick garolite. Loops were positioned with 2.5-mm spacing between elements with edges filleted to a 3.2-mm radius. Each loop included six 2.2 pF distributed capacitors (100 Series: American Technical Ceramics) located equidistantly apart along the perimeter of the loop elements. Variable capacitors (1 – 30 pF, Johanson Manufacturing, NJ) were placed at the drive point and opposite thereof for tuning and matching. Dipoles were nested inside each loop, located along the virtual ground of their respective loop element. Six wire-wound inductors were placed along equidistant breaks to inductively shorten the dipole length and ensure resonance at 297.2 MHz. Dipoles were matched to 50 Ω via low-pass Pi matching circuits utilizing two variable capacitors (1 – 30 pF, Johanson Manufacturing, NJ) and one variable inductor (25 – 34 nH, Coilcraft, IL). Shielded baluns were placed at drive points corresponding to both the dipole and loop elements.

Full-wave electromagnetic simulations were performed using commercially available software CST Microwave Studio (Darmstadt, Germany) for the loop/dipole array both with and without the decoupling circuits applied. Experimental S-parameters were measured with a network analyzer (Agilent Technologies, model E5071C). The coil system was loaded with two concentric, transversally aligned gel phantoms (14.6 cm in diameter and 8.6 cm in height, each), placed approximately 2 cm from the array. The gel phantoms were composed of gadolinium chloride, agarose, and sodium chloride, in concentrations intended to mimic the human head. Q-ratios were measured for a single element in isolation, with and without decoupling circuits present. Q-ratios of the isolated element were measured without a coaxial cable or balun attached to the element. All loaded Q measurements were acquired with the phantom described above.

S-parameters of the computed decoupling solution produced by the synthesis algorithm are presented in Fig. 3. These contour plots correspond to the S-parameters obtained by CST simulation in Fig. 4a, as well as the experimentally measured S-parameters available in Fig. 4b. Coupled and decoupled magnetic field patterns, located along the cut-plane displayed in Fig. 2e, are presented in Fig. 5a and Fig. 5b, respectively. Transmission profiles for the decoupled array, corresponding to cut-plane Fig. 2e, in the phantom are presented in Fig. 5c demonstrating distinct sensitivity profiles. Channel combinations for the same phase weightings are presented in Fig. 5d and Fig. 5e, for the array in coupled and decoupled states, respectively. Q-ratios (unloaded-to-loaded) for a loop and dipole in isolation were 3.9 and 4.4, respectively. Located inside the array, with decoupling circuits, Q-ratios were 3.6 and 4.2 for the loop and dipole, respectively. This corresponded to a mean decrease in Q-ratio of 6.1% when placing elements into the full array with decoupling circuits. Insertion loss of individual decoupling circuits was -0.25 dB.General coupling matrix optimization computes physically realizable designs for decoupling the nested loop/dipole arrangement and shows potential for realizing highly unconventional RF arrays. Judicial choice of decoupling circuit lumped elements will decrease insertion loss and minimize Q-factor spoiling.