Multi-tuned RF probes enables near-simultaneous interrogation of different nuclei within a single experiment.¹ This work describes the design and construction of a triple-tuned RF probe for detection of ²³Na, ³¹P and ¹H using traps. Traps may be used to dual-tune an RF probe, either by splitting the resonance of a single tuned circuit,² or by blocking coupling at the higher frequency when using a pair of resonant circuits.³ This work combines both methods to construct a triple-tuned probe consisting of a nested pair of loops (fig. 1). The probe is designed for use at 4T, with resonances at 45MHz (²³Na), 69MHz (³¹P) and 170MHz (¹H). The inner loop incorporates two traps, one to prevent coupling to the outer loop, which is tuned to ¹H, and a second to simultaneously tune the loop to ²³Na and ³¹P.

Circuit simulations were used to find trap component values that give strong blocking at the ¹H frequency and low loss at the ²³Na and ³¹P frequencies. The inner loop inductance and loaded Q-factor were 153nH and 30, respectively; trap inductors were assumed to have a Q-factor of 100. Two parameters were swept: the pass-frequency of the blocking trap (f₁), and the parallel resonance frequency of the splitting trap (f₂). Component values for the blocking trap were calculated first, with the blocking trap inductor (L_tr1) chosen to be twice the minimum possible value at each f₁. Splitting trap component values were then calculated, taking account of the frequency-dependent reactance of the blocking trap.

A pair of concentric square loops were then built using 6mm copper tape (inner loop 60×60mm², outer loop 100×100mm², inner dimensions). The loop response was measured using a network analyser (ZNB, Rohde & Schwarz) either directly connected to the loops or via a pair of loosely coupled sniffer loops. The inductor in the blocking trap was adjusted to minimise coupling to the outer loop at 170MHz. The splitting trap was then adjusted to correctly tune the 45MHz and 69MHz resonances. Q-factors were then measured while the probe was loaded with a phosphate-buffered saline phantom, and compared to those of isolated single-tuned loops of the same dimensions.

Figure 2 shows the circuit simulation results: (a,b,c) show calculated inductor and capacitor values for the splitting trap; (c,d) show Q-factors of the loaded trapped coil, relative loaded single-resonant loops of the same size, calculated as Q_rel = 1/(1+R_tr/R_loop), where R_tr and R_loop are the resistances of the trap and loaded loop, respectively; (e) shows the difference in trap resistance, relative to the loop resistance, at 45 and 69MHz. The circle indicates the operating point (f₁=31MHz, f₂=57MHz) chosen for the constructed probe, giving L_tr1=21nH, C_s1=238pF, C_p1=50pF, L_tr2=28nH, C_p2=270pF and C_s2=64pF. The resistance of the blocking trap is 0.04/0.08Ω at 45/69 MHz; for the splitting trap it is 0.48/0.73Ω, giving relative loaded Q-factors of 0.77 at both frequencies.

Measured S-parameters for the constructed probe (fig. 3) show the two resonant modes of the inner loop, the ¹H resonance of the outer loop, and low coupling between loops at the ¹H frequency. The table in fig. 4 presents the measured Q-factors of the triple-resonant probe, and those of the equivalent single-tuned loops at the three frequencies of interest.

Design of the blocking trap is straightforward, provided that the trap-mode frequency is placed well above the blocking frequency.⁴ There is a small reduction in trap resistance at the ²³Na and ³¹P frequencies as f₁ is reduced, accompanied by a small reduction in blocking efficiency at the ¹H frequency. Design of the splitting trap requires more attention; the trap resistance at the two resonance frequencies depends strongly on f₂, varying in opposite sense at the ²³Na and ³¹P frequencies. For the constructed probe, f₂ was chosen to balance the trap resistance at both frequencies.

The design approach demonstrated here, in which a dual-tuned X/Y resonator is decoupled from a separate ¹H resonator using a blocking trap, is well suited for use at UHF where the ¹H probe often uses a different design (dipoles, patches, etc.) to the non-proton resonator, which can use a more traditional approach such as a birdcage.

We have demonstrated a triple-tuned probe using a dual-tuned ²³Na/³¹P coil nested inside a ¹H resonator. While the blocking trap introduces very low loss at the lower frequencies, designing the splitting trap for low loss at both frequencies is difficult as these frequencies are relatively close together (γ_23Na /γ_31P=0.65). The design was demonstrated using ²³Na, ³¹P and ¹H, but is equally applicable to other combinations of nuclei.