Magnetic Resonance Microscopy. Группа авторов

is investigated with numerical simulations (CST Studio, Frequency Domain Solver), in the case of a disk whose first TE mode is located at 732 MHz (Eigenmode Solver of CST Studio). It was first checked that the excited mode was the TE_01δ mode: the magnetic field longitudinal component coincides in each configuration with the total magnetic field amplitude along the disk symmetry axis. As can be observed in Figure 2.11, the resonance frequency is slightly increased by the loop. The left column of Figure 2.11 shows that the maximum magnetic field at the center of the disk coincides with the minimum loop reflection coefficient, meaning the input power of the loop is transferred to the resonant mode of the disk. Also, but not shown in this figure, there exists an optimal loop diameter maximizing the magnetic field amplitude.

      Figure 2.11 Influence on (left) the reflection coefficient S₁₁ (minimum value), the magnetic field amplitude (maximum value) at the center of the resonator, and (right) the corresponding frequency of the feeding loop position relative to the dielectric disk (relative permittivity 530, loss tangent 8.10⁻⁴, diameter 18 mm, height 10 mm). (a) Varying lateral position. (b) Varying longitudinal position. There is an optimal position of the feeding loop with a minimum reflection coefficient, and a maximum transmitted power to the resonant mode.

      In the model proposed above (Section 2.3) for estimating the contribution of the ceramic probe to the SNR, the contribution of the feeding loop is neglected. In fact, this contribution depends on the loop’s geometry and material. For the geometry in Figure 2.11 and the considered probe prototype, the probe efficiency was compared for a lossy copper loop and an ideal loop in a perfect electric conductor; the difference was less than 0.2% between the two values.

The first hybrid mode can be excited as well by the same loop positioned on its lateral side. Another potential excitation source is an electric dipole antenna, generating a circulating magnetic field around the dipole axis and positioned on the lateral side of the resonator, perpendicular to the ring axis to excite the first TE mode, and parallel for the first hybrid electromagnetic (HEM) mode.

Figure 2.12 shows an example of an experimental setup exploiting a ceramic probe operating under its TE_01δ mode. In this setup, we use a circulating water blanket as a thermostat to homogenize and set the temperature of the ceramic probe at a given level. As depicted in Figure 2.13, the permittivity of the ceramic material and, as a result, the mode frequency, depend on its temperature, since reaching such a high permittivity value is enabled by working close to the Curie point of the ferroelectric material [7]. In [30], this dependence is used for real-time, accurate tuning of the probe to the Larmor frequency. However, in any other setup involving ceramic probes, even if another method is used for tuning, the temperature should be controlled to avoid the resonant frequency shifting away from the Larmor frequency.

      Figure 2.12 Experimental setup. Fixing the resonant frequency at the Larmor frequency is enabled by thermostating the resonator at a given temperature (left). Matching is performed by sliding the excitation loop (right). With permission from [30].

      Figure 2.13 Temperature dependence of the ceramic relative permittivity and therefore of the probe resonant frequency. With permission from [30].

      2.4.2 Performance

      The ceramic probe designed in [30] was experimentally compared to the reference probe, which is a solenoid coil optimally designed for the same required field of view.

The measured field patterns in Figure 2.14 confirm that the designed ceramic probe has the same region of homogeneity as the solenoid immersed in Fluorinert™ (the Fluorinert is used to improve both B₁ and B₀ homogeneity of the solenoid [31]), and induces higher transmit field intensity using less input power. Quantitatively speaking, for the same input power, the ceramic probe equivalently provides a field level 2.1 times higher than the solenoid immersed in Fluorinert.

      Figure 2.14 Measured transmit field pattern in a homogeneous liquid phantom for the reference optimal solenoid probe with Fluorinert (top) and for the designed ceramic (bottom), at 17.2 T. Data from [30].

The experimental SNR, measured in a tissue-mimicking liquid phantom, was also investigated for both probes in Figure 2.15 (left panel). With the same effective flip angle, the SNR was 2.2 times higher with the ceramic probe than with the solenoid immersed in Fluorinert. This has to be compared to the theoretical prediction: from Figure 2.7, for the designed ceramic probe (indicated as “proposed ceramics”), the semi-analytical method presented in Section 2.3 predicts an SNR gain of 2.5.

      Figure 2.15 Experimental comparison of the designed ceramic probe and the reference optimal solenoid at 17 T. (Left) signal and noise maps measured in a homogeneous tissue-mimicking liquid phantom. (Right) microscopy images obtained with both probes; (first line) Ilex aquifolium fruit; (second line) chemically fixed rat spinal cord; (third line) 3D rendering and image slice of a plant petiole (obtained with the ceramic probe). With permission from [30].

      Better performance of the ceramic probe can be observed in Figure 2.15 (right panel): microscopy images obtained with this probe demonstrate improved quality compared with those obtained using the reference probe, since more structural details can be distinguished.

      2.4.3 Dual Ceramic Coils

Since ceramic probes are based on dielectric resonators, it is possible to exploit the coupling phenomenon between such objects positioned close to each other. Figure 2.16 depicts the coupling of the TE_01δ modes of two ring resonators each holding a sample, but the principle is the same for empty rings or disks. When the resonators are placed one above the other, the electromagnetic coupling of the individual modes (at the same frequency f₀) gives rise to two coupled modes: a lower-frequency mode denoted (++), with a field distribution such that the magnetic field between the resonators is enhanced; and a higher-frequency mode (+−) such that the magnetic field in between is decreased. In [32] the (++) mode of two coupled disks is exploited to image a sample held between the disks. In [33] two samples are imaged simultaneously by using the (++) mode of two ring resonators, each of them holding a sample. The corresponding magnetic field distribution is depicted in Figure 2.16 (right field maps, second line).

Figure 2.16 Coupling model of the first TE modes of two ring resonators. From [33].

Due to coupling, for the same input power, the SNR in each sample of the coupled ceramic probe is degraded by a factor of compared with the single ceramic probe performance. However, in this study exploiting a prototype that had already overcome the reference solenoid probe by a