Modular Coils with Low Hydrogen Content Especially for MRI of Dry Solids

Abstract

Introduction: Recent advances have enabled fast magnetic resonance imaging (MRI) of solid materials. This development has opened up new applications for MRI, but, at the same time, uncovered new challenges. Previously, MRI-invisible materials like the housing of MRI detection coils are now readily depicted and either cause artifacts or lead to a decreased image resolution. In this contribution, we present versatile, multi-nuclear single and dual-tune MRI coils that stand out by (1) a low hydrogen content for high-resolution MRI of dry solids without artifacts; (2) a modular approach with exchangeable inductors of variable volumes to optimally enclose the given object; (3) low cost and low manufacturing effort that is associated with the modular approach; (4) accurate sample placement in the coil outside of the bore, and (5) a wide, single- or dual-tune frequency range that covers several nuclei and enables multinuclear MRI without moving the sample.

Materials and methods: The inductors of the coils were constructed from self-supporting copper sheets to avoid all plastic materials within or around the resonator. The components that were mounted at a distance from the inductor, including the circuit board, coaxial cable and holder were manufactured from polytetrafluoroethylene.

Results and conclusion: Residual hydrogen signal was sufficiently well suppressed to allow 1H-MRI of dry solids with a minimum field of view that was smaller than the sensitive volume of the coil. The SNR was found to be comparable but somewhat lower with respect to commercial, proton-rich quadrature coils, and higher with respect to a linearly-polarized commercial coil. The potential of the setup presented was exemplified by 1H/23Na high-resolution zero echo time (ZTE) MRI of a model solution and a dried human molar at 9.4 T. A full 3D image dataset of the tooth was obtained, rich in contrast and similar to the resolution of standard cone-beam computed tomography.

Fig 1. MRI of solids.

3D rendering of a skull found in a late roman settlement, estimated 300–400 A.D., and head coil acquired with an UTE sequence at 1.5 T (left). 3D maximum intensity projection of a ¹H-ZTE image of a quadrature mouse coil at 9.4 T (center, QR₂) and corresponding photograph (right).

Fig 2. Representative photograph of several loop-gap inductors that were constructed (LG_1–6). Note the ledges used for connection. Dimensions are provided in Table 2.

Fig 3. Photographs and schematics of the single-tune circuit boards (CB₁, left and CB₂, center) and dual-tune circuit board (CB₃, right) with and without loop-gap (LG) inductors.

Conventional BNC connectors (left and right) as well as a custom-made PTFE coaxial cable with low hydrogen content are shown (center). Values of variable capacitors are: C_V1,3-9 = 0.3–3.5 pF, C_V2,10,11 = 1.1–16 pF; fixed-value capacitors: C_F1 = 1 pF, C_F2 = 100 pF, C_F3 = 6.8 pF, C_F1 = 10 pF; inductors: L_1,2 = 22 nH.

Fig 4. Photograph of the complete coil assemblies, CB₁ with LG₄ (front) and CB₂ with LG₂ (back), ready to be mounted to the bores of the 7 T and 9.4 T MR systems.

Fig 5. ¹H-MR signal height of a tooth acquired with unlocalized spectroscopy as a function of pulse power using a constant flip angle (α ≈ 6°) at 9.4 T with CB₂ and LG₁.

Note that the MR signal decreased for powers exceeding 22 W.

Fig 6. 3D maximum-intensity-projection ZTE MRI and approximate coil positions (left) acquired at 9.4 T and photograph of coil assembled from CB₂ and LG₂ (right), either with a conventional BNC connector and coaxial cable (top), or a custom-made PTFE cable that was soldered to the circuit board (bottom).

Note that the signal originating from the BNC connector was removed using the custom build cable. In turn, signal originating from the variable capacitors appeared close to the noise level (FOV = 20 cm).

Fig 7. ¹H-ZTE images of a dried musca domestica acquired ex vivo with a commercial quadrature coil (left, QR₁) and new, hydrogen-poor coil composed of CB₁ and LG₂ (right) with identical settings at 7 T.

Note that the strong artifacts on the left were induced by signal sources from outside of the FOV that was (3 cm)³. In contrast, these artifacts did not appear when the modular loop-gap coil was used (right).

Fig 8. Photograph (top left), orthogonal reconstructions of a 3D ZTE MRI of an extracted human molar acquired at 9.4 T with CB₂ and LG₁ (top row) and 3D CBCT (bottom row).

Note that the brightness of the MRI on the right was adjusted to show the enamel. Dashed lines indicate the slice positions, dotted lines the approximate outline of the enamel.

Fig 9. ¹H- and ²³Na-ZTE MRI of an aqueous model solution containing 100 g NaCl / l acquired with LG₄ and CB₃ at 9.4 T.