Synopsis
Radiative dipole and monopole coil
arrays are increasingly used for ultrahigh field MRI, but few decoupling methods
have been proposed for radiative arrays. To overcome this problem, we
propose a Tunable-Defected-Ground-Structure (TDGS) method to decouple monopole arrays at 7T.
This concept was successfully validated by EM simulation, bench test and MR
experiments. By using the TDGS method, the cross-talk between two
closely-spaced monopoles was reduced from -7 dB to -25 dB. It was also found that
the TDGS method had little effect on the original B1 fields of the individual
monopole elements.Target
Audience
Researchers
interested in RF coil array design.
Introduction
Radiative
dipole and monopole coil arrays are uniquely simple to design and can penetrate
deep into the body, and are increasingly used for ultrahigh field MRI
1-3.
One technical challenge in building densely-spaced
radiative arrays with many coil elements is to suppress the strong
electromagnetic (EM) coupling between elements. In telecommunication applications, various
decoupling methods for monopole and dipole antennas have been described. These
approaches fall into three categories: (I)
Eigenmode feed networks
4; (II)
External resonators
5; (III)
Slits in ground the ground plane, i.e., defected ground structure (DGS)
6.
Compared with other decoupling methods, the DGS method causes the least disturbance
to the monopoles’ radiation patterns. However, direct implementation of this
method in ultra-high field MRI is limited by the relatively long MRI wavelengths,
which would require impractically deep slits. To address this limitation and
enable this decoupling method to be used for MRI, we propose a
Tunable-
Defected-
Ground-
Structure (TDGS) using
lumped elements, and demonstrate its capability to decouple monopole arrays at 7T.
Methods
Fig. 1A shows the diagram of traditional DGS using a
1/4 λ slit. With the combination of
distributed capacitance and inductance, the DGS can be thought of as a bandstop
filter and this effectively reduces current flow between the monopoles. Fig.
1B
shows the proposed TDGS method using a small slot and a capacitor (Cd) in series. This configuration is more suitable for MRI since it
provides tuning capability without compromising the ground plane.
To test the feasibility of the proposed
method, we comparatively investigated the coil performance of 2-ch monopole
arrays without and with TDGS. The width and length of each monopole element were
1 cm and 25 cm, respectively. The distance of the two monopole elements was about
8 cm. Simulations were performed using HFSS and Designer (ANSYS, Canonsburg, PA, USA). Based on
the simulation results, we also built a 2-ch TDGS-decoupled monopole array. This
array was used for both transmission and reception, and was matched to 50 Ω
and tuned to 298 MHz. Imaging was performed on
a human 7T Philips Achieva scanner (Philips Healthcare, Cleveland, Ohio, USA).
Simulation Results
Figs. 2A and 2B
show the simulated S
11 and
S
21 plots of two coupled
monopoles and two TDGS-decoupled monopoles,
respectively. As expected, S
21
is only -7 dB when no decoupling treatment was used; but it was reduced to -28 dB by using TDGS. Figs.
2C
and 2D
show the current distribution over the surface of
monopoles and ground, when the left element was driven with 1 W power, and the
right monopole was terminated to 50 Ω. Similar to the S-parameter
results, current flows on the right monopole become negligible when TDGS was utilized.
Instead, a large portion of surface current is trapped by the
capacitor-terminated slot, which is consistent with the bandstop filtering characteristic of L/C tank circuits.
Experimental
Results
Fig. 3A shows the constructed 2-ch monopole array with
TDGS. Fig. 3B
shows measured S-parameter plots. Similar to the simulation results, the S
21 was only about -25 dB and while S
11
was -28 dB. Fig. 3C shows gradient echo images of each channel on a cylindrical water phantom
(16 cm in diameter, 20 cm in height). The MR images were acquired with the
following parameters: FA = 25
0, TR/TE= 100/10 ms, FOV=180
Χ 180 mm
2, Voxel size= 0.94 Χ 0.94 Χ 2 mm
3. During the MRI experiments, the phantom was placed
very close to the monopole elements (distance
<1 cm). The
MR images show that the original individual
monopole image intensity profiles
are preserved using TDGS, reflecting that the TDGS method does not significantly
disturb the single-element EM fields.
This is different from the results with previous
ICE-decoupled monopole array
5.
Conclusion
We
proposed a Tunable-Defected-Ground-Structure
(TDGS) decoupling method for monopole arrays. The proposed concept was successfully
validated by EM simulation, bench test and MR experiments. By using the TDGS
method, the cross-talk between two closely-spaced monopoles was reduced from -7
dB to -25 dB. Based on the acquired MR images, it was found that the TDGS
method had little effect on the
original B
1 fields of the
individual monopole elements.
Acknowledgements
This
work was supported by NIH R01 EB016695.References
[1] A.J. Raaijmakers, et al, MRM.
66: 1488-1497 (2011). [2] G. C. Wiggins, et al, ISMRM. 541 (2012). [3] S. M.
Hong, et al. MRM, 71:1944-1052 (2014). [4] J. C. Coetzee, et al, IEEE TAP. 56: 1587-1593
(2008). [5] X. Yan, et al. QIMS. 4:79-86 (2014). [6] C. Chiu, et al, IEEE TAP.
55: 1732-1737 (2007).