(PDF) Lumped

TL;DR: This study replaced the dielectric substrate with discrete capacitors, which resulted in both SNR improvement and a tunable lumped‐element PSA (LPSA) whose dimensions can be optimized within broad constraints, for a given region of interest (ROI) and MRI frequency.

Abstract: The strip array (1) has a number of advantages over conventional loop-resonator arrays (2) in high-field and parallel MRI. For instance, it allows a large number of non-adjacent elements to simultaneously receive MRI signals, with minimal coupling between elements (3). Compared with a loop array, the more explicit spatial information in the phase of the signals from a strip array makes parallel reconstruction less susceptible to aliasing artifacts. Also, it enables phased-array applications in open-geometry magnets, where conventional loop arrays are limited by field orientation. The basic element of a planar strip array (PSA) is a microstrip with both substrate and superstrate (the electrical length of which is either π/2 or π) terminated in either an open circuit or a short circuit (1). Its geometric length must be a quarter or half of the resonant wavelength λ of the electromagnetic (EM) field at the MRI frequency, or integer multiples thereof. Under these conditions, the array has the unique advantage that the PSA elements are intrinsically isolated from each other. However, at commonly used fields of 1.5T (63.87 MHz), λ is around 4.7 m (if the substrate is air), so a PSA with λ/4 and λ/2 would generally be too long. The use of high-dielectric-constant substrates, instead of air, allows some reduction of λ to lengths practical for human imaging. However, the dielectric constant of a substrate is in general not easily varied, nor are there suitable materials available that possess a continuum of dielectric constants from which to choose. This limits our ability to arbitrarily adjust the PSA geometry in order to optimize its MRI performance in applications to a particular organ at a particular field strength. Nevertheless, the PSA elements are intrinsically isolated from one other, as long as their lengths are integer multiples of a quarter wavelength. Another disadvantage of the original PSA is that losses associated with the high-dielectric-constant substrate and large size of the distributed element can reduce the signal-to-noise ratio (SNR). If instead of using a high-dielectric constant substrate to reduce the wavelength, one creates an artificial or “reduced-length” transmission line (RTL) (4) by deploying two or more discrete high-Q shunted capacitors along the strip, two advantages can be realized. First, the discrete capacitors reduce the electrical length of the PSA transmission line strips, and potentially enable the design of PSAs with much smaller geometric dimensions that can be adjusted for a given field strength. Second, the losses associated with the dielectric substrate and superstrate can be reduced, thereby increasing the SNR. Moreover, the geometric configuration can be adjusted to maximize the SNR for each specific application without being bound by the λ/4 criterion as applied to the physical strip length. In this study we developed a lumped-element PSA (LPSA), employing RTLs for the array elements (5,6). In the LPSA, the electrical function of the substrate is replaced in part or substantially with two or more distributed shunt capacitors, yielding the equivalent electrical length of π/2 at 63.87 MHz with a much shorter strip length. At this frequency, an LPSA with a strip length of about 30 cm can be tuned with just two shunted 100 –200pF capacitors, while it can also be tuned with more uniformly distributed capacitors. A benefit of using only two capacitors in this situation (1.5T) is that the LPSA maintains a relatively homogeneous B1 field while it is tuned to an electrical length of π/2. The interstrip decoupling mechanisms of the LPSA are different from those of the PSA. In the PSA, because of its quasi-transverse EM (TEM) field distribution, contributions from incident and reflected waves along a λ/4 strip cancel one another (1), so that the PSA is inherently decoupled. However, in the LPSA, at 1.5T the physical length of each strip can be much shorter than λ/4, and the condition of intrinsic decoupling is difficult to achieve. In this study, we first apply a lumped-element circuit model to analyze the resonance condition and Q-factor of an RTL, and calculate its field and SNR. Second, interelement coupling, sensitivity profiles, and g-factor maps of the LPSA are evaluated in detail. Third, we describe two different schemes to isolate the RTL strips and substantially eliminate mutual coupling. In the past, shared (7) and interconnected (8) capacitors have been used to decouple loop MRI coils, as well as to couple transmission line sections (9), and in the first scheme we use interconnecting capacitors to decouple each nearest-neighbor pair. Low-input impedance preamplifiers are then deployed to decouple the remaining strips, as was implemented in the original phased array (2,10). The other scheme is to make the ratio of the strip spacing to the strip-to-ground distance large enough to achieve isolation (6). This approach limits the minimum separation between neighboring strips. Fourth, quantitative measurements of array characteristics and MRI experiments are used to demonstrate that the LPSA can produce high-quality in vivo MR images at 1.5 T with either phased-array or parallel data acquisition.