Cryogenic Single and Array Coils for Magnetic Resonance Systems

Abstract

The annual cost of cancer treatment in the United States of America and the European Union exceed 100 billion euro. Using a combination of magnetic resonance imaging (MRI) and dissolution dynamic nuclear polarization (dDNP) the potential worldwide savings are in the billions of euro annually. To make the techniques clinically viable an essential aspect is the design and implementation of radio frequency (RF) receive hardware optimized for maximal signal-to-noise ratio (SNR). This work investigates three primary topics within receiver hardware for MRI systems utilizing dDNP of 13C at 3 T: Preamplifiers, volume coils, and array coils. Preamplifiers for MRI arrays require a high input reflection coefficient, to enable decoupling of neighbouring coils, while exhibiting low noise figure. In this thesis a design procedure is presented enabling the implementation of ideal preamplifiers for MRI arrays by using the inherent feedback of the transistor. This causes the input impedance of the preamplifier to depend on the output matching circuit while the noise figure remains constant due to the high gain of the transistor. A procedure for designing cryogenic preamplifiers is presented. It is shown theoretically that a negative input impedance amplifier can be used for ideal coil decoupling. In practice, 50 dB decoupling was achieved using a cryogenic preamplifier design cooled with liquid nitrogen to 77 K with a 0.05 dB noise figure having an input impedance of −8+j533 Ω. Cryogenic volume coils are generally not viable candidates for human imaging due to high sample loading, which cannot be mitigated by cooling the coil. This is not necessarily the case for volume coils for small animals. Hence,sensitivity improvement of a birdcage coil for small animal imaging using cryogenic cooling is investigated. The implemented birdcage coil has a bore size of 50 mm with a length of 100 mm and is cooled to 77 K using liquid nitrogen. A dedicated, low cost, cryostat was also developed. The measured unloaded and loaded Q-factors of the cryogenic birdcage are 627 and 616, respectively. Using conventional formulas for estimating the SNR gain between the room temperature and the cryogenic birdcage coil results in an estimated SNR gain of approximately 2.5 times. However, the conventional analysis does not take into account the room temperature RF front end. An extended analysis is thus presented that takes into account the temperatures of the coil and the RF front end connected to the coil (hybrid coupler, transmit/receive switch, preamplifier). Thus, taking in to account the influence of the room temperature RF frontend, the expected SNR gain is 2 times. If instead a cryogenic RF front end is used the expected SNR gain is 2.4 times. Hence, it is vital when using cryogenic coils to also use a cryogenic RF front end. Controlling the dedicated transmit/receive switch and Q-spoiling circuits used for both the volume and array coils is achieved by a custom PIN diode driver, which is also detailed in this work. The PIN diode driver switches from the transmit to receive state in approximately 0.4 µs and from receive to transmit in under 2 µs. Further, the PIN diode driver supplies a constant current, regardless of characteristics of the PIN diode(s), in the receive state. While in the transmit state a negative voltage of -5 V is applied. To enable larger field-of-views and accelerated/parallel imaging an array of loop coils is employed. In this thesis, a 32 channel human brain coilis designed and implemented for clinical imaging focussing on the application of parallel imaging to decrease acquisition time of images. The performance is measured in the scanner versus a birdcage coil and shows an approximate SNR decrease in the center of a head phantom by approximately 48 %. The problem is noise coupling when using non-overlapped neighbouring elements and conventional 50 Ω noise figure optimized preamplifiers. By noise matching to a complex impedance, rather than the conventional 50 Ω, the noise coupled between non-overlapped coils can be decreased by approximately 50 %. Hence, using the newly proposed preamplifier design yields an SNR impairment of 19 %. This is, however, in the center of the phantom and closer to the surface a significant SNR increase is present. Further, the array coil enables parallel imaging, which is impossible with the birdcage coil. Especially for arrays, the SNR impairment caused by the preamplifiers due to noise coupling is not dominated by the noise figure, but rather the corresponding noise and current voltages and their correlation. Looking into the future, this work enables optimal preamplifiers for single and array coils for both room temperature and cryogenic operation.