Introduction

Development of cryogen-free low-field magnetic resonance (MR) systems can address accessibility issues often associated with MRI. These scanners, if built with high-performance gradients, can overcome the main field deficit and generate high quality images comparable to high-field scanners [1]. These systems rely on a cold head to draw heat away from the superconducting coils, instead of the coils being directly immersed in the cryogen. This improves accessibility because little or no cryogen is used, and due to the lower field of the system, some expensive specialized siting requirements are relaxed. However, the absence of a large heat sink of liquid helium could result in a greater variation in the possible temperatures of the coils when the magnet is subjected to gradient-induced eddy current heating, as the ability to regulate the temperature is in general more limited compared with traditional cryogen systems.
Gradient-induced eddy current heating is a source of potential interference to the main magnet system, and this is especially relevant for high-strength or asymmetric geometry gradient systems where unwanted field may ‘leak’ beyond the gradient into the magnet compared with traditional whole-body gradient coils (recent research-oriented gradient systems designed for imaging the head can reach up to 200 mT/m strength [2]). While gradient-induced eddy current fields can be corrected by pre-emphasis or post-processing, an important concern is heating of the magnet during imaging sequences, which could cause field stability problems if not appropriately characterized or mitigated, or worse the magnet may heat enough to quench. Other investigations of the thermal characteristics of cryogen-free MRI magnets have shown how main field stability can be affected by magnet temperature [3]. In this study, we explore the effect of gradient coil operation on the main magnet, starting with the effect on the temperature of the coils.

Methods

The temperature dynamics were investigated using set of 7 temperature probes placed throughout the main and shield coils of a compact cryogen-free superconducting magnet located in our labs designed to operate at a field of 0.5T. The temperature probe voltage outputs were exported using an NI DAQ system, using a LabVIEW program to collect data for 400 seconds. A prototype asymmetric gradient coil designed for head-only imaging with this magnet was connected to a gradient amplifier capable of 900A peak current. Two gradient pulse sequences were tested using the different gradient components: A single long trapezoidal pulse (similar to a diffusion gradient) and a series of bipolar trapezoidal lobes (similar to an EPI train) – all of which were meant to test the limits of the gradient system and give “worst case” heating effects. Further details on the gradient parameters tested are shown in Figures 1 & 3. The maximum gradient amplitudes ranged between 65 mT/m and 130 mT/m. Initially, a comparison of heating induced by the different gradient axes was measured to determine which had the largest heating effect by comparing heating after 400s of gradient operation. Using these results, the axis producing the strongest interactions (x-axis) was compared to itself at different gradient strengths. The analysis of the temperature shift was done using MATLAB and Python to optimize a fit using a A(1-e^x) shape to model the effect of varying gradient strength on the temperature of the main magnet over the duration of the gradient operation. The heating model was used to describe the predicted settling time (time until the temperature reached equilibrium).

Results

The magnet temperature dynamics for repeated diffusion gradient pulses along with the fit performed are shown in Figure 1. In Figure 2, the exponential fit parameters used to fit the Figure 1 data are described with the key term being the time constant which allows us to estimate the settling time until an equilibrium temperature is reached in the MR system. Further, Figure 3 shows the main magnet heating from the x-, y-, and z-axis gradients running an EPI style pulse sequence.

Discussion

This preliminary study has investigated the effect of gradient operation on magnet temperature over the course of a typical scan duration of five minutes. The induced heating in all cases was observed to be <0.015 K, making it unlikely that the gradient operation could induce a quench by heating the magnet. Although this induced heating is not likely to be a destabilizing factor in the operation of the main cryogen free magnet system, the effect of induced heating is still an important factor to consider in gradient design. The results shown in Figure 3 suggests that the axis that interacted the most with the main magnet was the x-axis. This may be due to the geometry of the gradient, the design of the main magnet itself causing resonances, or less effective shielding on this component of the gradient. Interaction effects seen in the x axis coil must be minimized and in future gradient-induced eddy current fields must be measured and characterized on this system.

Conclusion

The interactions between trapezoidal pulses run on the different gradient axes and the main magnet were investigated for a compact cryogen-free magnet coupled with an asymmetric head-only gradient coil, with the resultant temperature change in the main magnet coils being quantified and guiding further study.

Acknowledgements

The authors would like to acknowledge support form NSERC and the Ontario Research Fund.