The effect of coil geometry on the RF heating of saline phantoms: applications to in vivo NMR

Publisher Website
Google Scholar
Abstract
The lumped impedance model of RF losses in a weakly conducting sample, originally proposed by Gadian and Robinson, has been reformulated to show that the losses can be characterised by three parameters which depend on the geometry of the coil and the sample. Two of these parameters, K1 and K2, are particularly important since they represent the eddy current losses and the losses brought about by the conservative electric field in the high-frequency or low-conductivity limit, respectively. Both the verification of the predicted frequency dependence of the losses and the values of the parameters were obtained from experiments conducted between 18 and 100 MHz on a series of saline samples ranging in concentration from 7 mM to 3.5 M and irradiated by means of coils of approximately 2.5 cm in diameter. Solenoidal and saddle coils were investigated, one of the saddle coils being studied in its series and parallel configurations. A theoretical model is presented for estimating K1 and gives values which are in satisfactory agreement with experiment for each coil. A series-to-parallel comparison of the values of K2 and K3 for the same saddle soil shows that the currents caused by the conservative electric field are primarily at right angles to the coil axis in the case of the series saddle. Our experimentally measured parameters are used to estimate the frequency dependence of an upper limit for the rise in the average temperature of a sample of conductivity 0.6 S m-1 (typical of biological tissue) in a representative NMR imaging experiment and at sizes appropriate for animal work. It is found that the heating due to the conservative field is a factor of four larger when using a series saddle coil than when using the same coil in its parallel configuration.

This publication has 15 references indexed in Scilit: