Frequency cycling for compensation of undesired off-resonance effects in surface nuclear magnetic resonance

Abstract

To produce reliable estimates of aquifer properties using surface nuclear magnetic resonance (NMR), an accurate forward model is required. The standard surface NMR forward model assumes that excitation occurs through a process called on-resonance excitation, which occurs when the transmit frequency is set to the Larmor frequency. However, this condition is often difficult to satisfy in practice due to the challenge of accurately determining the Larmor frequency within the entire volume of investigation. As such, in situations where an undesired offset is present between the assumed and true Larmor frequency, the accuracy of the forward model is degraded. This is because the undesired offset leads to a condition called off-resonance excitation, which impacts the signal amplitude, phase, and spatial distribution in the subsurface, subsequently reducing the accuracy of surface NMR estimated aquifer properties. Our aim was to reduce the impact of an undesired offset between the assumed and true Larmor frequency to ensure an accurate forward model in the presence of an uncertain Larmor frequency estimate. We have developed a methodology where data are collected using two different transmit frequencies, each an equal magnitude above and below the assumed Larmor frequency. These data are combined, through a method we refer to as frequency cycling, in a manner that allow the component well-described by our estimate of the Larmor frequency to be stacked coherently, whereas the component related to the presence of an undesired offset is combined destructively. In synthetic and field studies, we have determined that frequency cycling is able to mitigate the influence of an undesired offset providing more accurate estimates of aquifer properties. Furthermore, the frequency-cycling method stabilized the complex inversion of surface NMR data, allowing advantages associated with complex inversion to be exploited.