Search

Dynamic nuclear polarization can be used to hyperpolarize the bulk of proton-free inorganic materials in magic angle spinning NMR experiments. The hyperpolarization is generated on the surface of the material with incipient wetness impregnation and from there it is propagated towards the bulk through homonuclear spin diffusion between weakly magnetic nuclei. This method can provide significant gains in sensitivity for MAS NMR spectra of bulk inorganic compounds, but the pathways of the magnetization transfer into the material have not previously been elucidated. Here we show how two-dimensional experiments can be used to study spin diffusion from the surface of a material towards the bulk. We find that hyperpolarization can be efficiently relayed from surface sites to multiple bulk sites simultaneously, and that the bulk sites also engage in rapid polarization exchange between themselves. We also show evidence that the surface peaks can exchange polarization between different sites in cases of disorder.

It has recently been shown how dynamic nuclear polarization can be used to hyperpolarize the bulk of proton-free solids. This is achieved by generating the polarization in a wetting phase, transferring it to nuclei near the surface and relaying it towards the bulk through homonuclear spin diffusion between weakly magnetic nuclei. Pulse cooling is a strategy to achieve this that uses a multiple contact crosspolarization sequence for bulk hyperpolarization. Here, we show how to maximize sensitivity using the pulse cooling method by experimentally optimizing pulse parameters and delays on a sample of powdered SnO2. To maximize sensitivity we introduce an approach where the magic angle spinning rate is modulated during the experiment: the CP contacts are carried out at a slow spin rate to benefit from faster spin diffusion, and the spin rate is then accelerated before detection to improve line narrowing. This method can improve the sensitivity of pulse cooling for Sn-119 spectra of SnO2 by an additional factor of 3.5. (C) 2019 Elsevier Inc. All rights reserved.