There are a variety of reasons for defective baselines in NMR spectra, often unavoidable due to the nature of a particular experiment; for example pulse "break-through". Often a good baseline is obtained by the simple expedient of using a large sweep-width,1 however, in multidimensional experiments this is usually undesirable since the digital resolution required may lead to excessively large data sets. For some experiments, such as those involving spin-locking (e.g. ROESY),2 there may be no experimental way to prevent some baseline problem. This poster describes two new methods for the removal of baseline artifacts and compares these to literature methods for some "typical" cases.
