Transverse Dephasing Optimised NMR Spectroscopy in Solids: Natural-Abundance 13 C Correlation Spectra GaèlDePaèpe, [a] Anne Lesage, [a] Stefan Steuernagel, [b] andLyndonEmsley* [a] Introduction Solid-state NMR is a central technique for the characterisation of molecular compounds or materials. [1] Notably, it plays a key role in the structural characterisation of solids that cannot be analysed by X-ray methods like glasses, heterogeneous cata- lysts, amorphous polymers, or plaque-forming proteins. For or- ganic compounds, establishing the connectivity of the carbon skeletonisoftenthefirstandmostimportantstepinthestruc- tural characterisation process. This is usually done by applying carbon correlation spectroscopy that relies either on the dipo- lar couplings to transfer magnetisation between neighbouring nuclei [2] or on the scalar couplings to establish connectivities between bonded carbon atoms. [3] The latter approach offers the advantage of yielding through-bond correlations, and therefore leads to the straightforward assignment of signals from 13 Cresonances. INADEQUATE has for many years been established as a method of choice for determining carbon±carbon connectivi- ties in liquids. [4] A few years ago, we demonstrated how these experiments could be successfully transferred to rigid organic solids in either crystalline [5] or disordered [6] compounds, by capitalising on earlier work on mobile systems [7] and inorganic materials. [8] Together with UC2QF-COSY, [9] also known as SUPER-COSY, [10] a closely related method, INADEQUATE experi- ments have since been demonstrated to be robust methods for establishing direct scalar carbon±carbon connectivities in isotopically enriched organic samples under magic-angle spin- ning (MAS) and have been applied to characterise the struc- ture of systems as complex as the polymer phase of CsC 60 , for example, [11] to detect hydrogen bonds in solids through 15 N± 15 N connectivities, [12] or to establish high-resolution correlations in disordered solids such as cellulose. [13] By way of illustration, Figure1b shows the two-dimensional (2D) 13 C INADEQUATE spectrum of fully 13 C-labelled l-tyrosine. Gradients were ap- plied in this experiment to select the desired coherence path- ways, so that this spectrum recorded with only around 10 mg of sample could be acquired in an extremely short experimen- tal time (13 minutes). As indicated in the figure, the assign- ment of the one-dimensional (1D) carbon spectrum is straight- forward from this 2D map because two bonded nuclei share a common frequency in the double quantum dimension at the sum of their individual frequencies. Applying through-bond correlation techniques on natural-abundance systems is, how- ever, a much more challenging problem because the sensitivity ofthe2Dexperimentsdropsbyafactor10 4 duetothe1%iso- tope concentration of spins. As a consequence, through-bond correlation techniques have so far been very rarely applied to rigid samples at natural abundance and were shown to require either large ( 250mg) sample volumes [5] or long ( 2weeks) experimental times. [14] We have recently shown [15] that heteronuclear decoupling sequences could be specifically optimised to provide long co- herence lifetimes by using a direct spectral optimisation proce- dure. [16] By using this approach, improvements in sensitivity up to a factor of five have been reported [15] for refocused INADE- QUATE experiments on fully labelled systems, compared to normal two-pulse phase modulation (TPPM) [17] or continuous- wave (CW) decoupling. Herein, we demonstrate experimentally that by using the transverse dephasing optimised spectrosco- py (TDOP) approach refocused INADEQUATE spectra can be re- corded on medium-sized molecules at natural abundance in reasonableexperimentaltimes(80mg/4days). Results and Discussion Optimisation of the Coherence Lifetimes As discussed in previous papers, [6,15] the efficiency of INADE- QUATE-type experiments depends mainly on the decay of the 13 Ctransverse magnetisation during the t±p±t blocks (the pulse sequences for the normal and refocused INADEQUATE experiments are indicated in Figure1a and Figure2d and in- clude, respectively, one and two t±p±t blocks). During these delays, because the p pulses refocus all the linear spin interac- tions and in particular the chemical-shift distribution, the life- time of transverse coherences (or the so-called refocused line width) is strongly dependent on the efficiency of heteronuclear decoupling. [18] We have recently shown that eDROOPY decou- [a] G. De Paèpe, Dr. A. Lesage, Prof. L. Emsley Laboratoire de Chimie (UMR-5182 CNRS/ENS) Laboratoire de Recherche Conventionnÿ du CEA (23V) Ecole Normale Supÿrieure de Lyon 46 Allÿe d'Italie, 69364 Lyon (France) Fax:(+ 33)472-728-483 E-mail:lyndon.emsley@ens-lyon.fr [b] Dr. S. Steuernagel Bruker BioSpin GmbH Silberstreifen, 76287 Rheinstetten (Germany) ChemPhysChem 2004,5,869±875 DOI: 10.1002/cphc.200301062 ¹2004Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim 869 NMR SPECIAL