High-sensitivity 2H NMR in solids by 1H detection.

The 2H quadrupolar coupling is an excellent nuclear magnetic resonance (NMR) probe of segmental orientation, molecular dynamics, or hydrogen-bonding in solids and liquid crystals.1,2 It has provided information on protein dynamics3 and structure,4 on motions in polymers,5 on bond-order parameters in liquid crystals,6 on dynamics of guest molecules in zeolites,7 clays,8 and inclusion compounds,9 and on hydrogen-bond lengths.2,10 However, due to large line widths and the relatively small magnetic moment of 2H (15% of that of 1H), the sensitivity of traditional 2H NMR in solids is relatively low, except for the special case of CH3 groups. We report here a novel method, proton inverse-detected deuteron (PRIDE) NMR, that can provide sensitivity enhancement in 2H NMR by an order of magnitude or more. It can be regarded as the solid-state NMR analogue of “inverse-detection” schemes in solution NMR,11,12 and recently demonstrated in fast-magicangle-spinning solid-state NMR.13 In these experiments, the spectra of low-sensitivity nuclei are detected indirectly, in the first dimension of a two-dimensional spectrum, via the modulation of the strong 1H signals. The sensitivity gain in the PRIDE NMR experiment comes primarily from the sensitivity of 1H detection, which is higher than that of 2H ()D) by a factor of (γH/γD) ) 108, where γH and γD are the gyromagnetic ratios of 1H and 2H, respectively.14 Other factors also contribute to sensitivity enhancement, in particular 1H line-narrowing by a solid-echo train (pulsed spinlock) used during detection.15 Due to the longer persistence of the 1H signal under the spinlock, where it decays with a time constant T1F,H, the line width ∆ν1H ∼ 1/T1F,H in the 1H spectrum is decreased, and the signal height increased. Combined with the indirect (two-dimensional) detection scheme, this results in a signal-to-noise gain of (T1F,H/τD)(tdet/tdw,2), where τD is the maximum acquisition time of the 2H time signal,13 tdw,2 is the dwell time between acquired points, and tdet is the detection window length. Typically, the sensitivity gain from this factor, relative to 2H detection, is ∼2-10. Compared to standard 1H wide-line detection, a more than 5-fold sensitivity enhancement is achieved. The larger electronic quality factor of the 1H resonance circuit also increases the relative sensitivity of the PRIDE experiment (∼2-fold). Other effects decrease it, e.g. the ratio of the spin-dependent prefactors in the magnetization expression,16 I(I + 1)/[S(S + 1)] ) /8, with the spin quantum numbers I ) /2 for 1H and S ) 1 for 2H. The detection efficiency per 2H, which combines several factors, was measured to be 0.7. The use of a double-resonance probe results in an estimated sensitivity reduction by 0.5-0.2. The combined factors of ∼600 * 0.08 ) 48 will still result in a large sensitivity enhancement. Instead of double cross polarization13 with its high 2H radiofrequency power requirements,17 we use a heteronuclear multiplequantum coherence (HMQC) approach11,12 with only two or three 2H pulses of ∼10 μs total duration.18 The pulse sequence is shown in Figure 1. After the initial 1H 90°-pulse, the 1H-2H dipolar coupling generates 1H-2H coherence. During this period τHD, 1H homonuclear decoupling must be applied; for simplicity, the MREV-8 cycle19 was used. Then, a “magic sandwich” consisting of four pulses20 is applied, so that 1H homonuclear dipolar evolution in the following long window is refocused into a magicsandwich echo (MSE).20 A 1H 180° pulse at the center of the window refocuses the chemical-shift evolution for the entire sequence until the start of signal detection. A 2H 90° pulse near the start of the window makes the heteronuclear coherence transverse in the 2H term, which is then modulated by the 2H quadrupolar coupling. The duration of the quadrupolar evolution time t1 and the lengths of the two long inner pulses of the magicsandwich are incremented synchronously to fulfill the MSE condition.20 The quadrupolar evolution is terminated by a 90° 2H pulse that makes the coherence again longitudinal on 2H. A third 2H pulse, shown dashed in Figure 1, partially compensates for the finite duration of the two other 2H pulses, by effectively creating a solid echo. This composite-pulse21 scheme provides efficient 2H excitation. Then MREV-8 decoupling is resumed, and the 1H-2H dipolar coupling converts the heteronuclear coherence modulated by the 2H quadrupolar coupling back into observable 1H magnetization.