Ultrafast Laplace NMR to study metal–ligand interactions in reversible polarisation transfer from parahydrogen

Laplace Nuclear Magnetic Resonance (NMR) can determine relaxation parameters and diffusion constants, giving valuable information about molecular structure and dynamics. Information about relaxation times (T1 and T2) and the self-diffusion coefficient (D) can be extracted from exponentially decaying NMR signals by performing a Laplace transform, which is a different approach to traditional NMR involving Fourier transform of a free induction decay. Ultrafast Laplace NMR uses spatial encoding to collect the entire data set in just a single scan which provides orders of magnitude time savings. In this work we use ultrafast Laplace NMR D–T2 correlation sequences to measure key relaxation (T2) and diffusion (D) parameters of methanolic solutions containing pyridine. For the first time we combine this technique with the hyperpolarisation technique Signal Amplification By Reversible Exchange (SABRE), which employs an iridium catalyst to reversibly transfer polarisation from parahydrogen, to boost the 1H NMR signals of pyridine by up to 300-fold. We demonstrate use of ultrafast Laplace NMR to monitor changes in pyridine T2 and D associated with ligation to the iridium SABRE catalyst and kinetic isotope exchange reactions. The combined 1440-fold reduction in experiment time and 300-fold 1H NMR signal enhancement allow the determination of pyridine D coefficients and T2 values at 25 mM concentrations in just 3 seconds using SABRE hyperpolarised ultrafast Laplace NMR.


S5.2 Reproducibility of SABRE HP ultrafast Laplace NMR D-T2
S6. References S1. Using ultrafast Laplace NMR to measure D and T 2 of concentrated thermally polarised pyridine A solution of pyridine (5 M) in methanol-d4 (0.6 mL) was used for determination of D and T2 using ultrafast LNMR D-T2 pulse sequences. D-T2 correlation spectra are shown in Figure S1a. Analogous measurements on a reference sample of methanol-d4 (0.6 mL) are shown in Figure S1b. A D-T2 correlation spectrum for a sample containing pyridine (5 M) in methanol-d4 (0.6 mL) recorded using D-T2 UF LNMR sequences with a PROJECT loop is shown in Figure S1c. Overlaid D-T2 spectra are shown in Figure S1d. S2. Using ultrafast Laplace NMR to measure D and T 2 of dilute solutions of thermally polarised pyridine A solution of pyridine (25 mM) in methanol-d4 (0.6 mL) was used for determination of D and T2 using ultrafast LNMR D-T2 pulse sequences. D-T2 correlation spectra are shown in Figure S2a. A D-T2 correlation spectrum recorded using D-T2 UF LNMR sequences with a PROJECT loop is shown in Figure S2b. Overlaid D-T2 spectra are shown in Figure S2c. This figure also contains an overlap with analogous measurements on a reference sample of methanol-d4 (0.6 mL), the which are shown in Figure S1b (with a CPMG loop).

S2.1 Effect of number of scans on ultrafast Laplace NMR D-T2
The effect of the number of scans on D-T2 correlation spectra for a solution of pyridine (25 mM) in methanol-d4 (0.6 mL) was determined by using ultrafast LNMR D-T2 pulse sequences with 1, 4, 32 and 256 scans. In D-T2 correlation spectra only the OH signal from HOD / HOCD3 can be resolved using 1 scan ( Figure S3a). When 4 scans are used the minor pyridine component can now be resolved, but its D and T2 values are unreliable ( Figure S3b). 32 scans are sufficient to generate reliable data ( Figure S3c) which is improved when greater number of scans are used ( Figure S3d).

S2.2 Effect of excitation frequency on ultrafast Laplace NMR D-T2
The effect of the central excitation frequency (o1p) on D-T2 correlation spectra for a solution of pyridine (25 mM) in methanol-d4 (0.6 mL) was determined by using ultrafast LNMR D-T2 pulse sequences with a varying o1p parameter. In all experiments presented in this work the o1p was set to 6.378 ppm. We expect the excitation pulses to excite all protons within the δ = 0 to 10 ppm range. Nevertheless, we investigated the effect of changing o1p on the experimental results. D-T2 correlation spectra recorded with an o1p set at 1.000 ppm (noise) and 8.570 ppm (pyridine ortho resonance) are shown in Figure S4a and S4b respectively. Figure S4c shows that when overlapped, spectra appear similar, but D and T2 values that are extracted can change depending on the o1p value used. The effect of changing o1p on D-T2 correlation spectra collected using a PROJECT loop in the pulse sequence (instead of a CPMG loop) is shown in Figure S5 and appears to be less influenced by the o1p value.

S3. Effect of J-modulation on 1D CPMG data
Ultrafast D-T2 sequences recorded using a CPMG loop can give rise to J-modulated artefacts (e.g. Figure S1a). Related effects can also be observed in 1D CPMG data recorded using a short echo time (4 ms) ( Figure S6). These effects can be reduced by using longer echo times (15 ms) ( Figure S7). We confirm that both 1D T2 data recorded using a CPMG or a PROJECT sequence with a 15 ms echo time (30 ms double spin echo time in the case of the PROJECT sequence) are consistent ( Figure S7). We note that for this T2 data recorded using a 15 ms echo time the pyridine meta site still appears to show the effects of J-modulation and this is likely responsible for the larger differences observed between PROJECT and CPMG sequences for this resonance ( Figure S7c). While it has been reported that the effect of J-modulation is generally more pronounced when longer echo times are used, [1] this was not visible when 1D CPMG data was collected using a 40 ms echo time ( Figure S8). Here, experiments with longer echo times record data points at sufficient time spacings that the majority of 1D CPMG data recorded with 40 ms echo times produce signal decay plots that do not exhibit J-modulated oscillations. They can therefore be fit to a mono-exponential decay function to yield a reliable T2. Any data containing oscillations were discarded and a T2 value was not extracted.

S4. Formation of SABRE catalysts examined using thermally polarised ultrafast Laplace NMR
Hydrogen gas was produced from the electrolysis of water using a desktop hydrogen generator (F-DGSi, Evry, France). This was used directly to make parahydrogen (pH2) using a BPHG 90 parahydrogen generator (Bruker) which passes the hydrogen gas over a spin-exchange catalyst at low temperature. The generator operates at ca 38 K and produces a constant flow of pH2 with ca 92% purity. Parahydrogen (3 bar) was added to NMR tubes containing a quick pressure valve using a home-built system shown in Figure  S9. This set-up contains 3 valves that allow the system to be opened to A) the NMR tube B) parahydrogen and C) a vacuum pump. This system was used to degas the sample. Samples were left to react with H2 (3 bar) for a period of several hours (usually around 16 hours overnight). Activation is usually indicated by a change in colour from orange to pale orange. Catalyst activation is indicated in 1 H NMR spectra by the formation of a peak corresponding to [Ir(H)2(IMes)(pyridine)3]Cl at δ= −22.74 ppm. [2] As catalyst activation proceeds, 1 H NMR signals for the COD ligand at δ = 3 and 4 ppm eventually disappear and a signal for cyclooctane at ca δ = 1.6 ppm becomes visible. [2][3][4]

Figure S9: Picture of a) hydrogen generator b) parahydrogen generator and c) home-built device for addition of parahydrogen to NMR tubes in which A, B and C refer to the valves that allow opening of the system to the NMR tube, parahydrogen line or vacuum pump respectively.
A sample of [IrCl(COD)(IMes)] (5 mM) and pyridine (5 equiv.) were dissolved in methanol-d4 (0.6 mL). This solution was then degassed and activated with 3 bar pH2 overnight to form the SABRE-active [Ir(H)2(IMes)(pyridine)3]Cl. During this process, T2 and D values were recorded using both 1D CPMG and DOSY sequences and UF LNMR (D-T2 containing a CPMG loop). In some of these D-T2 correlation plots a third signal may be discernible, in addition to those assigned as OH and pyridine. A representative spectrum containing this artefact peak with the largest intensity is shown in Figure S10a. This additional signal does not appear in measurements performed without the metal catalyst. This peak is often poorly resolved by the Laplace transform and corresponds to a diffusion coefficient of ca 7 × 10 -9 m 2 s -1 and T2 of ca 150 ms which do not show a consistent change over the activation period ( Figure S10b-c). This larger D relative to pyridine and solvent makes this signal unlikely to correspond to species such as [Ir(H)2(IMes)(pyridine)3]Cl, which has a diffusion constant of ca 6 × 10 -10 m 2 s -1 according to 1D DOSY measurements. Assignment due to other molecules present in these mixtures, such as cyclooctane, which is formed during catalyst activation, [3,5] is also unlikely as 1D DOSY and CPMG measurements reveal D and T2 values of 1.5 × 10 -9 m 2 s -1 and ca 3 s respectively for this molecule in these mixtures. T2 values for [Ir(H)2(IMes)(pyridine)3]Cl can not be determined using 1D CPMG sequences due to low signal intensity and peak overlap. It is most likely that this additional signal is an artefact that results from spectral noise or homonuclear J-coupling. [6][7][8]

S5. Improving MR sensitivity of UF LNMR by using SABRE hyperpolarisation
Samples containing the active [Ir(H)2(IMes)(pyridine)3]Cl SABRE catalyst were prepared by activating a solution containing [IrCl(COD)(IMes)] (5 mM) and pyridine (5 equiv.) in methanol-d4 (0.6 mL) with H2 (3 bar) overnight. This sample was then shaken manually with fresh 3 bar pH2 for 10 seconds in a 65 G magnetic field before insertion into the 9.4 T spectrometer and collection of D-T2 excitation-detection profiles. The shake and drop method was employed for recording hyperpolarised NMR spectra. This involves filling NMR tubes with fresh pH2 (3 bar) as described in section S4 before shaking them vigorously for 10 seconds in a 6.5 mT (65 G) magnetic field. [2] We find that the stray field of our shielded 9.4 T magnet is no larger than 2 mT, therefore we use an electromagnetic coil powered by a Blanko PS-3005 0-30V 0-5A switching power supply to provide the necessary magnetic fields for SABRE polarisation transfer. The current and voltage of the power supply can be altered to achieve a magnetic field inside the coil of ca 6.5 mT. Magnetic fields were measured using a Hirst GM04 Gaussmeter and we estimate that the sample experiences a magnetic field of 6.5 ± 1.0 mT during the manual shaking process. This set up is shown in Figure S11 and is placed as close to the NMR spectrometer as possible to reduce transfer time after the 10 second shaking period. The sample is inserted into the spectrometer as rapidly as possible, this is facilitated by pre-emptively turning off the lift function on the spectrometer. 1 H NMR pulse sequences are modified to include an autosuspend function such that radiofrequency excitation occurs immediately upon sample insertion.

S5.1 Effect of delayed acquisition on SABRE HP ultrafast Laplace NMR D-T2
D-T2 Laplace NMR sequences containing a CPMG loop were used with a variable delay between insertion of the HP sample into the magnet and spectral acquisition. This delay was introduced to examine the effect of potential movement of the solution on the appearance of these spectra. (For example, at short delay times the solution may still be moving and this may influence spatia l and/or diffusion encoding). The effect of this time delay is shown in Figure S 12. There appears no significant difference between spectra collected with a 0, 1 or 2 second time delay, although spectral clarity does improve when a 3 second time interval is used. Generally, the intensity of the signal excitation detection profile appears to decrease as the delay time is increased which is consiste nt with a lower hyperpolarised 1 H NMR signal intensity due to greater relaxation as the delay is increased.

S5.2 Reproducibility of SABRE HP ultrafast Laplace NMR D-T2
The reproducibility of SABRE HP D-T2 correlation spectra was investigated by repeating the shaking process with fresh pH2 shaking on the same sample under the same conditions. These give similar D-T2 correlation spectra ( Figure S13) although the intensity of the excitation detection profile can change more significantly (ca 50 % difference) which is likely related to variation in 1 H NMR signal enhancements, although generally variation in the efficiency of SABRE polarisation transfer can be reduced by contr olling more precisely the polarisation transfer field [9] or automated bubbling systems compared to manual shaking. [10][11][12]
The reproducibility of SABRE HP D-T2 correlation spectra collected using a PROJECT loop was also investigated by repeating the shaking process with fresh pH2 under the same conditions using the same sample. This yields consistent D-T2 correlation spectra ( Figure S14). The reproducibility of SABRE HP D-T2 (CPMG loop) correlation spectra collected using two separate samples both containing the same amount of SABRE precatalysts and substrate, hyperpolarised in the same manner and recorded using the same spectral acquisition parameters was also investigated and found to yield consistent results (Figure S1 5).