Windowed cross polarization at 55 kHz magic-angle spinning

Cross polarization (CP) transfers via Hartmann-Hahn matching conditions are one of the cornerstones of solid-state magic-angle spinning NMR experiments. Here we investigate a windowed sequence for cross polarization (wCP) at 55 kHz magic-angle spinning, placing one window (and one pulse) per rotor period on one or both rf channels. The wCP sequence is known to have additional matching conditions. We observe a striking similarity between wCP and CP transfer conditions when considering the ﬂip angle of the pulse rather than the rf-ﬁeld strength applied during the pulse. Using ﬁctitious spin-1/2 formalism and average Hamiltonian theory, we derive an analytical approximation that matches these observed transfer conditions. We recorded data at spectrometers with different external magnetic ﬁelds up to 1200 MHz, for strong and weak heteronuclear dipolar couplings. These transfers, and even the selectivity of CP were again found to relate to ﬂip angle (average nutation).

Hartmann-Hahn conditions [1] for MAS NMR can be efficient for both zero-quantum (ZQ) and double-quantum (DQ) conditions [33][34][35][36].They occur, respectively, when the difference or the sum of the simultaneously applied nutation frequencies (rf-field strengths) matches the rotor frequency or twice the rotor frequency: where m I and m S are applied rf-field strengths on I and S spins, respectively; m R is a MAS rate and n ZQ ; n DQ =±1,±2.Eq. ( 1) describes the Hartmann-Hahn conditions for CP transfers between a pair of spins.Hartmann-Hahn conditions can also be simultaneously achieved for two different heteronuclear pairs.[37][38].
In 1993 Hediger et al. [70] investigated amplitude-modulated cross-polarization [71][72][73] (AMCP) schemes for improvement of polarization transfer characteristics for MAS rates of 5 to 7 kHz.Using a sample of adamantane, it was shown that the AMCP sequences can help broaden CP conditions.One of the schemes, S-AMCP, initially proposed by Barbara and Williams [72], con-tained a CW-spin lock pulse on one of the channels and rotor synchronized windowed rf-field on the other.At 5-7 kHz, this windowed sequence (wCP) resulted in improved transfer at n = 0, but ineffective spin lock, leading to relaxation.Here, we explore this sequence, as well as an extension of if with windows on both channels, at 55 kHz.The rotor period is then about 10 times shorter than in previous work, which can be expected to result in more efficient transfer due to a more effective spin lock.This motivates revisiting this sequence on modern NMR spectrometers at 600 to 1200 MHz where wCP might become a more competitive heteronuclear transfer element.In addition, it motivates the theoretical reconsideration and reanalysis with simplified definition of the optimal condition.
In this article we investigate windowed cross polarization (wCP), in which windowed rf-field is applied on one or both channels, without phase alternation.We analyze the wCP sequence with respect to the pulse flip angle values and derive optimal FLip ANgle conditions (FLAN) under which zero-quantum (ZQ) and double-quantum (DQ) transitions are observed.We provide theoretical, numerical and experimental analysis of the transferred signals with CP and wCP and show the connection between Hartmann-Hahn and FLAN conditions.For most cases wCP demonstrates similar efficiency as CP.We find that a windowed version of spectrally induced filtering in combination with cross polarization (SPECIFIC-CP) [44] can provide more efficient transfers, which is demonstrated with lipid bilayer samples of the membrane protein Influenza A M2 using 55.555 kHz MAS.

Result and discussion
Fig. 1A shows the wCP pulse sequence, where a single pulse with a window every rotor period is applied on at least one channel.The wCP sequence was proposed along with other AMCP sequences by Barbara and Williams [72] in 1992 and investigated by Hediger et al. [70] in 1993.The arrangement of the windowed pulses can also be different, for example, a pair of pulses with a single window between them.However, we find that these two approaches provide the same efficiency.While in simulations and experiments we applied both approaches (see the supplemental information), for simplicity, the following considers only the pulse sequence shown in Fig. 1A.
In this work we apply a well-known approach using fictitious spin-1/2 formalism [74] and average Hamiltonian theory (AHT) [75].Note that Floquet formalism provides an alternate approach to simulation of AMCP sequences as intensively investigated by Hediger et al in 1993 [70], 1995 [76] and 1997 [77], and Nielsen ).Red and blue stars represent positive and negative conditions.Red and blue lines represent the flip angles for CP (the pulse length is one rotor period on both channels), and the sum or difference between rf-field values match Hartmann-Hahn conditions [1].The diagonal (cyan crosses) represents a special case, where positive transfer is observed if tI-tS.(C) and (D) wCP 1D (HN)H spectra as a function of windowed pulse parameters on proton and nitrogen channels.(C) with the condition tH ¼ tN and (D) as a function of windowed pulses on the nitrogen channel, tN.In both panels, the 1 H? 15 N transfer was kept constant: CP with linear amplitude modulation 80 % ?100 % of 102 kHz and 44 kHz on 1 H and 15 N channels, respectively, and 800 ls mixing.wCP was applied for 15 N? 1 H transfers.To calculate the power levels of pulses in watts, the reference values of 90°optimized pulses with lengths of 2.25 ls (111.111kHz, proton channel) and 3.12 ls (80.128 kHz, nitrogen channel) were used.(C) 10 wCP 15   [58].In the Supplementary Information (SI), two approximate solutions are derived.The first, Eq.S16, can qualitatively describe CP and wCP profiles with the same pulse lengths and an arbitrary set of flip angles applied on the two channels.
(By wCP profile, we mean the transfer efficiency with respect to two rf-field strengths or flip angles).The second, Eq.S39, can qualitatively describe wCP profiles with different pulse lengths and flip angles.Figure S1 compares the approximate theoretical and numerically exact curves.Here we show the key points for the first approximate solution.
The transferred signal for wCP with a fictitious spin-1/2 formalism [74] and in the tilted rf-field frame [78] can be described as follows: where the integration over orientation (XÞ indicates the usual averaging over powder Euler angles, a; b; c ð Þ [78].U ð1;4Þ tot and U ð2;3Þ tot are DQ and ZQ unitary propagators.The derivation of the Eq. ( 2) is shown in the SI (Eq.(S1-S12)).
The first order ZQ and DQ Average Hamiltonian [79] (FOH) can be calculated for the following ZQ flip angle (FLAN) conditions, where a is the flip angle acquired by the spin I or S during one rotor period: Then ZQ and DQ FOH are: On the other hand, with the following DQ FLAN conditions: ZQ and DQ FOH are: The equation governing the signal transferred via wCP (Eq.( 7)) has the same structure as for CP reported previously by Zhou and Ye [81]: the signal is composed of ZQ and DQ contributions with opposite signs.Such identity in the structure shows the connection between the CP and wCP sequences.
The schematic presentation of FLAN conditions for positive (red) and negative (blue) polarization transfers are shown in Fig. 1B.Red and blue stars show FLAN conditions for wCP, while red and blue lines show the set of flip angles for CP (the pulse length is one rotor period) Hartmann-Hahn conditions.Note the additional transfer conditions that appear for wCP.We note here a similarity to the related sequence, delays alternating with nutation for tailored excitation (DANTE).[82][83] The nutation behavior of the windowed pulses (for spins within the excitation bands of DANTE) during one rotor period can be approximated with the average nutation frequency over this period.
Along the diagonal of the wCP profile (Fig. 1B), the cyan crosses represent a special case where signal transfer occurs away from Hartmann Hahn conditions, but unlike other conditions, transfer is only observed if the pulse lengths applied on the two channels differ.In other words, for wCP if a I ¼ a S and t I ¼ t S the transferred signal is zero.This can be explained by the fact that under these conditions the rf-field Hamiltonian with a fictitious spin-1/2 formalism [74] only depends on DQ operators.Then the total ZQ Hamiltonian only depends on dipolar interaction, which results in U ð2;3Þ tot ¼ 1.The modified Eq. ( 2) can be written as: tot is calculated in the same way as before (Eqns.S8-S15 in SI).If we apply the diagonal flip angle conditions: and: Substituting Eq. ( 12) into Eq.( 9), we obtain that the measured signal is zero: In summary, wCP results in transfer along the diagonal of the profile (a I ¼ a S ) when t I -t S .
Fig. 1C and D confirm this conclusion, where we show the experimental wCP 1D (HN)H spectra as a function of pulse lengths and flip angle values.
When we keep a H ¼ a N ¼ 90 and t H ¼ t N (cyan spectra in Fig. 1C), we observe a noise level signal.By contrast, for the condition a H ¼ 270 , a N ¼ 90 and t H ¼ t N , negative DQ transfer is observed (blue spectra in Fig. 1C).
Fig. 1D shows the dependence of the pulse length difference:   13 C? 1 H transfer, with continuous (blue) and windowed (red) pulses on the 13 C channel.Linear ramped amplitude modulation, 100%?80%, was applied on the 1 H channel.In both cases, the 1 H? 13 C transfer was kept constant, while 13 C? 1 H conditions were varied.The carrier frequency was set to 100 ppm.The mixing time for both transfers was 360 ls. and t H , the positive signal (ZQ dominating) increases.Note that the signal in panels C and D are not to scale, and that the signal along the diagonal of the profile is typically not the most efficient (vide infra).
We next focus on conditions where a I -a S .
In Fig. 2 we compare the numerical and FOH (Eq.( 7)) curves.Fig. 2A shows the comparison of the numerical (solid lines) and analytical (gray stars) curves as a function of mixing time for different pulse lengths on both channels (keeping t I ¼ t S ¼ t p ).We observe a good agreement between numerical and FOH curves.While the size of the window has an influence on the effective dipolar scaling factor, for some pulse lengths the theoretical efficiency of wCP reaches 75%.Fig. 2B shows numerical curves with the pulse length on one channel kept constant (t S = 0.5T R ) and the pulse length of the second channel varied from 0.5T R to T R (no window).In this case, we also observe similar transfer efficiencies, but with equal pulse lengths on both channels wCP achieves a slightly better performance.
While in numerical simulation, the rf-field strength is not restricted, for experiments high power cannot be reached for low gamma spins.We therefore limited our experimental profiles to the areas indicated with dashed rectangles in Fig. 3A for wCP and Fig. 4A for CP.Fig. 3B-E and 4B-E compare the theoretical (B), numerical (C-D) and experimental (E) wCP (Fig. 3) and CP (Fig. 4) profiles.For Figs. 3 and 4 we observe a good agreement between theoretical (B) and two spin system simulated (C) profiles, as well as three spin system simulation (D) and experimental (E) profiles.
Fig. 3 shows that over a broad range of flip angle values, wCP provides polarization transfers and the main effect depends on the sums or differences between flip angles.Note that since the sequence is rotor synchronized, the dependence on flip angle could equally be described as a dependence on the average rf-field strength (average nutation frequency).For Fig. 4 there is a difference between three spin system (D) and two spin system (C) simulations.In these two figures we observe additional gaps, which are not seen in two-spin system simulations (Fig. 4C).These gaps are related to homonuclear rotary resonance recoupling at about 55 and 110 kHz for the I spin.
While Fig. 3E and Fig. 4E demonstrate wCP and CP profiles at a specific mixing time, Fig. S2A shows the comparison of the experimental CP and wCP curves as a function of the mixing times.Similar build up curves are observed for both cases.
The ratio between rf-field strength values (in kHz) of CP and wCP sequences has an inverse dependence on the normalized pulse length: t p =T R À Á À1 .However, the requirement of the rf-field power values in Watts for windowed pulses is squared in comparison to the ratio in kHz, leading to a ratio for rf-field power values of t p =T R À Á À2 .Therefore, the application of this sequence might be limited at ultra-fast MAS rates, especially for ZQ transfers.On the other hand, faster MAS rates are typically achieved through the use of a smaller rotor and coil, leading to an improvement in the maximum possible rf-field strength.A low power condition also exists in the wCP sequence (vide infra) with matching flip angles on the two channels.It is not particularly efficient under the conditions reported here, but might be interesting for future developments, e.g.> 200 kHz MAS where the required rf-field strength for CP becomes extreme.In Figs. 3 and 4 we kept similar rf-field strength ranges.In this case, CP and wCP profiles appear to be different.However, according to the derivations in the SI, if we rescale according to the flip angle, CP and wCP profiles should share common features.This is demonstrated in Fig. 5.For wCP (Fig. 5B) the windowed pulses were applied on the S spin only and continuous rf-field power was applied on the I spin.
With CP (Fig. 5A) we observe the expected one DQ and two ZQ conditions within this region, according to the Hartmann Hahn conditions [1].Comparing CP (A) and wCP (B) profiles we obtain a full match for these one DQ and two ZQ conditions in the flip angle domain.However, an additional positive transfer and an additional negative transfer are observed for wCP.
For wCP the additional positive transfer (red) is observed when equal flip angle values are applied on both channels, e. g. at a I ¼ a S ¼ 270 .As was explained above (and demonstrated exper-imentally with Fig. 1D) such transfers are only observed when the pulse lengths on the two channels differ.The additional negative transfer (blue) is observed with the next flip angle set, a I ¼ 810 ; a S ¼ 270 , according to FLAN conditions (Fig. 1B).
To experimentally demonstrate the matching conditions visible in Fig. 5, two slices from the CP and wCP profiles were recorded (the dashed white lines of Fig. 5).The experimental data is shown in Fig. 6, where the upper spectra were recorded with CP (blue) and the lower with wCP (red).As in simulation, we observe three transfer conditions for CP, and 5 for wCP.
One of the additional transfer conditions for wCP is observed in Fig. 6 when the applied pulses on different channels have the same flip angles.This transfer condition is an additional example of the those seen along the cyan diagonal in Fig. 1A.Note that second order CP [84][85] also lies along the diagonal, but is typically less efficient than ZQ or DQ CP.
Fig. 6C compares CP (red) and wCP (blue) (HC)H spectra with the highest performance from each profile.In that case we observe better performance for CP, especially in the methyl region.However, note that the relatively short transfer time might not be optimal for wCP, given the reduced scaling factor as compared with CP (Fig. 1 and Fig. S2A).
To improve performance despite rf-field inhomogeneity, a small ramped amplitude modulation (from 95% to 100%) can be additionally applied on the second channel for windowed pulses.Similar to ramped CP [41] a linear ramped amplitude improves the performance.The benefit of amplitude modulation was also demonstrated in RESPIRATION CP experiments [86].Although we mostly found that the amplitude modulation on one of the channels was enough to reduce the influences of rf-field inhomogeneity and offset, in some cases we applied an additional small ramped amplitude modulation on the second channel to further improve the performance.Fig. S3 demonstrates such a case, comparing 1D (HC)H spectra with amplitude modulation on both channels.
The previous comparison of CP and wCP was made using strong rf-field strengths, applied on both channels.Under these conditions we did not find any clear advantage of using wCP with respect to CP.In contrast, we found that wCP can outperform CP when low rf-field strength (small flip angle) is applied on at least one of the channels.Baldus et al. [44] in 1998 and Laage et al. [87] in 2008 demonstrated that low rf-field strength applied on ): for CP 67 kHz and 19 kHz (80%?100%, linear ramp) were applied on 13 C and 1 H channels, respectively, while for wCP 64 kHz (no windows, 95%?100% linear ramp) and 57 kHz (6 ls pulses with 80%?100% linear ramp) were applied on 13 C and 1 H channels, respectively.360 ls contact time was used.The data were acquired at an 800 MHz spectrometer with 55.555 kHz MAS and using S31N M2.Additional experimental details are given in the SI.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)the 13 C channel allowed transfer to select regions of the spectrum.The method was named SPECIFIC-CP.Such conditions result in a 13 C flip angle of around 90°for the SPECIFIC-CP conditions.The application of wCP with about 90°pulses on 13 C channel also preserves the band-selectivity despite the higher rf-field strength.This is demonstrated in Figs.7-9.Fig. 7 shows 1D (HC)H spectra as a function of 1 H flip angle (and rf-field strength) for CP and wCP while maintaining a 90°flip angle on the 13 C channel.The shape of the profile is maintained.The pulse lengths on the 13 C channel are 18 ls (A, CP), 9 ls (C, wCP) and 4.5 ls (E, wCP).For wCP higher rf-field strengths are applied to achieve the same flip angle value as for CP.The rf-field strength of these pulses on the 13 C channel required to maintain a 90°flip angle were then: 14.2 kHz (A), 28.3 kHz (C) and 56.6 kHz (E).With respect to rf-field power levels in Watts (W) these values were: 0.85 W, 3.42 W and 13.6 W, respectively.Fig. 7F compares CP (green) and wCP (blue, red) (HC)H spectra with the highest performance from each profile.Note that there appears to be an improvement in ZQ polarization transfer for wCP (positive signals) and a decrease in performance for DQ polarization transfer.Fig. S4 demonstrates similar experiments, acquired at an 800 MHz spectrometer.
We next take a closer look at the band-selectivity with application of 90°pulses on the 13 C channel in wCP experiments.Fig. 8A-B demonstrates the dependence of carbonyl (A) and aliphatic (B) regions on offset values with 44.6 kHz rf-field strength on the 13 C channel.As can be seen, beyond a 23 kHz offset the carboxyl (A) and aliphatic (B) region are not observed.Despite high rffield strength, the band-selectivity is preserved.Similar to SPECIFIC-CP, the application of windowed 90°pulses on the 13 C channel can be called a SPECIFIC-wCP.
When in reverse 90°pulses are applied on the 1 H channel (Fig. 8C), more complete band excitation is observed: all regions of the 13 C spectrum have similar intensities.Under these conditions, 67 kHz is applied on the 13 C channel for the CP case, and the flip angle is 450°for both CP and wCP.The carrier frequency was set to 100 ppm.The comparison of CP (red) and wCP (blue) spectra shows better performance for wCP under these conditions.
The higher performance for SPECIFIC-wCP with respect to SPECIFIC-CP can be more clearly observed from 2D (H)CH experiments (Fig. 9), where the carrier frequency is set to the aromatic region (135 ppm, A) or aliphatic region (B, 25 ppm).Selected slices from these data are shown at the right.For all regions, SPECIFIC-wCP has an improved performance as compared with SPECIFIC-CP.In particular, in slice 4, a twofold improvement is observed for carbonyl to amide proton transfer.In general, a larger improvement is seen towards the edges of the specific band.
In the SI we show 2D wCP (H)CH spectra acquired at a 1200 MHz spectrometer, where $ 90°pulses were applied on the 1 H channel (Figure S5).Also in that case we observe more complete excitation of the 13 C region.CP and wCP profiles for weak dipolar coupling pairs (e.g. 13 C-15 N) are compared in Figure S6.

Conclusions
Cross polarization transfers between heteronuclear spins is a fundamental building block used in MAS NMR, where the transfers via dipolar interactions are usually explained with Hartmann-Hahn conditions [1].These conditions declare that maximal transfer occurs when the difference or the sum between applied spinlock rf-field strengths is an integer, n, times the MAS rate (n=±1,±2).In this article we theoretically investigated windowed CP sequences (wCP) that had a single window per rotor period, and derived the flip angle conditions (FLAN), under which zeroquantum and double-quantum transfers were observed.A theoretical approximation of the transfer, as well as numerically exact simulations of two to three spins, were in good agreement with experimental CP and wCP profiles.We showed that the CP and wCP transfer conditions can be best compared with respect to the rf pulse flip angles, or equivalently, the average spin-lock rffield strengths.This revealed a close connection between wCP and CP, namely that the wCP sequence shares conditions of efficient transfer with CP, and also has additional transfer conditions of lower efficiency.Given the similarity between the two sequences, it is perhaps unsurprising that they performed equally well most of the time (note we only considered fast MAS conditions where windows do not cause problems with spin locking).Nevertheless, a benefit in transfer efficiency was observed for wCP for a SPECIFIC-CP transfer condition with small 13 C flip angle.Surprisingly, the higher rf-field strength on 13 C for wCP experiments did not eliminate the band-selectivity.The wCP sequence also showed an improvement in transfer efficiency with 90°pulses on the 1 H channel and 450°pulses on the 13 C channel, conditions under which the transfer has a broad bandwidth.

Experimental methods
The full experimental details are given in the SI.

Fig. 1 .
Fig. 1. (A) wCP.Each rotor period contains a single pulse on each channel with flip angles, lengths and rf-field strengths (aI; tI; mI) and (aS; tS; mS) on channels I and S, respectively.(B) Flip angle values for which the FOH was calculated (Eqns.(1 and 2).Red and blue stars represent positive and negative conditions.Red and blue lines represent the flip angles for CP (the pulse length is one rotor period on both channels), and the sum or difference between rf-field values match Hartmann-Hahn conditions [1].The diagonal (cyan crosses) represents a special case, where positive transfer is observed if tI-tS.(C) and (D) wCP 1D (HN)H spectra as a function of windowed pulse parameters on proton and nitrogen channels.(C) with the condition tH ¼ tN and (D) as a function of windowed pulses on the nitrogen channel, tN.In both panels, the 1 H? 15 N or D are the time integrated functions, which depend on the sums (for DQ) or the differences (for ZQ) between rf-field strength values and dipolar time dependent functions [78,80].Their explicit forms are shown in the SI.For both cases the measured ZQ (positive, D) and DQ (negative, R) signals are described with the next simple Eqn.:

Fig. 2 .Fig. 3 .
Fig. 2. wCP curves.For both (A) and (B) the MAS was 55.555 kHz (18 ls rotor period) and the spin system was two spins with 6 kHz dipolar coupling.The same flip angle set, ðaI ¼ 630 ; aS ¼ 270 Þ, was applied.(A) First Order Average Hamiltonian Approximation (gray stars) and numerical simulation (solid lines) as a function of mixing time under different pulse lengths (equal lengths on each channel, t I = t S = t p ).The pink solid line represents CP experiments that satisfy the Hartmann-Hahn condition.(B) Numerical simulations, where the pulse length on S channel was the same (t S = 0.5T R ) and the pulse length on the I channel was varied as labeled in the legend.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4 .
Fig. 4. CP profiles: the transferred signal intensity (t mix = 0.18 ms) as a function of applied rf-field.Broad numerical profile (A), first order Hamiltonian (B), numerical (C-D) and experimental (E) transferred signal intensity (t mix = 0.18 ms) as a function of applied rf-field, obtained with continuous pulses.The dashed rectangle indicates the experimentally investigated area, which was investigated theoretically, numerically and experimentally (B-E).(B) Analytical CP profile was obtained with first order approximation of the Hamiltonian (Eq.S16).(A, C) The CP simulations considered an IS spin system with heteronuclear dipolar couplings of 11 kHz.(D) The CP simulations considered an I 2 S spin system with heteronuclear dipolar couplings of 11 kHz and 3.2 kHz and a homonuclear dipolar coupling of 5 kHz.(E) Experimental data were recorded in 1D following (HN)H transfer, with polarization transfer conditions from 1 H to 15 N kept constant (ramp CP with 0.72 ms mixing) and the polarization transfer condition for

Fig. 5 .
Fig. 5. Simulation of CP (A) and wCP (B).For both cases, the numerical simulations considered an I 2 S spin system with heteronuclear dipolar couplings, 21 kHz and 7.2 kHz and a homonuclear dipolar coupling of 8 kHz.The mixing time is 90 ls and the MAS rate 55.555 kHz.(A) Continuous pulses on I and S spins.(B) Continuous pulses on I spins and windowed pulse every rotor period with 9 ls length on the S spin.The dashed white lines show the slices that were experimentally recorded and shown in Fig. 6.

Fig. 6 .
Fig. 6. (A) CP (blue) and wCP (red) 1D (HC)H CP spectra as a function of 1 H rf-field strength with different 13 C pulse lengths / rf-field strengths: blue -continuous / 43.16 kHz and red 9 ls / 84 kHz.(B) Schematic presentation of the pulse sequence for 13 C? 1 H transfer, with continuous (blue) and windowed (red) pulses on the 13 C channel.Linear Figure C compares 1D CP (red) and wCP (blue) (HC)H spectra with the best performance for each case.The data were acquired at a 600 MHz spectrometer with 55.555 kHz MAS and the M2 sample.Additional experimental details are given in the SI.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 7 .
Fig. 7. CP (A) and wCP (C, E) 1D (HC)H spectra as a function of 1 H rf-field strength (for 1 H? 13 C transfer) with different 13 C rf-field strengths and pulse lengths: (A) CP: 14.2 kHz; (C) wCP: 28.3 kHz and 9 ls and (E) wCP: 56.6 kHz and 4.5 ls.Figure F compares 1D CP and wCP (HC)H spectra with the best performance for each case.(B), (D) and Figure F compares 1D CP and wCP (HC)H spectra with the best performance for each case.(B), (D) and (F) depict the 1 H? 13 C cross polarization pulse sequence, which was used in each case.The ramped 13 C? 1 H cross polarization step was kept the same with an optimized ZQ Hartmann-Hahn condition.The spectra were acquired at a 600 MHz spectrometer with 55.555 kHz MAS and the M2 sample.The carrier frequency was set to 135 ppm.Further experimental details are given in the SI.E. Nimerovsky, S. Becker and L.B. Andreas Journal of Magnetic Resonance 349 (2023) 107404

Fig. 8 .
Fig. 8. (A)-(B) Experimental 1D wCP (H)C spectra as a function of offset with respect to carbonyl (A, 175 ppm is on resonance) and aliphatic regions (B, 35 ppm is on resonance).The 1 H channel conditions were 15 ls / 78.38 kHz pulses and linear ramp (100%?90%) and the 13 C channel conditions were 6 ls / 44.6 kHz pulses with linear ramp (95%?100%).In all cases 540 ls contact time was used.(C) Experimental 1D CP (red) and wCP (blue) (H)C spectra with the same flip angle sets (aH 90

Fig. 9 .
Fig. 9. SPECIFIC-CP (red) and SPECIFIC-wCP (blue) 2D (H)CH spectra.(A) The carrier frequency position was set to 135 ppm.(B) The carrier frequency position was set to 25 ppm.The data were acquired at an 800 MHz spectrometer with 55.555 kHz MAS and using S31N M2.The full experimental details are given in the SI.The slices from 2D experiments are indicated with numbers 1-8.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)