Recent 1D and 2D TD–NMR Pulse Sequences for Plant Science

Time domain nuclear magnetic resonance (TD–NMR) has been widely applied in plant science in the last four decades. Several TD–NMR instruments and methods have been developed for laboratory, green-house, and field studies. This mini-review focuses on the recent TD–NMR pulse sequences applied in plant science. One of the sequences measures the transverse relaxation time (T2) with minimal sample heating, using a lower refocusing flip angle and consequently lower specific absorption rate than that of conventional CPMG. Other sequences are based on a continuous wave free precession (CWFP) regime used to enhance the signal-to-noise ratio, to measure longitudinal (T1) and transverse relaxation time in a single shot experiment, and as alternative 2D pulse sequences to obtain T1–T2 and diffusion-T1 correlation maps. This review also presents some applications of these sequences in plant science.


Introduction
The fast and non-invasive methods to determine oil content in oilseeds were the first applications of time domain nuclear magnetic resonance (TD-NMR) in plant science, and they have since been used in germplasm evaluation and plant breeding programs [1]. The analyses of oilseeds, with low moisture content, are performed by measuring the intensity of free induction decay (FID) after a radiofrequency (rf) pulse. The measurements are performed at 50 µs to avoid interference of solid components (proteins, carbohydrates). Spin-echo sequence has been used to measure moisture and oil content. The oil content is calculated by the echo intensity at 7 ms and the moisture content by the difference between the FID (at 50 µs) and echo intensities [2].
In the lasts two decades, one-dimensional (1D) and two-dimension (2D) pulses sequences have been proposed for TD-NMR [3,[12][13][14][15]. This mini-review focuses on sequences, such as CPMG with low refocusing pulses, which minimize the sample-heating problem, and 1D and 2D sequences based on a special regime of steady state free precession, known as continuous wave free precession (CWFP).

CPMG with Low Refocusing Pulses
Most TD-NMR studies in plants are based on the measurement of T2 relaxation time obtained by the CPMG pulse sequence, which consists of a π/2 pulse in the x axis followed by time (τ) and a train of refocusing π pulses in the y axis separated by time 2τ (Figure 1 using θ = π ) [5]. The CPMG pulse sequence is very robust with negligible dependence on pulse imperfections and magnetic field inhomogeneity; in addition, a single scan provides the full relaxation decay [5,10,16]. The CPMG signals have been used to study plants in laboratories, greenhouses, and in the field [17][18][19][20][21][22]. Normally, the CPMG decay is measured for 5T2 to obtain suitable discrete or continuous fitting. Since T1 and T2 are similar in low field TD-NMR spectrometers, the scans can be performed without signal saturation, even with a recycle delay (RD) of a few milliseconds. This procedure is important to being able to carry out experiments in short measuring times for static samples or in moving samples on a conveyor belt. When RD is in the order of milliseconds, the π pulses are applied almost continuously, which can cause probe and power amplifier overheating [17]. In some cases, mainly for living tissues, a continuous high power rf pulse train may exceed the specific absorption rate (SAR), which can damage the sample by excessive heating and also lead to erroneous T2 values. The sample heating problem is a critical issue, for example, for oilseeds, as T2 varies exponentially with temperature [19,23].
To minimize instrumental or sample-heating problems, T2 measurements have shown accuracy using CPMG with a low refocusing flip angle (LRFA-CPMG) ( Figure 1 using θ << π), which may lower rf power by up to one order of magnitude [16]. In LRFA-CPMG, T2 accuracy depends on experimental parameters, such as τ value and magnetic field inhomogeneity. Figure 2 shows the experimental CPMG decay using π, 3π/2, π/2, and π/4 refocusing pulses, for a soybean oil sample, using two τ values in two magnetic fields with different field homogeneity. Figure 2A Figure 2C,D show CPMG signals for a signal with 100 Hz line width and τ values of 0.1 ms ( Figure 2C) and 0.4 ms ( Figure  2D). These results show that for a quite homogeneous magnet, a very low refocusing pulse with short τ values gives a signal with the same intensity and T2 values as that of classical CPMG with π pulses. Signal intensity decays for longer τ values and magnetic field inhomogeneity; however, T2 values are still rather accurate [16]. The CPMG signals have been used to study plants in laboratories, greenhouses, and in the field [17][18][19][20][21][22]. Normally, the CPMG decay is measured for 5T 2 to obtain suitable discrete or continuous fitting. Since T 1 and T 2 are similar in low field TD-NMR spectrometers, the scans can be performed without signal saturation, even with a recycle delay (RD) of a few milliseconds. This procedure is important to being able to carry out experiments in short measuring times for static samples or in moving samples on a conveyor belt. When RD is in the order of milliseconds, the π pulses are applied almost continuously, which can cause probe and power amplifier overheating [17]. In some cases, mainly for living tissues, a continuous high power rf pulse train may exceed the specific absorption rate (SAR), which can damage the sample by excessive heating and also lead to erroneous T 2 values. The sample heating problem is a critical issue, for example, for oilseeds, as T 2 varies exponentially with temperature [19,23].
To minimize instrumental or sample-heating problems, T 2 measurements have shown accuracy using CPMG with a low refocusing flip angle (LRFA-CPMG) (Figure 1 using θ << π), which may lower rf power by up to one order of magnitude [16]. In LRFA-CPMG, T 2 accuracy depends on experimental parameters, such as τ value and magnetic field inhomogeneity. Figure 2 shows the experimental CPMG decay using π, 3π/2, π/2, and π/4 refocusing pulses, for a soybean oil sample, using two τ values in two magnetic fields with different field homogeneity. Figure 2A Figure 2C) and 0.4 ms ( Figure 2D). These results show that for a quite homogeneous magnet, a very low refocusing pulse with short τ values gives a signal with the same intensity and T 2 values as that of classical CPMG with π pulses. Signal intensity decays for longer τ values and magnetic field inhomogeneity; however, T 2 values are still rather accurate [16]. Plants 2021, 10, x 3 of 13 Figure 2. Experimental CPMG signals using refocusing angle pulse of π, 3π/4, π/2, and π/4 for soybean oil in different homogeneities (Δν = 15 Hz (A,B) and 100 Hz (C,D)), τ = 0.1 ms (left) and 0.4 ms (right). Adapted from publication [16]. Copyright (2011), with permission from Elsevier. Figure 3A shows CPMG decays of a castor bean seed (oil signal) obtained with the CPMG pulse sequence with π (red line) and π/2 (black line) refocusing pulses. These experiments were performed in a Minispec instrument (Bruker 20 MHz for 1 H) using τ = 0.5 ms. Figure 3B shows the continuous T2 distribution of Figure 3A signals, obtained by inverse Laplace transform (ILT) [24]. These figures show that LRFA-CPMG, with π/2 refocusing pulses, shows identical T2 values when compared with those of the standard CPMG sequence. As B1∼P 1/2 , where P is the rf power and B1 the amplitude of the magnetic field of the applied pulse, the LRFA-CPMG with π/2 pulses reduced to one quarter of the energy deposited in the sample [16] and also obtained similar T2 results to those obtained by the conventional CPMG method ( Figure 3).  Experimental CPMG signals using refocusing angle pulse of π, 3π/4, π/2, and π/4 for soybean oil in different homogeneities (∆ν = 15 Hz (A,B) and 100 Hz (C,D)), τ = 0.1 ms (left) and 0.4 ms (right). Adapted from publication [16]. Copyright (2011), with permission from Elsevier. Figure 3A shows CPMG decays of a castor bean seed (oil signal) obtained with the CPMG pulse sequence with π (red line) and π/2 (black line) refocusing pulses. These experiments were performed in a Minispec instrument (Bruker 20 MHz for 1 H) using τ = 0.5 ms. Figure 3B shows the continuous T 2 distribution of Figure 3A signals, obtained by inverse Laplace transform (ILT) [24]. These figures show that LRFA-CPMG, with π/2 refocusing pulses, shows identical T 2 values when compared with those of the standard CPMG sequence. As B 1 ∼P 1/2 , where P is the rf power and B 1 the amplitude of the magnetic field of the applied pulse, the LRFA-CPMG with π/2 pulses reduced to one quarter of the energy deposited in the sample [16] and also obtained similar T 2 results to those obtained by the conventional CPMG method ( Figure 3).  Figure 3A shows CPMG decays of a castor bean seed (oil signal) obtained with the CPMG pulse sequence with π (red line) and π/2 (black line) refocusing pulses. These experiments were performed in a Minispec instrument (Bruker 20 MHz for 1 H) using τ = 0.5 ms. Figure 3B shows the continuous T2 distribution of Figure 3A signals, obtained by inverse Laplace transform (ILT) [24]. These figures show that LRFA-CPMG, with π/2 refocusing pulses, shows identical T2 values when compared with those of the standard CPMG sequence. As B1∼P 1/2 , where P is the rf power and B1 the amplitude of the magnetic field of the applied pulse, the LRFA-CPMG with π/2 pulses reduced to one quarter of the energy deposited in the sample [16] and also obtained similar T2 results to those obtained by the conventional CPMG method ( Figure 3).    Figure 4 shows the high correlation (r = 0.98) between T 2 obtained by LRFA-CPMG performed with a π/2 refocusing flip angle with T 2 measured by conventional CPMG for oil seeds of 30 plant species [16]. . Correlation between T 2 values obtained with the CPMG and LRFA-CPMG methods using π/2 as a refocusing and different oilseed species; r = 0.98. Adapted from Publication [16]. Copyright (2011) with permission from Elsevier.
The magnitude of the CWFP signal (|M ss |) is constant (dashed red line on top of the CWFP signal in Figure 6B) and for ψ = (2n+1)π, it is dependent on M 0 , T 1 , and T 2 according to Equation (2). Equation (2) also shows that (|M ss |) is not dependent only on T 1 , as observed in conventional NMR sequences, but on T 1 /T 2 . For example, when T 1 = T 2 (|M ss |) = 0.5M 0 .
Therefore, in the CWFP sequence, the pulse interval can be as short as possible (Tp < T 2 *), which allows the acquisition of thousands of scans per second [26,27]. Consequently, the CWFP sequence is used to enhance the signal-to-noise ratio (SNR) for more than Plants 2021, 10, 833 6 of 13 one order of magnitude in the TD-NMR, without increasing the experimental time [26,27]. The CWFP signal also shows linear correlation with concentration for a sample with similar T 1 /T 2 , allowing fast quantitative measurements [27]. The CWFP sequence has been used to measure oil content in seeds with low oil content, such as peas and maize, without increasing measuring time [26,33].
The CWFP sequence is also used to determine the oil content in seeds moving at 13 cm/s on a conveyor belt. In this experiment, each seed gives a signal where the intensity is proportional to the oil content. Therefore, it is a high-throughput method to determine oil content in more than 20,000 seeds per hour [33].

CWFP Sequence to Measure T 1 and T 2 in a Single Experiment
The CWFP sequence can also be used to measure both relaxation times (T 1 and T 2 ) in a single shot sequence [15,26,27,32,34]. These measurements require acquiring the amplitude of the NMR signal from the first pulse to the CWFP regime, using θ = π/2 and ψ = (2n+1)π (Figure 7). After the first π/2 pulse, the signal magnitude is at a maximum and is proportional to M 0 . After the following pulses, the signal oscillates in a transient period depending on T 2 * (Figure 7, dark gray region). When oscillation stops, the signal reaches a quasi-stationary state (QSS), indicated by a red arrow in Figure 7. From the QSS, the signal decays to the CWFP regime with a time constant T* that depends on θ and on relaxation times, according to Equation (3).
Plants 2021, 10, x 7 of 13 Figure 7. Typical CWFP signal (magnitude) obtained from the first pulse. The dark grey region indicates the initial signal that oscillates for a time that is dependent on T2*. When the oscillation stops, a quasi-stationary state (QSS) signal is observed (red arrow). The light grey region shows T* decay from the QSS to the steady state regime |Mss| (white region).
The CWFP sequence results in signals with the highest variation in T* amplitude or the highest dynamic range (DR) for samples with T1 >> T2. On the other hand, DR is minimal when T1~T2, which makes it difficult to determine T* in low SNR signals. To overcome this situation, a second CWPF sequence was developed to measure T1 and T2 for samples with T1~T2 [35]. This sequence has a π/2 pulse separated by a Tp/2 before the CWFP pulse train (ϕ = x) (Figure 8). It is a Carr-Purcell sequence with π/2 refocusing pulses and it is named CP-CWFP [35]. Recently, the use of a π phase alternation, on the CWFP loop of pulses (ϕ = −x), has been proposed ( Figure 8) [15]. This sequence, known as CP-CWFPx−x, has been the best CWFP sequence due to its highest DR for both conditions T1~T2 and T1 > T2 and the signal can also be measured on resonance (ψ = 0) [15]. Typical CWFP signal (magnitude) obtained from the first pulse. The dark grey region indicates the initial signal that oscillates for a time that is dependent on T 2 *. When the oscillation stops, a quasi-stationary state (QSS) signal is observed (red arrow). The light grey region shows T* decay from the QSS to the steady state regime |Mss| (white region).
The CWFP sequence results in signals with the highest variation in T* amplitude or the highest dynamic range (DR) for samples with T 1 >> T 2 . On the other hand, DR is minimal when T 1~T2 , which makes it difficult to determine T* in low SNR signals. To overcome this situation, a second CWPF sequence was developed to measure T 1 and T 2 for samples with T 1~T2 [35]. This sequence has a π/2 pulse separated by a Tp/2 before the CWFP pulse train (φ = x) (Figure 8). It is a Carr-Purcell sequence with π/2 refocusing pulses and it is named CP-CWFP [35]. Recently, the use of a π phase alternation, on the CWFP loop of pulses (φ = −x), has been proposed ( Figure 8) [15]. This sequence, known as CP-CWFPx−x, has been the best CWFP sequence due to its highest DR for both conditions T 1~T2 and T 1 > T 2 and the signal can also be measured on resonance (ψ = 0) [15]. Typical CWFP signal (magnitude) obtained from the first pulse. The dark grey region indicates the initial signal that oscillates for a time that is dependent on T2*. When the oscillation stops, a quasi-stationary state (QSS) signal is observed (red arrow). The light grey region shows T* decay from the QSS to the steady state regime |Mss| (white region).
The CWFP sequence results in signals with the highest variation in T* amplitude or the highest dynamic range (DR) for samples with T1 >> T2. On the other hand, DR is minimal when T1~T2, which makes it difficult to determine T* in low SNR signals. To overcome this situation, a second CWPF sequence was developed to measure T1 and T2 for samples with T1~T2 [35]. This sequence has a π/2 pulse separated by a Tp/2 before the CWFP pulse train (ϕ = x) (Figure 8). It is a Carr-Purcell sequence with π/2 refocusing pulses and it is named CP-CWFP [35]. Recently, the use of a π phase alternation, on the CWFP loop of pulses (ϕ = −x), has been proposed ( Figure 8) [15]. This sequence, known as CP-CWFPx−x, has been the best CWFP sequence due to its highest DR for both conditions T1~T2 and T1 > T2 and the signal can also be measured on resonance (ψ = 0) [15].

CWFP Sequence to Measure T 1 in a Single Shot Experiment
CWFP sequences have also been used to measure T 1 in a single shot experiment [14]. According to Equation (3), T*, of the CWFP signal, can be used to measure T 1 , when very low flip angles are used. To obtain the maximum DR, the CWFP-T 1 sequence (Figure 9) starts with a π pulse, which inverts the magnetization, and it is followed by a train of low flip angle pulses [14]. CWFP sequences have also been used to measure T1 in a single shot experiment [14]. According to Equation (3), T*, of the CWFP signal, can be used to measure T1, when very low flip angles are used. To obtain the maximum DR, the CWFP-T1 sequence (Figure 9) starts with a π pulse, which inverts the magnetization, and it is followed by a train of low flip angle pulses [14].  [14]. Copyright (2016), with permission from Elsevier. Figure 10 shows that the single shot CWFP-T1 signal (θ~π/20) has a profile similar to the IR experiment (square symbols) [14]. The low SNR in the CWFP-T1 signal is due to the small flip angles in the pulse sequence. However, the SNR of the CWFP-T1 signals can be enhanced using post-acquisition digital filters, such as Savistky-Golay or wavelet filters [36]. Figure 10 shows that the CWFP-T1 signal has similar results to those obtained  [14]. Copyright (2016), with permission from Elsevier. Figure 10 shows that the single shot CWFP-T 1 signal (θ~π/20) has a profile similar to the IR experiment (square symbols) [14]. The low SNR in the CWFP-T 1 signal is due to the small flip angles in the pulse sequence. However, the SNR of the CWFP-T 1 signals can be enhanced using post-acquisition digital filters, such as Savistky-Golay or wavelet filters [36]. Figure 10 shows that the CWFP-T 1 signal has similar results to those obtained with the standard IR method; however, in a shorter measuring time [14].

Figure 9.
Pulse diagram for the CWFP-T1 pulse sequences. Adapted from publication [14]. Copyright (2016), with permission from Elsevier. Figure 10 shows that the single shot CWFP-T1 signal (θ~π/20) has a profile similar to the IR experiment (square symbols) [14]. The low SNR in the CWFP-T1 signal is due to the small flip angles in the pulse sequence. However, the SNR of the CWFP-T1 signals can be enhanced using post-acquisition digital filters, such as Savistky-Golay or wavelet filters [36]. Figure 10 shows that the CWFP-T1 signal has similar results to those obtained with the standard IR method; however, in a shorter measuring time [14]. Figure 10. Inversion recovery signal (black points) and the CWFP-T1 signal (gray line) obtained for a water sample at 27 °C. Adapted from publication [14]. Copyright (2016), with permission from Elsevier. Figure 10. Inversion recovery signal (black points) and the CWFP-T 1 signal (gray line) obtained for a water sample at 27 • C. Adapted from publication [14]. Copyright (2016), with permission from Elsevier.

Two-Dimensional Methods Using the CWFP-T 1 Sequence
Two-dimension (2D) experiments are used in TD-NMR to reduce peak overlap and obtain a correlation between relaxation times (T 1 -T 2 ) [37][38][39] and between diffusion and T 2 (D-T 2 ) [40,41]. Most 2D sequences use the CPMG (T 2 ) to detect the direct dimension ( Figure 11A) signal, which is modulated by T 1 , T 2 , or D sequences.
Recently, it has been demonstrated that a single shot T 1 method can also be used in the direct dimension of the 2D T 1 -T 2 TD-NMR experiment, using the CWFP-T 1 sequence [13]. This sequence was denominated CPMG-CWFP-T 1 and the pulse diagram is as shown in Figure 11B. The 2D-CPMG-CWFP-T 1 maps show higher resolution in the T 1 dimension than do methods that use CPMG as a direct dimension. Moreover, the CPMG-CWFP-T 1 method has a shorter experimental time (up to an order of magnitude) than does IR-CPMG for the samples with T 1~T2 [13]. Figure 12 shows the T 1 -T 2 correlation maps of the signals acquired with castor bean oil at 40 • C ( Figure 12A,B) and green banana at 22 • C ( Figure 12C,D) inverted by a fast multidimensional Laplace inversion (FLI) [37,38] algorithm. IR-CPMG maps were obtained using 32 logarithmically spaced T 1 recovery intervals, from 0.05 to 900 ms for the castor bean oil sample and from 3 to 4,500 ms for the banana sample. The CPMG data were collected using echo numbers (n- Figure 11A) equal to 2400 and 12,000 echoes for the castor bean oil and banana, respectively. For both samples, the τ used on the CPMG method was 0.3 ms. The T 2 -T 1 map measured with the CPMG-CWFP-T 1 sequence used 32 logarithmically spaced CPMG echoes from 2 to 2400 for the castor bean oil and from 2 to 12,000 for the banana sample. CWFP-T 1 Tp was 0.3 ms, for both samples, and the number of CWFP-T 1 loops was 1500 for castor bean oil and 7500 for banana. Using these experimental parameters, for both samples, the CPMG-CWFP-T 1 method ( Figure 12A,C) provides T 1 -T 2 maps with higher resolution in a shorter experimental time (up to an order of magnitude) than those produced with the IR-CPMG sequence ( Figure 12B,D).
Recently, it has been demonstrated that a single shot T1 method can also be used in the direct dimension of the 2D T1-T2 TD-NMR experiment, using the CWFP-T1 sequence [13]. This sequence was denominated CPMG-CWFP-T1 and the pulse diagram is as shown in Figure 11B. The 2D-CPMG-CWFP-T1 maps show higher resolution in the T1 dimension than do methods that use CPMG as a direct dimension. Moreover, the CPMG-CWFP-T1 method has a shorter experimental time (up to an order of magnitude) than does IR-CPMG for the samples with T1~T2 [13].   [37,38] algorithm. IR-CPMG maps were obtained using 32 logarithmically spaced T1 recovery intervals, from 0.05 to 900 ms for the castor bean oil sample and from 3 to 4,500 ms for the banana sample. The CPMG data were collected using echo numbers (n- Figure 11A) equal to 2400 and 12,000 echoes for the castor bean oil and banana, respectively. For both samples, the τ used on the CPMG method was 0.3 ms. The T2-T1 map measured with the CPMG-CWFP-T1 sequence used 32 logarithmically spaced CPMG echoes from 2 to 2400 for the castor bean oil and from 2 to 12,000 for the banana sample. CWFP-T1 Tp was 0.3 ms, for both samples, and the number of CWFP-T1 loops was 1500 for castor bean oil and 7500 for banana. Using these experimental parameters, for both samples, the CPMG-CWFP-T1 method ( Figure 12A,C) provides T1-T2 maps with higher resolution in a shorter experimental time (up to an order of magnitude) than those produced with the IR-CPMG sequence ( Figure 12B,D). Figure 12A shows two peaks of the castor bean oil, which were assigned to the two distinct pools of protons on the fatty acid chain with different mobility or inhomogeneous structural organizations [42,43]. In the CPMG-CWFP-T1 map ( Figure 12A), the population of protons with the highest mobility (T1-T2 = 0.19-0.14 s) represents 15.2%, while the signal of the proton with the lowest mobility (T1-T2 =0.067-0.038 s) corresponds to the  Figure 12A shows two peaks of the castor bean oil, which were assigned to the two distinct pools of protons on the fatty acid chain with different mobility or inhomogeneous structural organizations [42,43]. In the CPMG-CWFP-T 1 map ( Figure 12A), the population of protons with the highest mobility (T 1 -T 2 = 0.19-0.14 s) represents 15.2%, while the signal of the proton with the lowest mobility (T 1 -T 2 = 0.067-0.038 s) corresponds to the greater 2D map proportion of 84.8%. The IR-CPMG map ( Figure 12B) does not allow this same quantification because of its poor resolution.
In the CPMG-CWFP-T 1 map of a banana, the strongest signal ( Figure 12C: T 1 -T 2 = 0.55-0.39 s) was assigned to water in the vacuole and corresponds to 78.1% of the signals. The intermediary intensity signal ( Figure 12C: T 1 -T 2 = 0.19-0.14 s) is attributed to water in the cytoplasm and corresponds to 19.3% of the signal. The 2D map of banana obtained by the IR-CPMG method showed no separation between water signals in the vacuole and cytoplasm. The other signals with low intensity, in IR-CPMG and CPMG-CWFP-T 1 bananas maps, can be related to water in the cell walls or inside the starch granule [34].
The CWFP-T 1 sequence was also used to obtain D-T 1 correlation maps [12]. Figure 13 shows the diagram of the 2D PGSE-CWFP-T 1 pulse sequence to obtain a D-T 1 correlation map. The PGSE-CWFP-T 1 method shows similar results to the conventional IR-PGSE and SR-PGSE methods; however, PGSE-CWFP-T 1 provides a faster analysis and low SAR due to the small flip angle at the CWFP-T 1 method [12].
In the CPMG-CWFP-T1 map of a banana, the strongest signal ( Figure 12C: T1-T2 = 0.55-0.39 s) was assigned to water in the vacuole and corresponds to 78.1% of the signals. The intermediary intensity signal ( Figure 12C: T1-T2 = 0.19-0.14 s) is attributed to water in the cytoplasm and corresponds to 19.3% of the signal. The 2D map of banana obtained by the IR-CPMG method showed no separation between water signals in the vacuole and cytoplasm. The other signals with low intensity, in IR-CPMG and CPMG-CWFP-T1 bananas maps, can be related to water in the cell walls or inside the starch granule [34]. The CWFP-T1 sequence was also used to obtain D-T1 correlation maps [12]. Figure  13 shows the diagram of the 2D PGSE-CWFP-T1 pulse sequence to obtain a D-T1 correlation map. The PGSE-CWFP-T1 method shows similar results to the conventional IR-PGSE and SR-PGSE methods; however, PGSE-CWFP-T1 provides a faster analysis and low SAR due to the small flip angle at the CWFP-T1 method [12]. . PGSE-CWFP-T1 diagram pulse sequence to obtain fast D-T1 correlation maps. Adapted from publication [12]. Copyright (2020) with permission from Elsevier. Figure 14 shows the D-T1 correlation maps obtained by IR-PGSE (A) and PGSE-CWFP-T1 (B) sequences for asparagus stems oriented parallel (in black) and perpendicular (in red) to the pulse field gradient direction. Both maps show similar results; however, PGSE-CWFP-T1 is much faster (up to an order of magnitude) than the IR-PGSE experiment [12]. Applications of the PGSE-CWFP-T1 sequence to measure D-T1 correlation in fruits and vegetables are showing promising results. Figure 13. PGSE-CWFP-T 1 diagram pulse sequence to obtain fast D-T 1 correlation maps. Adapted from publication [12]. Copyright (2020) with permission from Elsevier. Figure 14 shows the D-T 1 correlation maps obtained by IR-PGSE (A) and PGSE-CWFP-T 1 (B) sequences for asparagus stems oriented parallel (in black) and perpendicular (in red) to the pulse field gradient direction. Both maps show similar results; however, PGSE-CWFP-T 1 is much faster (up to an order of magnitude) than the IR-PGSE experiment [12]. Applications of the PGSE-CWFP-T 1 sequence to measure D-T 1 correlation in fruits and vegetables are showing promising results. from publication [12]. Copyright (2020) with permission from Elsevier. Figure 14 shows the D-T1 correlation maps obtained by IR-PGSE (A) and PGSE-CWFP-T1 (B) sequences for asparagus stems oriented parallel (in black) and perpendicular (in red) to the pulse field gradient direction. Both maps show similar results; however, PGSE-CWFP-T1 is much faster (up to an order of magnitude) than the IR-PGSE experiment [12]. Applications of the PGSE-CWFP-T1 sequence to measure D-T1 correlation in fruits and vegetables are showing promising results. Figure 14. D-T1 correlation maps obtained by (A) IR-PGSE and (B) PGSE-CWFP-T1 methods for asparagus stems oriented parallel (in black) and perpendicular (in red) to the PFG. Adapted from publication [12]. Copyright (2020) with permission from Elsevier.

Conclusions
The recent 1D methods based on the CWFP regime or the low-power CPMG sequence have potential as alternatives to CPMG to enhance SNR and resolution and minimize instrumental or sample-heating problems, especially when fast analysis protocols are necessary for plant science studies. The new 2D pulse sequences (CPMG-CWFP-T1 and PGSE-CWFP-T1) based on the direct detection of CWFP-T1 signals are an alternative to obtain T2-T1 and D-T1 correlation maps, with better resolution and lower SAR than those from the standard 2D sequences and are an important advancement in the use of TD-NMR for plant science studies.   [12]. Copyright (2020) with permission from Elsevier.

Conclusions
The recent 1D methods based on the CWFP regime or the low-power CPMG sequence have potential as alternatives to CPMG to enhance SNR and resolution and minimize instrumental or sample-heating problems, especially when fast analysis protocols are necessary for plant science studies. The new 2D pulse sequences (CPMG-CWFP-T 1 and PGSE-CWFP-T 1 ) based on the direct detection of CWFP-T 1 signals are an alternative to obtain T 2 -T 1 and D-T 1 correlation maps, with better resolution and lower SAR than those from the standard 2D sequences and are an important advancement in the use of TD-NMR for plant science studies.