Figure 9.3

Figure 9.3

The second term gives 2Ix Sx, a mixture of DQC and ZQC which is not observable in the FID. The INEPT transfer term, 2SxIz, is multiplied by cos(^a t1), which is the amplitude modulation of our13 C FID according the "history" of the1 H(I spin) chemical-shift evolution during the t1 delay. This is how the "encoding" of the chemical shift of the proton comes out in precise mathematical terms.

We can diagram this experiment in a way that summarizes the flow of magnetization (Fig. 9.3). First draw the two nuclei that are being correlated, and indicate the relationship that will lead to a crosspeak: a 1H bonded to a 13C. The "relationship" is a single bond, which leads to a large J coupling (~150 Hz). We start the flow with excitation of the proton (step 1: preparation). Next we measure the chemical shift of the proton indirectly during the t1 delay (step 2: evolution, shown by a dotted circle around the 1H and the label "t1"). Then we transfer the coherence from 1H to 13 C via INEPT transfer (step 3: mixing, indicated by an arrow from 1H to 13C labeled "1 Jch"). Finally, we measure the chemical shift of the 13 C directly by recording a 13 C FID (step 4: detection, shown by a solid circle around the 13C labeled "t2"). We call the time domain of the directly detected FID t2 because it comes after t1 in the sequence. This leads to the name F2 for the directly detected frequency domain (horizontal axis of the 2D spectrum) and F1 for the indirectly detected frequency domain (vertical axis in the 2D spectrum). We will use these diagrams throughout our discussion of 2D NMR as a quick way of showing how the experiment works.

Now let's consider a situation where we have three different carbon resonances in our 13C spectrum, each with one proton attached: -C1H1-C2H2-C3H3-. We do the same INEPT experiment with the variable delay t1 and record a series of 13 C spectra. Starting with the first FID, obtained with t1 = 0, we Fourier-transform each FID and load the resulting 13C spectra (with three peaks) into successive rows moving up in a 2D matrix of data. This gives us a stack of 13 C spectra starting with the t1 = 0 spectrum at the bottom and moving upwards as we increment t1: 1, 2, 3 ms, and so on (Fig. 9.4). This intermediate 2D matrix, after the F2 Fourier transform, is sometimes called an interferogram. The height of each 13C peak oscillates (+, 0, —, 0, etc.) at the frequency of its attached proton, and decays due to T2 relaxation of the attached proton, as we look at successive spectra with increasing t1 delay values (moving up in the data matrix). Carbon C1 oscillates at a high frequency because proton H1 has a downfield chemical shift. Carbon C3 oscillates slowly in t1 because its attached proton, H3, has an upfield (lower frequency) chemical shift. In Figure 9.4 the trace of data in each of the three columns is shown to the side of the data column—in each case we have a decaying sinusoidal signal: a t1 FID. It may have taken several minutes in

Figure 9.4


Figure 9.4

real time to acquire the FID for one point of this column (e.g., 16 scans of 13C acquisition), but in the reconstructed time domain of t1 the time difference is only the t1 increment (e.g., 1 ms) between successive data rows.

The second Fourier transform is performed on each of the columns of the data matrix, starting with the first one on the left side and moving to the right side. Most of the columns in Figure 9.4 contain noise, but when we reach the column at the chemical shift of C1 we load the t1 FID (shown to the left side of the data column) and Fourier-transform it to obtain a 1H spectrum with a peak at the frequency of H1. This spectrum is put back into the data column, replacing the t1 FID. Likewise, when we pass through the 13C shift of C2 we transform its t1 FID and obtain the 1H spectrum of H2, and so forth. The completed data matrix (the 2D spectrum) is shown in Figure 9.5. Now for each data column corresponding to a peak in the 13C spectrum we have a 1H spectrum (shown to the side of the data column) of just the one proton attached to that carbon. We have a two-dimensional map that correlates the 1H spectrum (on the vertical or F1 axis) with the 13C spectrum (on the horizontal or F2 axis).

Figure 9.6 shows a 2D HETCOR spectrum with the 1D 13C spectrum displayed at the top (horizontal or F2 dimension) and the 1D 1H spectrum displayed vertically on the left side (vertical or F1 dimension). From any peak (resonance) in the 1H spectrum, we can follow a horizontal line until we encounter a spot or blob of intensity in the 2D data matrix. These clusters of intensity represent correlations in the 2D spectrum and are called crosspeaks. From the crosspeak we move up along a vertical line and run into the 13C peak in the 1D 13C spectrum corresponding to that proton's personal carbon atom (the one it is directly bonded to). In this way we can pair up each proton peak in the 1H spectrum with a carbon peak in the 13C spectrum—a process called chemical-shift correlation.

Figure 9.5

Figure 9.7


Figure 9.7

Figure 9.7 shows the HETCOR spectrum of sucrose. In Chapter 8 we were able to assign nearly all of the !H resonances of sucrose using the selective 1D TOCSY experiment, using the H-g1 and H-f3 doublets as unambiguous starting points. Now we can "transfer" these 1H assignments to the 13C spectrum using the one-bond correlations mapped out in the HETCOR spectrum. All 10 protonated (nonquaternary) carbons give clearly resolved crosspeaks, leading us from each 13C peak to a precise chemical shift in the 1H spectrum, even if the 1H resonance is overlapped in the 1D 1H spectrum. Although the C-g3 and C-g5 peaks are very close to each other in the 13 C spectrum, we can clearly connect the downfield peak of the pair to the H-g3 triplet. Because all of the other protons except H-g5 have been assigned, the upfield peak of the pair must "point" to H-g5, an overlapped resonance between the H-f5 multiplet and the large, broad singlet representing H-g6 and H-f6. All three CH2OH crosspeaks (f1, f6, and g6) appear in a tight triangle at the right side, consistent with the upfield location of CH2 groups relative to CH groups with the same inductive (electron-withdrawing) factors. The steric effect on 13C shifts gives us general locations of 50-60 ppm for CH3-O, 60-70 for CH2-O, 70-80 for CH-O and 80-90 for Cq-O. The top peak of the triangle can be assigned to f1, since the H-f1 singlet (integral area 2) was identified by an NOE across the glycosidic linkage from H-g1. The two other peaks have nearly identical 1H shifts (the peak on the left is slightly lower, or downfield on the 1H shift scale) and we cannot assign them unambiguously. Note that the C-f2 quaternary carbon does not give any crosspeak in the HETCOR spectrum because it is not directly connected to a proton. Only one-bond relationships between 1H and 13C lead to a correlation in this experiment (Fig. 9.3).


Let's apply the general design principles of 2D NMR to the one-bond correlation of 13C (in F2) to 1H (in F1). The specific steps for the HETCOR experiment are as follows:

1. Preparation: a 90o nonselective 1H pulse rotates 1H z magnetization into the x-y plane.

2. Evolution: the 1H magnetization precesses in the x-y plane for a period t1, encoding its chemical shift as a function of t1.

3. Mixing: an INEPT sequence converts the 1H magnetization into antiphase magnetization with respect to its attached 13C nucleus, and then transfers the 1H magnetization to 13C magnetization by simultaneous 90o pulses on both 1H and 13C channels.

4. Detection: The 13C FID is recorded.

The simplest possible pulse sequence would involve a 90o 1H pulse followed by a t1 delay, and then an INEPT sequence (Fig. 9.8). Now we need to make refinements, thinking carefully about what we want to happen during various delays and what we want to prevent or suppress. During the evolution (t1) period, we only want chemical-shift evolution. We would like to refocus the /-coupling (1 /CH) evolution, and it would be nice if we could also refocus the homonuclear (1H-1H) /-coupling evolution. That way the only information that will be encoded during the evolution period is the information that we want to show up in the F1 (QH) dimension of the 2D spectrum: the 1H chemical shift. Suppressing the 1 /CH coupling is accomplished simply by inserting a 13 C 180o pulse into the center of

Figure 9.8

the t1 period (Fig. 9.9). This reverses the /-coupling evolution so that it refocuses during the second half of ti. The 1H chemical shifts continue to evolve because there is no 1H 180° pulse. This is the reverse of the strategy used in Chapter 6 to refocus /-coupling evolution during a 13C evolution period (Fig. 6.33). Refocusing of homonuclear (1H-1H) /-coupling evolution is a bit more complicated and will be discussed last.

Next, we need to examine the 1/(2/) period of the INEPT sequence. In an INEPT experiment we usually allow only the /-coupling evolution, which is required to generate antiphase magnetization, to occur during the 1/(2/) delay. The chemical-shift evolution is suppressed so that the phase of detected peaks is not screwed up by off-resonance chemical-shift effects. We could accomplish this by inserting simultaneous 180° pulses on 1H and 13C in the center of the 1/(2/) delay. But in this experiment, we will not worry about phase. The data will be looked at in "magnitude mode," which calculates a single number for each data point from the real and imaginary parts of the spectrum:

Because chemical-shift evolution just rotates the magnetization vector in the X-/ (real-imaginary) plane without affecting its magnitude, the phase of the detected 13 C magnetization is not important. So we can live with a simple 1/(2/) delay.

As our sequence stands so far, the detected signals will be fully coupled antiphase peaks. This could be messy since CH groups will appear as antiphase doublets (intensity ratio 1, — 1), CH2 groups as antiphase triplets (1,0, -1), and CH3 groups will appear as antiphase quartets (intensity ratio 1, 1, —1, —1) in the F2 (13C) dimension. We need to use proton decoupling during the FID acquisition ("detection") period. This will collapse all of the 13 C multiplets into single peaks. But as we saw in the simple INEPT experiment (Chapter 7), the intensities of the antiphase multiplets add up to zero. So turning on the proton decoupler will lead to a complete loss of all 13C signals in F2. This is not good. The solution is to allow the antiphase magnetization to undergo /-coupling evolution for an additional delay period so that the individual multiplet components come back together into normal (in-phase) multiplets. The optimal refocusing period is 1/(2/) for the CH carbons, 1/(4/) for the CH2 carbons, and about 1/(5/) for the CH3 carbons. The best compromise that allows observation of all three kinds of carbons (Fig. 9.10) is a delay of 1/(3/). Again, we do not worry about13C chemical-shift evolution during this delay because we will display the data in magnitude mode.

Figure 9.11

Finally, we will deal with the question of 1H-1H coupling during the evolution period. This is frosting on the cake, and I only bring it up to show the beautiful things you can accomplish with pulse sequence building blocks. The only protons we are interested in are those bonded to a 13C, because these are the only ones that can transfer magnetization to 13C. If we consider the other 1H nuclei that have homonuclear coupling to the (13C-bound) 1H that is undergoing chemical-shift evolution during the t1 period, we see that they have a 99% chance of being bonded to a 12C, since the abundance of 13C is only 1%. The only exception is the case of geminal coupling (2Jhh), where the proton is attached to the same carbon, which has to be a 13C. The trick is to apply a "magic" selective 180° pulse that only affects protons bonded to 12C, at the same time as the 13C 180° pulse in the middle of the t1 period (Fig. 9.10). From the point of view of the proton we are observing in F1, it is undergoing chemical-shift evolution during the t1 period. The 180° 13C pulse in the center of t1 reverses the 1JCH coupling evolution from its directly bound 13C, and the 180° 1H "magic pulse" on the 12C-bound protons reverses its J-coupling evolution from the vicinal 1H coupling.

How do you generate a 180° 1H pulse that only hits 12C-bound protons? This magic is accomplished by a spin-echo sequence called bilinear rotation decoupling (BIRD), which takes advantage of the different J-coupling evolution of the 13C-bound protons and the 12C-bound protons (Fig. 9.11). For the 12C-bound 1H magnetization, this is just a 180° inversion pulse (two 90° pulses with a spin echo sandwiched in between). The spin-echo part is effectively invisible because the 180° 13C pulse has no effect and the 1/J delay does not lead to any J-coupling evolution (J refers to 1JCH as before). Starting with Iz, we rotate the magnetization to — y', and the spin-echo returns it to — y', reversing any chemical-shift evolution. The final 90°x pulse rotates it from — y' down to — z. This is just what we want because we wish to reverse any J-coupling evolution due to this proton. For the 13C-bound 1H magnetization, however, the spin echo leads to 1JCH evolution for a total of 1/J s, rotating the magnetization from — y' to +y', just as it does in the APT experiment (Chapter 6).The final 90°x pulse rotates it back up to +z, so the 13C-bound 1H nuclei are not affected by this sequence. We can use product operators to verify that for the 13C-bound protons the BIRD sequence is equivalent to a simple 180° 13C pulse, just like the center of the t1 delay in Figure 9.10:

Iz —90°° (1h)^

— Iy —1/(2/)^

2Ix Sz —180y (1h)/180° (13c)^

2Ix Sz —1/(2/Iy —90°° (1h)^


Ix ^

Ix ^

2Iy Sz ^

—2Iy Sz ^ Ix ^


Iy ^

Iz ^

Iz ^

—Iz ^ —Iz ^


2Ix Sz ^

2Ix Sz ^

Iy ^

Iy ^ —2Ix Sz ^

—2Ix Sz

2Iy Sz ^

2Iz Sz ^

2Iz Sz ^

2Iz Sz ^ 2Iz Sz ^

—2Iy Sz

Figure 9.12

Every possible component of 13C-bound 1H net magnetization is affected the same way it would be affected by a simple 180° 13C inversion pulse. For the 12C-bound net magnetization:

Iz ^ -Iy ^ -Iy ^ -Iy ^ -Iy ^ -Iz Ix ^ Ix ^ Ix ^ Ix ^ Ix ^ Ix

Iy Iz Iz Iz Iz Iy

The sequence is exactly like a 180°y pulse on the1H channel. Throughout we are ignoring 1H chemical-shift evolution during the 1/(27) delays, because the 180° proton pulse in the center will refocus it. We will see other variants of this strategy later. By reversing the phase of the final 1H 90° pulse, we can do just the opposite: deliver a 180° 1H pulse to the

13C-bound protons while leaving the 12C-bound protons alone. If we change the 90° pulses to 45° pulses, we can selectively deliver an overall 90° 1H pulse to the 13C-bound protons alone or to the 12C-bound protons alone. This variant is called TANGO.

So the BIRD sequence is effectively a selective 180° 1H pulse that applies only to 12C-bound protons. The final version of our HETCOR pulse sequence is shown in Figure 9.12. This is beginning to look pretty complicated, but most of the details just have to deal with controlling what is refocused and what is allowed to evolve during the t1 period. The resulting spectrum will show only single peaks in F1 for each proton chemical shift, and single peaks in F2 for each 13C chemical shift (Fig. 9.7). This leads to a remarkable simplification of heavily overlapped regions of the proton spectrum. We will see in Chapter 11 that the HETCOR experiment is largely obsolete, replaced by the more sensitive inverse experiment HSQC, which correlates 1H to directly bound 13C by transferring in the reverse direction, from 13C (F1) to 1H (F2).


The 2D data matrix is just an array of numbers (intensities) arranged in rows and columns: 2048 columns and 1024 rows is typical. The numbers themselves can be positive or negative

Stack«] Piot Intensity Plot

Figure 9.13

Stack«] Piot Intensity Plot

Figure 9.13

and are typically very large (e.g., -2 x 109 to +2 x 109). How can we represent this "third dimension"—the intensity values—on a two-dimensional piece of paper or computer screen? We could use a stacked plot of 1D spectra (Fig. 9.13), with a horizontal offset to keep peaks from one spectrum from falling on top of the same peak in the next spectrum. Software can even "whitewash" the spectra behind a peak to make the peak stand out like a three-dimensional object. The nice thing about a stacked plot is that we can see the noise level, but for any degree of complexity it is not practical. An intensity plot is just a color code of each data element ("pixel") according to a color key at the side of the spectrum. Typically red and yellow colors are used for positive values and blue colors are used for negative values. This is very fast for a computer to display (Varian dconi command) but when we look at the details of a crosspeak it is difficult to see the fine structure. The standard for display of 2D spectra today is the contour map or contour plot. If you are a hiker, you are familiar with the topographic maps that show elevation as a "third dimension" of the map. If you imagine that the world is flooded to the level of 1000 ft. above sea level, for example, the shoreline is drawn on the map as the 1000 ft. contour. A mountain is represented as a series of concentric circles, successive shorelines around the "island" that would be left if the area were flooded. We can do the same thing to describe an NMR crosspeak, with contours showing positions of equal intensity. For example, we might connect all "pixels" or data cells that have intensities of 2 with the "2" contour line. All data values of 4 would be connected to create the next higher contour, and data values of 8 would be used for the next (Fig. 9.13). Because data values are digitized in discreet "boxes," we need to interpolate if we cannot find the exact value we are looking for. In topographic maps we use even intervals of elevations (e.g., 40 or 80 ft contour intervals), but in NMR we use a geometric series (e.g., 2, 4, 8, 16, 32) because it fits the Lorentzian lineshape better. A Lorentzian mountain would be a real challenge for hikers! For an NMR contour plot, we need to decide on a contour threshold (everything below this value is ignored), a number of contours and a contour interval or multiplier. If the threshold is 100 and the multiplier is 1.5, we have contour levels of 100, 150, 225, 338, 506, and so on, up to the number of contours. For some 2D data, negative values are important, so we might choose to show both negative and positive contours. On the computer screen, a typical display consists of 10 positive and 10 negative contours, with a multiplier of 1.20 between levels. If the threshold is 1000, we have contour levels of:

In white: 1000 1200 1440 1728 2074 2488 2986 3583 4300 5160 In red: -1000 -1200 -1440 -1728 -2074 -2488 -2986 -3583 -4300 -5160

A contour plot hides many evils. Because we can set the threshold as high as we want, we can eliminate impurities and artifacts without being accused of fraud! To be honest (and avoid missing important details) it is best to move the threshold down until a little bit of noise is visible. It's like trying to see a boat in a choppy ocean: a supertanker is easy to see without showing any of the waves, but if you want to find a canoe you'd better bring your threshold down until you see just the tops of the waves. Beginners in the world of 2D NMR always want to print out their 2D data on paper, but this is difficult because you really need to look at each part of the spectrum at several different contour thresholds, first displaying only the most intense peaks and then moving down until the noise is visible. Aligning peaks and comparing different 2D experiments is extremely difficult on paper. I remember a protein NMR lab in 1990 that had a room full of full-sized drafting tables, with researchers working on table-sized printouts of 2D NMR data using very sharp pencils to align crosspeaks. Soon afterward all of this was replaced with computer programs with sophisticated graphic displays that allow side-by-side display of different regions of the same 2D spectrum or corresponding regions of two different experiments, with correlated crosshairs that move simultaneously in both spectra under mouse control. Although the NMR instrument manufacturers all have software for 2D NMR data processing and analysis, a number of "third party" (neither Bruker nor Varian) programs are available that are far superior to software on the spectrometers: Felix (Accelerys, Inc.), MestRec, NUTS (Acorn NMR), and NMRpipe/NMRview (freeware) are examples.

9.3.2 2D Data Acquisition and Processing in General

The raw data from a 2D experiment consist of a series of FIDs, each acquired with a slightly longer t1 delay than the previous one. Varian creates an array of FIDs, with the t1 delay (parameter d2) arrayed (e.g., d2 = 0, 0.001, 0.002, 0.003, and so on). Bruker puts the FIDs together in a "serial file" (filename ser) and uses the parameter d0 (d-zero) for the ti delay in the pulse sequence, increasing it by the increment in0 with the pulse program command id0 (increment d-zero). Keep in mind that the heart of the 2D experiment is the transfer of magnetization from nucleus A to nucleus B during the mixing step. The first step in processing a 2D dataset is to Fourier-transform each of the FIDs in the array. The resulting spectra are loaded into a data matrix (like a spreadsheet) with the rows representing individual spectra in order of t1 value, with the smallest t1 value as the bottom row. The horizontal axis is labeled F2, which is the chemical shift observed directly in each FID, and the vertical axis is t1, the evolution delay. Each row in the 2D matrix represents a spectrum acquired with a different t1 delay, and each column in the matrix represents either noise (if the F2 value of that column is in a noise region of the spectrum) or, if F2 is the frequency of nucleus B (^b), the column is a t1 "FID" with maximum intensity at the bottom and oscillating in a decaying fashion as we move up to higher t1 values (Fig. 9.14). The frequency of this oscillation is just the chemical shift (^a) of nucleus A. Of course, a real sample has more than one peak in its spectrum, so there would be other columns containing different t1 FIDs.

The second step in processing the 2D data is to perform a second Fourier transform on each of the columns of the matrix. Most of columns will represent noise, but when we reach a column which falls on an F2 peak, transformation of the t1 FID gives a spectrum in F1, with a peak at the chemical shift of nucleus A (Fig. 9.15). The final 2D spectrum is a matrix of numerical values that has a pocket of intensity at the intersection of the horizontal line F1 = and the vertical line F2 = and has an overall intensity determined by the efficiency of transfer of magnetization from nucleus A to nucleus B. This efficiency tells us something about the relationship (J value or NOE intensity) between the two nuclei

Figure 9.14

within the molecule. Simultaneously with the process we described, other pairs of nuclei are undergoing the same evolution, mixing and detection process resulting in other crosspeaks at the intersections of the appropriate chemical-shift lines and with characteristic intensities representing the efficiency of transfer of magnetization. The 2D spectrum thus represents a complete map of all interactions that lead to magnetization transfer, with the participants in the interaction addressed by their chemical shifts.

The important concept here is the "labeling" of the magnetization with the chemical shift of nucleus A during the evolution period, and the subsequent unraveling of this information to link nucleus A to nucleus B. Let's look at this process in more detail. Each FID of the 2D experiment samples a single point in the indirect time domain t\, in the same way that the ordinary 1D FID is sampled at discreet, evenly spaced time points by the analog to digital converter (ADC) during a direct (real-time) acquisition. During the evolution (ti) period, the x! component of the nucleus A magnetization is a cosine or sine function with amplitude A and angular frequency Qa that decays due to dephasing of individual nuclei with time constant :

Now assume that the x component of nucleus A magnetization is transferred to nucleus B by the mixing sequence (most mixing schemes can only transfer one component of the magnetization). The transferred component might be z magnetization or antiphase magnetization or even DQC, but for simplicity we will use Mx. Magnetization transfer gives B magnetization whose intensity is modulated by the factor MX(ti) that depends on the offset of nucleus A and the value of ti at the moment of magnetization transfer. This B magnetization then precesses during the directly observed FID, inducing a signal in the probe coil. This signal is the normal FID of nucleus B, multiplied by the modulation (nucleus A) factor and the efficiency of transfer factor:

FID = A cos (^ati)exp(-ti/T2a) x Gab x cos(^bt2)exp(-t2/Tb)

where Gab is the efficiency of magnetization transfer (a function of Jab or rab, depending on the type of experiment), and t2 is the direct time domain of the FID. It is this function of multiplying the directly observed FID that "labels" the nucleus B information with the chemical shift of the original nucleus, A. If we sample the ti values in the same way we sample the t2 values, starting with zero and incrementing by a "dwell time" short enough to distinguish all of the frequencies expected, we have all the information needed to determine and ^b, the chemical shifts of nuclei A and B. This is all of the information we can expect to obtain from a 2D NMR experiment: the chemical shifts of each pair of nuclei involved and the intensity of their interaction. To get the information out we need to do two Fourier transforms: the first in t2, the second in t1.

Each individual FID is an oscillating and decaying function of t2, with the first two terms above equal to a constant. Fourier transformation gives a spectrum of B multiplied by the same constant:

Spectrum(ti, F2) = Acos(^ati)exp(-ti/7f) x Gab x Spectrum^ F2)

We have a different spectrum of B for each ti value, differing only in the value of the first term. For each column in the data matrix, we have a function of ti for a fixed value of F2 (Fig. 9.i4, right). Now the first term is the variable (function of ti) and the last term is a constant. Fourier transformation of the column converts the ti FID into a spectrum of A in the indirect frequency domain Fi:

Spectrum(Fi, F2) = Spectruma(Fi) x Gab x Spectrumb(F2)

For the data matrix, this means pulling out each column of the matrix in succession, treating it as an FID in ti, performing a Fourier transform and putting the resulting Fi spectrum back into the matrix at the same column position. When all the columns of the matrix (all the ti FIDs indexed by the frequency F2) have been transformed into A spectra, we have a 2D spectrum, which is a function in Fi and F2. If we fix F2 at the offset (chemical shift) of B (F2 = ^b), we are looking at a vertical slice through the crosspeak, which is Spectruma (Fi) (Fig. 9.i5, right). If we fix Fi at the offset of A (Fi = ^a), we have a horizontal slice through the crosspeak, which is SpectrumB (F2). The intensity of the crosspeak is determined by the efficiency of magnetization transfer Gab. Depending on the mixing scheme used (i.e., sequence of pulses and/or delays between evolution and detection) the selected relationship (interaction) might be a proximity in space (NOE) or a small number of bonds separating the two nuclei (J coupling).

9.3.3 Taxonomy of 2D NMR Experiments

By now you should be in the habit of thinking of a 2D NMR experiment as a transfer ("jump") of magnetization from nucleus A (the F1 frequency on the vertical axis) to nucleus B (the F2 frequency on the horizontal axis). All of the 2D experiments in use can be classified by the two types of nuclei detected in the direct (F2) and indirect (Fi) dimensions and the criteria for magnetization transfer during the mixing step. The mixing pulse sequences are designed to select for certain types of interactions between nuclei and can be divided into two categories: magnetization transfer based on a J-coupling interaction and magnetization transfer based on an NOE interaction. We can further divide the 2D experiments into homonuclear (experiments that transfer magnetization from one nucleus to another nucleus of the same type, usually 1H to 1H) and heteronuclear (experiments that transfer magnetization between two different types of nucleus, e.g., 1H and 13C).

Homonuclear Experiments (1H-1H)


F1 Nucleus


F2 Nucleus

0 0

Post a comment