1. TopSpin Training Course, 22nd November 2006

2. Overview Introduction to TopSpin Acquiring data through IconNMR
Processing data
Plotting/storing data
Interpretation – a worked example
Questions

3. What is the same?
Before looking at the new features it is worth mentioning what has not changed in TopSpin:
Parameters and Data format: you can still process TopSpin data with XWINNMR (not from AVII hardware)
Most text commands are either the same or aliased to the equivalent TopSpin command (i.e. you still type efp, …)
IconNMR has a few new features, but broadly remains the same – automation users will not need re-training, and most will never notice any difference!
XWINPLOT is essentially the same, so if you are already using it in xwinnmr you are ready for TopSpin Plot Editor!
At some future point (TS 2.?) the disk/data/user/nmr directory structure will be dropped.
The Avance II hardware (DRU) introduces some new digital filtration which unfortunately will not be understood by previous software versions during FT. Processed (already transformed) data will always be OK, and if required you can perform convdta to generate time-domain data in a back compatible format.

4. What is New? Integrated data browser
Command line history (and concatenation)
Configurable interface
Full support for handling multiple datasets
New interactive tools for Phasing, Integration, Peak Picking,…
Interactive guides (acquisition, processing, relaxation,…)
Acquisition status bar and tools
Extendable via AU programs, Python programs, etc. (including access to user interface)
…and much much more!

5. Acquiring Data Sample preparation is key
Selecting the right experiment:
What information do you need?
Which experiment(s) will give you this information?
Sample preparation: think about your poor spectroscopist and instrument!
use a nice clean sample tube, no broken tube tops
filter your sample to remove particulates
ensure a minimum of 4 cm sample depth, no bubbles, no floaters
wipe the sample tube and the spinner with clean tissue before placing on sample changer/lift
If you have trouble meeting these criteria talk to your spectroscopist who may be able to assist – they may have some tricks up their sleeve to help you!

6. IconNMR

7. Manipulating your experiment
Depending on the settings input by your service manager you may be able to manipulate the number of scans, for example.

8. Main types of experiment
Proton
Carbon
HMQC/
HSQC
COSY
HMBC
NOE 1H
13C
1H-13C correlation
1H-1H correlation
Long range 1H-13C correlation
Through space 1H-1H correlation
1 - 5 mg
5 – 25 mg
5 –10 mg
1 – 5 mg mg
1 – 5 mg
1 – 5 min
10 – 30 min
10 – 20 min
5 – 10 min
2 min – 6 hours* OH H C OH H C OH H C OH H OH
CH2
H2C
CH3 H H
*Depending on 1D or 2D methods being used..…
The amounts shown here are guidelines only and are based on MWt <1500. Some spectroscopists like to use the rule of thumb of ensuring there is 1mg of sample per carbon you are trying to observe for C13 experiments.
The time requirements are also very rough rules-of-thumb and assume gradient assisted techniques being used.

9. Example of a PROTON spectrum1 2 3 5 6 4 1H dq
J = 6.9, 15.5 Hz 1H dq
J =1.7, 15.5 Hz 2H q
J =7.1 Hz 3H dd
J =1.7, 6.9 Hz 3H t
J = 7.1 Hz
Five proton environments in ethyl crotonate - five distinct resonances are seen in the spectrum.
Integral ratio: 1 : 1 : 2 : 3 : 3
This immediately helps us assign the CH2 resonance, without even looking at the other evidence. There is only one CH2 in the molecule and therefore only one set of protons which the resonance at 4.18ppm can correspond to. There are two methyl groups, corresponding to the highest field resonances and the two alkene protons are in the lower field region, with respective integrals of 3 and 1 each. This is in keeping with chemical shift predictions, with the deshielding effects of electronegative atoms and p-electron clouds (ie: the alkene portion)
By analysing the splitting patterns and the coupling constants thereof, we can now easily assign the remaining resonances. Firstly, the triplet at 1.28ppm must be the methyl group 6, because it is only split into a triplet and with the same coupling constant by which the quartet at 4.18ppm (proton 5) is split. The other methyl resonance at 1.87ppm is split into a doublet of doublets, with a very small coupling constant and a larger one, more typical of alkyl chain coupling constants, so corresponding to proton 1 (being split by proton 2 with a coupling constant of 6.9 Hz and by proton 3 with a coupling constant of 1.7 Hz).
This leaves us with the two alkene resonances, which are both double quartets. One might expect a large coupling constant between the two olefinic protons themselves due to their torsion angle (Karplus equation!), and indeed this is seen with a coupling constant of 15.5Hz for the doublet element of the multiplets. The other coupling constants are then due to the methyl group 1, meaning that proton 2 must be the lower field resonance (6.97ppm) with a normal neighbouring alkyl coupling constant and the resonance at 5.84ppm corresponding to proton 3, which experiences a longer range (therefore smaller) coupling with methyl group 1. 2 3 1 5 6

10. Example of CARBON spectra4 5
1/6
2/3 M e O 1 2 3 5 6 4
DEPT stands for Distortionless Enhancement by Polarization Transfer. By running a standard C13 and one or more DEPT experiments you can uniquely distinguish between the different multiplicities of carbon (i.e: C, CH, CH2 or CH3). As a side benefit there is signal enhancement from DEPT that means it is more sensitive that a normal carbon, but of course only for proton-coupled carbons.
By far the most common edited C13 experiment performed is DEPT-135 (the number behind the DEPT phrase refers to the width of the refocusing pulse) which shows CH and CH3 peaks up and CH2 peaks down (note that phasing is relative – it can be set absolutely either by looking at the signal for TMS (or other known resonance), or by comparison to previously phased spectra). DEPT-90 will only show CH peaks,
Very occasionally, a signal may appear wrongly phased – this is usually attributable to a poor proton pulse angle. You could say that a DEPT spectrum is a good way of checking your probe tuning and pulse calibrations! If your pulse angles are fine, it is still possible for a peak to appear on the wrong side of the baseline, but only in very unusual circumstances (ie: an unusually large or small carbon-proton coupling constants for the signal you are trying to observe).
Six carbon environments in the molecule, six distinct resonances in the spectrum.
Note that the quaternary carbon at centre 4 is much reduced in intensity. This was discussed in the previous slide, but in summary, the quaternary carbon does not benefit from the nuclear Overhauser enhancement and it has a longer relaxation time than proton-attached carbon centres.
The olefinic carbon chemical shifts ( ppm) are consistent in that they give +ve signals in DEPT spectrum (C2 and C3). Methyl carbon centres (0-30 ppm) are consistent in that they give +ve signals in DEPT (C1 and C6).
An educated guess could be made as to the exact identity of each of the peaks, much along the same lines as the chemical shift information in the proton spectrum. It is likely that the higher chemical shift resonance corresponds to carbon 2 (144ppm) and the lower one to carbon 3 (123ppm), equally with the methyl resonances, it is likely that the most deshielded (higher ppm) corresponds to carbon 1, while the lower chemical shift resonance corresponds to carbon 6.
In order to “prove” this, we would need to run a carbon-proton correlation experiment (HSQC or HMQC). More on this later in the presentation.

11. Example of a COSY spectrum2 3 5 1 6 M e O 1 2 3 5 6 4
COrrelation SpectroscopY, COSY, is one of the most frequently used 2D NMR methods. It allows you to determine which protons are spin-spin coupled and hence, connectivity can be determined in a given system. It is very useful for providing an element of dispersion in crowded spectra by “spreading” into the second dimension: helping you assign which resonances are overlapping, if any.
A COSY spectrum will show two types of peak: diagonal peaks appear at the point of every chemical shift and can effectively be ignored. Cross peaks (any peaks off the diagonal) occur between spin coupled resonances and are the interesting peaks.

12. Example of an HSQC spectrum2 3 5 1 6 M e O 1 2 3 5 6 4
This spectrum shows simply the direct couplings between proton resonances (along the top) and carbon resonances (along the left side of the spectrum).
Proton-carbon correlation experiments are not only useful in assigning carbon resonances by virtue of an assigned proton spectrum, but like COSY, they can provide an element of dispersion in the proton dimension. Proton-carbon correlation is sometimes known as carbon-COSY and HETCOR as well as by the more modern names of HMQC and HSQC. Older versions of XWinNMR (pre 3.5) refer to INV4 experiments for HMQC and INVI for HSQC.
Most modern techniques rely on “inverse” detection, that is to say, even in the carbon dimension they are acquired by proton observation. Note that inverse techniques such as HMQC and HSQC do NOT require an inverse probe. (The geometry of an inverse probe is such that it is optimized for 1H observation, improving the sensitivity for protons but it is not essential).
Whilst it can be useful to have the carbon spectrum to project in the F1 dimension, it is not necessary to have run one and one can still obtain the carbon chemical shifts (at least for the nuclei with attached protons). This can be a useful time saving exercise, especially where samples are mass limited.

13. Example of an HMBC experiment2 3 5 1 6 M e O 1 2 3 5 6 4
Correlating coupled heteronuclear spins across multiple bonds.
The long range proton-correlation experiment is a powerful tool for linking together structural fragments. Where the long range coupling really comes into its own is when following a set of spins through a molecular sidechain or fragment of a molecule linked by a quaternary carbon. Like the HSQC or HMQC experiment, the HMBC experiment an inverse technique, so there is a sensitivity gain through observing in proton, rather than carbon. This means you can sometimes detect a quaternary carbon you might otherwise not have seen.
Since the HMBC experiment is fundamentally based on coupling constants between carbon and proton, the results should always be approached with some caution: not seeing a signal in the spectrum does not necessarily mean there is no long range correlation. By adjusting the long range coupling constant “cnst13” (typical range 4-15Hz) in the acquisition parameters and running the experiment again, other correlations may be seen.

14. An introduction to processing
Most spectra run through IconNMR will be processed
(and perhaps printed) already
You could simply open the processed spectrum and look at the data
You could reprocess - fundamentally:
multiplier, Fourier transform and phase
(more on this later)
Then you can integrate, peak pick, calibrate etc.

15. File handling, print, copy/paste, last 1D/2D/3D
Data Display Buttons
Hz or ppm
datasets
Y-scale
Measure distance
File handling, print, copy/paste, last 1D/2D/3D
grid
See later…
All
Last expansion
Retain scale
Horizontal position
Vertical position
Vertical scale
Horizontal scale
Buttons seen will depend upon data-type (1D or 2D), and any customisation!

16. Processing 1D data Processing guide (automatic/manual)
Processing -> Data Processing Guide
Command line automatic: “xaup” – this will probably also print a spectrum
Command line manual: em, ft, apk/apks
Button: if you have one set up….
…of course you could set one up…

17. Processing 2D data Processing guide (automatic/manual)
Processing -> Data Processing Guide
Command line automatic: “xaup” – this will probably also print a spectrum
Command line manual: xfb, abs1, abs2
Button: if you have one set up….
… of course you could set one up …

18. Other Buttons – interactive menus
calibration
peak picking
multiple display
zoom overview
integration
baseline correction
phase correction

19. Output “screen dump” Printer icon File -> Print Ctrl P
Data output onto paper or into transferable files such as PDF, JPEG or WMF/EMF is best handled through the TopSpin Plot Editor.
Whilst perhaps not immediately intuitive, it is a powerful tool for making data look presentable for hard copy or data transfer. It is worth getting to grips with and should be seen as a graphical interface only. I.e.: ensure all data processing is done in TopSpin, before opening the Plot Editor.

20. Worked example 1H spectrum DMSO
In order to see the actual use of all the methods discussed, we are going to work through an example. When faced with a proton spectrum as complex as this one it can be difficult to know where to start. The molecule camphor gives rise to such complex proton spectrum due to the rigidity of the ring system, causing each proton to be in a unique chemical environment and therefore each appears as an individual peak, split by all its neighbouring protons.
All we can say with certainty is that the three singlets correspond to the three methyl groups (A,B and C), and we could make a guess at the “triplet” at 2.05ppm corresponding to proton 1, purely down to its splitting pattern. To confirm this, we need to look at more spectral data.
DMSO

21. HSQC CH
CH3
CH2
This HSQC spectrum has been presented with the DEPT-135 spectrum in the F1 dimension. Along the top, we see the associated proton spectrum. The DEPT spectrum is phased such that CH and CH3 signals are pointing left and the CH2 signals are pointing right.
Immediately obvious from the DEPT spectrum is that there are 3 CH2’s in this system. The HSQC spectrum allows us to quickly assign which of these protons are CH2 “pairs” by virtue of the fact they are attached to the same carbon atom. The protons at 1.3 and 1.7ppm are obviously one pair, and the signals at 1.25 and 1.65 ppm are another. The remaining pair is slightly more difficult to assign, with the CH and CH2 carbon resonances being so close together.

22. HSQC expansion 1 CH
CH3
CH2
Looking at an expansion of this region will reveal that it is the 2.3 and 1.8 ppm resonances which are the CH2 pair. And we can confirm our suspicion that the 2.05ppm resonance is the CH.
At this stage we could hazard an educated guess as to the identity of the two protons responsible for the CH2 pair seen here, simply due to the multiplicity. Looking at the molecule there is only one proton which could appear as a doublet, which is proton 2, because of the torsional angle it is at relative to proton 1. Since it is at virtually 90deg it experiences no coupling from it. Therefore the only coupling it experiences is from its geminal proton 3.
However, it would be nice to see this confirmed in a proton-proton correlation (COSY). What we would expect to see is a strong proton – proton correlation between proton 1 and the multiplet at 2.3ppm, and also a strong correlation between the multiplet and the doublet at 1.8ppm.

23. COSY 1 6 7 2 3 5 4
This COSY spectrum is just looking at the regions in which the multiplets appear. The three methyl singlets at low chemical shifts have been left off for clarity. The lines linking resonances on the proton spectrum are the CH2 pairs we identified earlier.
Starting from the single CH peak (proton 1) we identified from the proton-carbon correlation experiment, we can clearly see there are two strong proton-proton correlations in this spectrum. You can trace the correlation from the diagonal up and down and across to the left hand projection, or across in both directions and up to the top projection.
Either way, you see a correlation to the multiplets at 2.3ppm and 1.9ppm. This means they must correspond to protons 3 and 5.
In other words, our assertion that the multiplet at 2.3ppm corresponds to proton 3 is now confirmed, and we also see the strong correlation from here to the doublet at 1.8ppm. So, we can now assign protons 2 and 3.
Going back to the correlation between proton 1 and the multiplet at 1.9ppm: the direct correlation must mean that the resonance at 1.9ppm is due to proton 5, and since we have identified its geminal proton to be the left hand side of the resonance at 1.3ppm, this is proton 4.
We are now left with two remaining signals which are protons 6 and 7. The correlation from proton 5 the strongest to the resonance at 1.65ppm, meaning this must be proton 6. In this case the dihedral angle between protons 5 and 6 is virtually 0, therefore a large coupling is observed, whereas proton 7 is at virtually 90º to proton 5 and therefore we see virtually no correlation.
Now that we have assigned all the peaks around the ring, we can turn our attention to the methyl resonances, protons A, B and C. There are a number of ways to do this: we discussed earlier the possibility of running long-range carbon-proton correlations, which may give us the answers we are looking for. Or, we could go down the route of establishing which protons are close in space, by using NOE methods.

24. HMBC 3 1 6 C 5 4 7 2 CH
CH3
CH2 C
Starting with the HMBC spectrum, we can immediately see it is complicated. Understandably, there are many long range correlations in a molecule such as this, and we need to approach the spectrum with some care.
In this case the projection along the left hand side is a DEPTq spectrum, and as you can see two new signals have appeared relative to our earlier projection on the HSQC spectrum. The chemical shift region in which they appear strongly suggests they are due to the two aliphatic quaternary carbon atoms in the molecule, not the ketone carbon. This we would expect to see at a chemical shift around 200ppm and is not shown on this spectrum.
Many of the long range correlations on this spectrum tell us little about which of the methyl groups is which. For example, the correlation between proton resonance 2 and the lowest chemical shift methyl carbon could be either A or C, both of which are three bond correlations. However, if we look at the correlation between the middle methyl proton resonance and the CH2 at ~28ppm: we know, working backwards from our assigned proton spectrum and the HSQC this resonance relates to the carbon atom to which protons 6 and 7 are attached. Therefore, we can say reasonably confidently the middle methyl proton resonance is due to methyl group C.

25. 2D NOESY 2 C 3 1 5 6 4 7
This is a typical 2D NOESY spectrum you might expect to see, where there is a great deal of information, not all of it useful. Generally you will find the spectra quite noisy and it can be difficult to pick out the peaks you are looking for. A 2D NOESY spectrum will almost always require manual post-processing in order to get the information required.
In this case, we can see some clear differences in the through-space correlations made by the methyl protons. Looking at the resonance which we assigned to methyl group C, we can see there is indeed a through space interaction with proton 6, but with little else. To see in more detail we can zoom in on the region of interest.

26. 2D NOESY expansion 1 2 3 5 6 B C A
What is clearly visible here is that the lowest chemical shift methyl resonance has through-space interactions with protons 3 and 1, whereas the higher shift one at ca. 0.9ppm shows a spatial interaction with protons 1, 5 and 6. This clearly points at methyl group A being the resonance at 1.9ppm and the resonance at 0.75ppm being methyl group B.
This molecule has given us a nice example of how all the techniques we have mentioned in today’s talk can be used to obtain a full assignment of the spectra. In particular, this NOESY spectrum has given us the final answers with regard to the more difficult to identify resonances. This spectrum took 6.5 hours to accumulate and took some careful processing and phasing, to be able to see the relevant information.

27. 1D selective NOESY C B 1 A 3 5 2 6 4 7
On this stacked plot we can see the normal proton spectrum on the top, and two selective 1D NOE experiments below it. In the middle spectrum, the methyl resonance at 0.9ppm has been irradiated, as indicated by the inversion of the peak. The effect of this irradiation on the surrounding peaks is clearly visible: a positive NOE effect is seen on protons 1, 5 and 6. The inversion of proton 3 (bottom spectrum) only shows a positive NOE effect on the lowest chemical shift methyl resonance.
This leads us to conclude that the resonance at 0.9ppm corresponds to methyl group A and the resonance at 0.76ppm is due to methyl group B.
This is the same conclusion the 2D NOESY spectrum led us to, but these spectra took a maximum of 10 minutes to collect.

28. Interpretation summary
Techniques
13C NMR
DEPT variations
1H NMR
COSY
HSQC/HMQC
HMBC
NOESY (1D and 2D)
Key Uses
Chemical environment, carbon count
Multiplicity determination
Chemical environment, quantification
Spin-spin coupled nuclei
Directly bonded 1H and 13C
1H and 13C correlation across multiple bonds
Interactions through space

29. Summary
Think about data you need … don’t waste instrument time by submitting every sample for all experiments
Remember sometimes you can obtain C13 chemical shifts from combination of 2D spectra when sample is limited
Use processing guides !!
Remember you can create a button for tasks you often repeat
“Save” is implicit in TopSpin – any processing changes you make will be remembered for next time
“Prnt” gives you quick and dirty screen dump
Plot Editor gives you pretty spectra for hard/electronic copy
………….any questions??