Logic Puzzle #11: Pairwise Correlation Confidence … Solution

Long range 2D NMR experiments do not necessarily provide information about all the connectivities. The following structure elucidation problem set is one such example.

Based on the 1H-13C HMBC shown below, there is no evident correlation between the 1H at 5.31 ppm and the 13C at 21.1 ppm. Note the green box describes the region of interest.

The correlation between the methyl 1H at 2.15 ppm and 13C at 84.7 ppm (indicated by the purple arrow and the red bonds below) is a weak 4J coupling (also denoted as a W-coupling or M-coupling). The coupling arises from the individual methyl protons rotating and interacting through a W-relationship (or M-)with the carbon. This coupling is not evident for the methine 1H at 5.31 ppm and the 13C at 21.1 ppm.

 
Thank you Serge for your comment.

Logic Puzzle #11: Pairwise Correlation Confidence

Typical for long range 2D NMR experiments, spectral data may exhibit more than one correlation for two coupled nuclei (e.g. A to B and B to A). The pairwise correlations offer an extra degree of confidence in the interpretation.  

For the following fragment, an 1H-13C HMBC correlation exists for the 1H 2.15 ppm to 13C 84.7 ppm (represented by the purple arrow in the diagram below).

 
Is the pairwise correlation for the 1H 5.31 ppm to 13C 21.1 ppm evident?

 
A special thanks goes to Gene M. for pointing me to the data.

Logic Puzzle #10: Deciphering the Fragment Pattern using 2D NMR Data … Solution

Like any new process, it takes some practice to extract, understand and convert the information presented from a set of experimental NMR datasets into a fragment.

The only fragment that can accommodate the set of restrictions from a 1H-13C HSQC and HMBC is 2,3-dimethylbutane-1,1-diyl. The green arrows illustrate the 2-3JCH coupling responses extracted from an HMBC experiment.

 

Logic Puzzle #10: Deciphering the Fragment Pattern using 2D NMR Data

The goal of this puzzle is to conceptualize a fragment(s) from the given information.

In the following example, a set of protonated sp3 carbons were extracted from an HSQC experiment (not shown). The green arrows represent the 2-3JCH coupling responses extracted from an HMBC experiment. Based on these restrictions, what fragment(s) supports the data?

Note there is an open valence off one of the carbon atoms.

Logic Puzzle #7: Almost Missed It … Solution 2

With intense solvent signals present on a spectrum, a smaller signal(s) can easily be missed. If 2D NMR data is available, then this extra information can assist in clarifying whether a small signal(s) is obscured by larger signals.

On the 1H-13C HMBC below, the correlations for CDCl3/CHCl3 (due to 1J coupling responses and more) are more intense in comparison to the weak correlation at approximately 6.9 and 77.3 ppm. In this case, the weak correlation is attributed to a quaternary carbon obscured by a set of intense solvent signals.

 

Logic Puzzle #2: How to link 3 Fragments

The goal of this puzzle is to logically combine a set of fragments using valence and NMR information.

In this puzzle, three fragments are correlated through 2-3J coupling responses (represented by a green arrow) that were extracted from a 1H-13C HMBC data (spectrum not shown). The carbon atoms with the 13C chemical shifts displayed in blue indicate the presence of an adjacent heteroatom. Based on these criteria, what 'complete' fragment(s) supports the data and is there anything missing?

In order to accommodate these restrictions, a logical fit is to consider a trivalent atom, e.g. nitrogen.

Logic Puzzle #2: How to link 3 Fragments

The goal of this puzzle is to logically combine a set of fragments using valence and NMR information.

In this puzzle, three fragments are correlated through 2-3J coupling responses (represented by a green arrow) that were extracted from a 1H-13C HMBC data (spectrum not shown). The carbon atoms with the 13C chemical shifts displayed in blue indicate the presence of an adjacent heteroatom. Based on these criteria, what 'complete' fragment(s) supports the data and is there anything missing?

In order to accommodate these restrictions, a logical fit is to consider a trivalent atom, e.g. nitrogen.

Logic Puzzle #1: The Missing Link

A great skill to master is the capability to conceptualize a fragment or structure directly off a spectrum without resorting to paper-and-pen work. This skill is learnt through lots of practice. Whenever partial information is available, an elucidator can conjure up a mental image of possibilities and should it be required instinctively hunt for any missing data.

In the following example, a set of fragments including 13C and 1H chemical shifts and long-range coupling information were extracted from an HMBC experiment (not shown). The green arrows represent the 2-3J coupling responses between the 3 equivalent methyl groups and the carbonyl’s quaternary carbon. Based on these restrictions, what fragment(s) support the data and is there anything missing?

To accommodate these restrictions, three potential fragments, assigned A, B and C, are shown below. Fragment A can be disregarded on the basis of the carbon valence. Fragment B is not a good candidate because the CH3 chemical shifts do not support the presence of an adjacent heteroatom. Fragment C seems to be the most logical choice. However, there is a missing quaternary carbon. The next step is to re-evaluate the NMR data in search of a weak 13C signal at ~40 ppm.

Logic Puzzle #1: The Missing Link

A great skill to master is the capability to conceptualize a fragment or structure directly off a spectrum without resorting to paper-and-pen work. This skill is learnt through lots of practice. Whenever partial information is available, an elucidator can conjure up a mental image of possibilities and should it be required instinctively hunt for any missing data.

In the following example, a set of fragments including 13C and 1H chemical shifts and long-range coupling information were extracted from an HMBC experiment (not shown). The green arrows represent the 2-3J coupling responses between the 3 equivalent methyl groups and the carbonyl’s quaternary carbon. Based on these restrictions, what fragment(s) support the data and is there anything missing?

To accommodate these restrictions, three potential fragments, assigned A, B and C, are shown below. Fragment A can be disregarded on the basis of the carbon valence. Fragment B is not a good candidate because the CH3 chemical shifts do not support the presence of an adjacent heteroatom. Fragment C seems to be the most logical choice. However, there is a missing quaternary carbon. The next step is to re-evaluate the NMR data in search of a weak 13C signal at ~40 ppm.

Will the correct structure please stand up? … Part 2

Part 1 presented a challenge to determine an experiment to distinguish two very similar products from each other, namely 3-methyl-5-(pyridin-2-yloxy)pyridine and 5'-methyl-2H-1,3'-bipyridin-2-one. The products have identical formula weights and the LC/MS and 1H NMR are too similar to draw any conclusion from.

 

The first step is to determine what is different between the two products and then identify an experiment specifically designed to focus on that difference. The obvious difference between the two products is the position of the oxygen atom—an ester group verse a carbonyl group. An FT-IR experiment, as commented by the reader Felipe A., can be used to sort out the products.

 

Other experiments can include the use of reducing agents, 15N NMR, 1H -13C HMBC, 1D NOE, 1H-1H TOCSY, MS2, etc. Note free water, acids and sample concentration can inhibit the use of some of these experiments.

 

A 13C NMR experiment may appear to be another good choice when trying to identify a carbonyl group. However, the carbonyl is part of a conjugated system and so the 13C chemical shift is expected around 160 ppm, which also happens to be expected for the 13C chemical shift of the O-C=N group on the other product.

Will the correct structure please stand up? … Part 2

Part 1 presented a challenge to determine an experiment to distinguish two very similar products from each other, namely 3-methyl-5-(pyridin-2-yloxy)pyridine and 5'-methyl-2H-1,3'-bipyridin-2-one. The products have identical formula weights and the LC/MS and 1H NMR are too similar to draw any conclusion from.

 

The first step is to determine what is different between the two products and then identify an experiment specifically designed to focus on that difference. The obvious difference between the two products is the position of the oxygen atom—an ester group verse a carbonyl group. An FT-IR experiment, as commented by the reader Felipe A., can be used to sort out the products.

 

Other experiments can include the use of reducing agents, 15N NMR, 1H -13C HMBC, 1D NOE, 1H-1H TOCSY, MS2, etc. Note free water, acids and sample concentration can inhibit the use of some of these experiments.

 

A 13C NMR experiment may appear to be another good choice when trying to identify a carbonyl group. However, the carbonyl is part of a conjugated system and so the 13C chemical shift is expected around 160 ppm, which also happens to be expected for the 13C chemical shift of the O-C=N group on the other product.

Correlating 1D NMR data on an HMBC

There are many challenges present in deciphering correlations on an HMBC experiment. Barring any data collection issues, most of the challenges will arise from signals that are overlapping, thus, adding an element of ambiguity.   

The 1H -13C HMBC spectrum below shows a correlation between the 13C at 71.2 ppm and the overlapping 1H signals at 2.08 (d) and 2.12 (m) ppm. There is nothing usually here, right?

A nearby correlation on the same 1H -13C HMBC spectrum (see below) shows a correlation between the 13C at 75.01 ppm and the 1H at 2.28 ppm. The splitting pattern for this correlation suggests that the previous correlation might also be splitting. If this is the case, then the correlation is centered on 71.26 and 2.10 (m) ppm, and thus, there is no correlation to the doublet at 2.08 ppm. In the end, further analysis of the overlapping multiplets between 2.00-2.15 ppm is needed to resolve the ambiguity.

Correlating 1D NMR data on an HMBC

There are many challenges present in deciphering correlations on an HMBC experiment. Barring any data collection issues, most of the challenges will arise from signals that are overlapping, thus, adding an element of ambiguity.   

The 1H -13C HMBC spectrum below shows a correlation between the 13C at 71.2 ppm and the overlapping 1H signals at 2.08 (d) and 2.12 (m) ppm. There is nothing usually here, right?

A nearby correlation on the same 1H -13C HMBC spectrum (see below) shows a correlation between the 13C at 75.01 ppm and the 1H at 2.28 ppm. The splitting pattern for this correlation suggests that the previous correlation might also be splitting. If this is the case, then the correlation is centered on 71.26 and 2.10 (m) ppm, and thus, there is no correlation to the doublet at 2.08 ppm. In the end, further analysis of the overlapping multiplets between 2.00-2.15 ppm is needed to resolve the ambiguity.

Distinguishing Impurities … Conclusion

In the blog series entitled Distinguishing Impurities, several types of NMR experiments were outlined to offer some insight into what to be on the look out for. Although there is no set guideline in this process, the best advice I may offer is to keep track of all the solvents, be aware of the possibility for sample degradation or side products, and finally if a signal does not look right, trial and error may be the only safe bet to ensure nothing is overlooked.

Below is a summary of the each part in the series:

Part 1 covered the aspects a 1H NMR whereby 2 singlets seemed out of place in relation to the remaining 1H signals.

Part 2 described some prep work to check for differences in the integrals of the 1H signals.

Part 3 depicted a 1H-13C HSQC with no obvious indicators that any of the 1H signals were impurities.

Part 4 illustrated how a 1H-1H DQF COSY was a bad choice when dealing with 1H singlets.

Part 5 had a 1H-1H TOCSY spectrum lacking any long range correlations with the 1H singlets.

Part 6 explained how the 1H-13C HMBC lacked any correlation trails with the 1H singlets.

Distinguishing Impurities … Conclusion

In the blog series entitled Distinguishing Impurities, several types of NMR experiments were outlined to offer some insight into what to be on the look out for. Although there is no set guideline in this process, the best advice I may offer is to keep track of all the solvents, be aware of the possibility for sample degradation or side products, and finally if a signal does not look right, trial and error may be the only safe bet to ensure nothing is overlooked.

Below is a summary of the each part in the series:

Part 1 covered the aspects a 1H NMR whereby 2 singlets seemed out of place in relation to the remaining 1H signals.

Part 2 described some prep work to check for differences in the integrals of the 1H signals.

Part 3 depicted a 1H-13C HSQC with no obvious indicators that any of the 1H signals were impurities.

Part 4 illustrated how a 1H-1H DQF COSY was a bad choice when dealing with 1H singlets.

Part 5 had a 1H-1H TOCSY spectrum lacking any long range correlations with the 1H singlets.

Part 6 explained how the 1H-13C HMBC lacked any correlation trails with the 1H singlets.

Distinguishing Impurities … Part 6

Although a 1H-13C HMBC experiment may take a long time to collect with adequate signal-to-noise, it can offer an idea as to which signals belong 'together'. Generally, signals belonging to the same structure leave a trail of connectivity information via long range correlations. If the atoms belong to the same structure, then a typical HMBC pattern is as follows: proton A is correlated to carbon B, carbon B is correlated to proton C, proton C is correlated to carbon D, etc.

The 1H-13C HMBC experiment below shows a lack of long range connectivity information for the singlets at 1.68 and 2.14 ppm to any of the other signals. This leaves two possibilities: either the singlets do not pertain to the main unknown structure or the protons are separated far enough from any carbon atoms to show any correlations.

Distinguishing Impurities … Part 6

Although a 1H-13C HMBC experiment may take a long time to collect with adequate signal-to-noise, it can offer an idea as to which signals belong 'together'. Generally, signals belonging to the same structure leave a trail of connectivity information via long range correlations. If the atoms belong to the same structure, then a typical HMBC pattern is as follows: proton A is correlated to carbon B, carbon B is correlated to proton C, proton C is correlated to carbon D, etc.

The 1H-13C HMBC experiment below shows a lack of long range connectivity information for the singlets at 1.68 and 2.14 ppm to any of the other signals. This leaves two possibilities: either the singlets do not pertain to the main unknown structure or the protons are separated far enough from any carbon atoms to show any correlations.

Identifying Peak Overlap on an HMBC Spectrum … Part 2

One of the trickiest parts of interpreting a 1H-13C HMBC is deciding whether a 13C resonance is coinciding with another 13C resonance, i.e. overlapping 13C peaks. A past blog, Part 1, describes a specific case where the possibility of two coinciding 13C resonances can be deciphered based on a high carbon correlation count. Herein, we present a specific case for a 1H-13C HMBC with paired 1J responses to distinguish peak overlap.

The 1H -13C HMBC spectrum below illustrates 2 protons multiplets at 1.89 and 2.37 ppm and 2 carbon resonances at 26.8 and 43.0 ppm. The paired 1J responses at 43.0 ppm indicate a CH2 group with the proton multiplet at 1.89 ppm. Similarly, the carbon at 26.8 ppm shows a paired 1J response with the proton multiplet at 2.37 ppm (CH group). In addition, a single correlation at exactly 2.37, 26.8 ppm indicates a quaternary carbon coinciding with the carbon resonance for the CH group.

The quaternary carbon and CH group are in the vicinity of each other.

Identifying Peak Overlap on an HMBC Spectrum … Part 2

One of the trickiest parts of interpreting a 1H-13C HMBC is deciding whether a 13C resonance is coinciding with another 13C resonance, i.e. overlapping 13C peaks. A past blog, Part 1, describes a specific case where the possibility of two coinciding 13C resonances can be deciphered based on a high carbon correlation count. Herein, we present a specific case for a 1H-13C HMBC with paired 1J responses to distinguish peak overlap.

The 1H -13C HMBC spectrum below illustrates 2 protons multiplets at 1.89 and 2.37 ppm and 2 carbon resonances at 26.8 and 43.0 ppm. The paired 1J responses at 43.0 ppm indicate a CH2 group with the proton multiplet at 1.89 ppm. Similarly, the carbon at 26.8 ppm shows a paired 1J response with the proton multiplet at 2.37 ppm (CH group). In addition, a single correlation at exactly 2.37, 26.8 ppm indicates a quaternary carbon coinciding with the carbon resonance for the CH group.

The quaternary carbon and CH group are in the vicinity of each other.

Limiting the Atom Ranges

In a previous blog, the atoms C, H, O, and N were limited to a specific range, 0-50, 0-100, 0-10 and 0-10, respectively. Starting from a wide atom count range ensures molecular formulae are not overlooked. The ranges for the atom count can be restricted further using additional information such as NMR, IR, etc.

A common tactic for elucidating compounds containing nitrogen is to acquire a 15N NMR experiment to assist in setting a narrower range for nitrogen count. The 1H-15N HMBC experiment below indicates the presence of at least 2 nitrogen atoms. Therefore, the nitrogen range can be set to 2-10 atoms and thus reduce the number of candidate molecular formulae.