Friday, December 21, 2007

How Can I Get a Quantitative 13C NMR Spectrum?

With conventional acquisition parameters, 13C NMR spectra of liquids are rarely quantitative and usually not integrated. There are two main reasons for this. Proton decoupling is normally used to simplify the spectrum and as a way to build up sensitivity as a result of the nuclear Overhauser effect. Also, typical T1's for 13C are very long. One can still get quantitative data and a proton decoupled spectrum if a very long recycle delay (perhaps 5 minutes) is used and the 1H decoupling is applied only during the acquisition time which is very short with respect to the recycle delay. Using such parameters however makes the collection of data prohibitively time consuming. To avoid this problem, you can add a paramagnetic relaxation agent to the sample to reduce the relaxation times.

In the spectra below the data were collected with a 10 second recycle delay. In the left hand spectrum, the duty cycle of the decoupler was 100 % and the spectrum exhibits an NOE. In the middle spectrum, the NOE is greatly reduced by decoupling only during the acquisition time (decoupling duty cycle = 10 %), however the long T1's prevent the spectrum from being quantitative. The spectrum on the right was acquired similar to that of the one in the middle except some Cr(acac)3 was added to the sample to reduce the relaxation times for the carbons. This spectrum is quantitative.

Thursday, December 20, 2007

Improving Your Baseline Correction

Spectral baselines are often corrected in Bruker's software by using the "abs" command after Fourier transformation. This correction is not always well suited to the data. Another method of getting flatter baselines is to define specific frequency limits "absf1" and "absf2" over which the baseline will be corrected and then use the "absf" command rather than the "abs" command. The smaller the difference between "absf1" and "absf2", the better the correction will be. Below is an example where "absf1" and "absf2" are defined by the region in pink.

Wednesday, December 19, 2007

31P - 109Ag HMQC

Silver has two spin I = 1/2 isotopes (109Ag and 107Ag) with natural abundances of 48.18% and 51.82%, respectively. Although 109Ag has a lower natural abundance, it is the preferred isotope to observe as its gyromagnetic ratio is slightly higher than that of 107Ag. Direct observation of 109Ag is problematic in that the Larmor frequency is low, the NOE's to protons are negative and most importantly, the T1's are very long (minutes to hours). In some silver phosphine complexes with 109Ag - 31P coupling, the problems of observing silver directly can be circumvented by using indirect detection via a 31P detected 31P - 109Ag HMQC. The advantages of using this approach are that the directly observed 31P frequency is 8.7 times higher than that of 109Ag and the relaxation delay in the experiment depends on the 31P T1's which tend to be orders of magnitude shorter than those of 109Ag.
Below is an example of a 31P detected 31P - 109Ag HMQC for a silver phosphine complex. These data were acquired without 109Ag decoupling. In the 31P trace on the top of the spectrum one can see a single 31P resonance with coupling to both 109Ag and 107Ag. The 109Ag is indirectly observed in the F1 domain.

Monday, December 17, 2007

Better 180 Degree Pulses

Many pulse sequences use 180 degree pulses. Standard hard pulses are not perfect at inverting all of the lines in a spectrum with a wide spectra width. In such cases a better choice may be a shaped adiabatic pulse. A comparison is shown below for the 13C spectrum of ethyl acetate. Each spectrum was acquired with the same number of scans. The trace on the left was acquired with a hard 180 degree pulse and the spectrum on the right was acquired with the CRP60comp.4 pulse from the Bruker pulse shape library. Note the better performance resulting in improved nulls.

Wednesday, December 12, 2007

Collecting 13C NMR Spectra of Highly Paramagnetic Compounds

I am sometimes asked to acquire 13C NMR spectra of highly paramagnetic compounds. Those of you attempting to acquire such spectra for the first time should be aware of the following:
1. You should use a large spectra width as the 13C chemical shifts can be very extreme. In order to get a wide uniform excitation, use very short high power pulses.
2. Both the T1's and the T2's are short, so you can use short acquisition times and short recycle delays.
3. The 1H chemical shift range for paramagnetic compounds can span hundreds of ppm. The 1H decoupling schemes used to collect 13C data do not permit such a broad decoupling bandwidth and you may have to collect several 13C spectra with different 1H offsets. It is always best to collect a proton spectrum first to evaluate the 1H decoupling needs in the 13C spectrum.
4. The lines are often very broad and therefore many scans are needed to build up the signal-to-noise ratio.
5. The 2H lock signal may be shifted by the paramagnetic compound, so don't be surprised if your automatic locking routines do not work.

Tuesday, December 11, 2007

13C - 14N J coupling

It has always puzzled me how students so willingly accept the fact that 13C is J coupled to a spin I = 1 nucleus like 2H (eg. the multiplets for the 13C spectra of deuterated solvents) however they do not question the fact that 13C almost never shows a J coupling to 14N (another spin I = 1 nucleus). The reason that 13C exhibits J coupling to 2H is that the relaxation among the three energy levels of 2H is rather slow and each 13C "sees" the 2H in each of its three Zeeman states, splitting the 13C resonance into 3 lines of equal intensity. The efficiency of the relaxation among the energy levels of quadrupolar nuclei (among other things) depends on the magnitude of the quadrupolar coupling constant - the larger the quadrupolar coupling constant, the faster the relaxation. Unlike 2H, which has a very small quadrupolar coupling constant, 14N (and most other quadrupolar nuclei) has a substantial quadrupolar coupling constant and therefore the relaxation among its energy levels is very fast. The 13C therefore "sees" the 14N in a single "average" state and as a result is a singlet. There are however compounds where the 14N is in a very symmetric environment. The high symmetry make the quadrupolar coupling constant much smaller and the relaxation therefore much slower. In these cases one can observe the 13C - 14N J coupling. Depending on the relaxation of the 14N, the 13C lines can vary from being a sharp 1:1:1 triplet, to a broad unresolved triplet, to a very sharp singlet. Below is an example of a case where the 14N is in a very symmetric environment and the 13C - 14N J couplings are resolved.

Monday, December 10, 2007

Enhancing 29Si NMR Spectra with DEPT

Measuring 29Si NMR spectra of liquids can be very time consuming as 29Si nuclei tend to have very long T1 relaxation times and therefore very long delays must be left between scans to allow for relaxation. Also, since 29Si has a negative gyromagnetic ratio, the NOE's with protons diminish the signals and inverse gated decoupling should be used. For molecules where there is a measurable 29Si - 1H J coupling, one can use a DEPT sequence to enhance the 29Si signal. The advantages are two-fold: first, there is an enhancement on every scan from the DEPT sequence and second, the repetition rate of the experiment depends on the 1H T1 rather than the 29Si T1. The second advantage is very significant as the 1H T1's can be shorter than the 29Si T1's by two orders of magnitude. The example below demonstrates the advantages.

The traces on the left and right were both acquired in 1 minute with a 2 second recycle delay. One can see the tremendous signal-to-noise-ratio advantage of using a DEPT sequence. For comparison, the trace in the middle was acquired in 16.4 minutes with inverse gated decoupling and a 60 second recycle delay. (Note that the DEPT signal could have been enhanced even more if a 35 degree rather than a 24 degree flip angle was used for the final 1H pulse in the DEPT sequence)

Friday, December 7, 2007

Shaped Pulses for Selective Excitation

Short rectangular pulses have broad excitation profiles (see BLOG entry for November 20, 2007) and are used for non-selective uniform excitation. These are often called "hard" pulses. Longer low power shaped pulses are used to excite specific spectral regions for selective excitation. An example of this is shown below.These are often called "soft" pulses.

Thursday, December 6, 2007

Removing t1 Noise from Heteronuclear Correlation Data

Heteronuclear correlation data can often contain vertical stripes of noise, called t1 noise. This noise can be reduced by taking the projection a group of rows containing no signals (only noise) and subtracting the projection from every row of the data. This is illustrated below for an HMQC spectrum. The spectrum on the right is the data after Fourier transformation. The data contains a considerable amount of t1 noise. All of the rows highlighted in yellow, containing only noise, were used to prepare a projection. The projection was subtracted from every row of the data. The result is the spectrum on the right. The t1 noise is much reduced.

Wednesday, December 5, 2007

A New BLOG Dedicated to Processing and Predicting NMR Data

Carlos Cobas, president and co-founder MESTRELAB RESEARCH has started an excellent BLOG on processing and predicting NMR data. Given Carlos' expertise in this area, it promises to be a very valuable resource to the NMR community.

Visit Carlos's BlOG here.

http://nmr-analysis.blogspot.com/

Tuesday, December 4, 2007

Sensitivity Improvement from Cross Polarization

Yesterday's post showed the pulse scheme for cross polarization in solids. Below is a comparison of the single scan 39.7 MHz 29Si MAS spectra for solid tetrakis(trimethylsilyl)silane. The trace on the left was acquired with a simple one pulse sequence with high power CW proton decoupling during the acquisition (Bloch decay). The trace on the right was acquired with cross polarization and high power CW 1H decoupling. In both cases the sample was spinning at 2 kHz. The signal-to-noise improvement resulting from cross polarization is obvious.

Monday, December 3, 2007

Cross Polarization

The sensitivity of NMR spectra for solids containing protons can be increased dramatically by using cross polarization. Cross polarization is a technique where magnetization is transferred from an abundant proton source to a dilute (isotopically of chemically) nucleus, X, during a "contact" period. During the contact time, rf fields for both 1H and X are turned on. The ratio of power levels between 1H and X must be equal to the ratio of gyromagnetic ratios between X and 1H. The enhanced magnetization of the dilute isotope is then detected while the abundant protons are decoupled. The maximum gain in sensitivity is equal to the ratio of gyromagnetic ratios between 1H and X. The technique has the additional advantage in that the relaxation delay can be chosen based on the T1(rho) for 1H rather than the T1 of the dilute isotope, which is often larger by at least an order of magnitude.