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.

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.

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.

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.

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.

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.

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)

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.

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.

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/

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.

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.