The next blog in the characterization mini-series was prepared by my colleague Tom Early. Tom has been an Nuclear Magnetic Resonance spectroscopist at GE for nearly thirty years and has been doing NMR for over forty years. Tom loves creating software for analyzing chemistry data and has also been an active Six Sigma Master Blackbelt for fourteen years. When not in the lab, Tom loves bicycling and being with his family, especially his grandchildren.
1. What is it?
Nuclear magnetic resonance (NMR) has been widely recognized as a powerful tool for molecular structure. A sample placed in a very high magnetic field will absorb and re-emit radio frequency radiation. The RF signals that the molecules generate are loaded with information about the chemical structure of the molecule. With just ten milligrams of an unknown and an afternoon of NMR spectrometer time, an experienced NMR spectroscopist can unequivocally determine the molecular structure of that unknown. Using multidimensional NMR experimental techniques, these signals from the molecule show how the magnetic nuclei of that molecule interact with each other, either as felt through the bonding electrons in the molecule, or felt through space.
The figures above show the result of two popular 2D NMR structural experiments. On the left, a COSY spectrum shows which hydrogen nuclei have shared electrons (J-coupling). The 1D hydrogen spectrum is plotted along the top while correlated spins appear in the 2D map. On the right are two hydrogen-carbon correlation experiments with 1D carbon and hydrogen spectra plotted on the vertical and horizontal, respectively. Direct hydrogen-carbon bonds appear as black correlations in an HSQC experiment while longer range (2, 3 and 4-bond) interactions are plotted as purple correlations using an HMBC experiment.
While molecular structure determination by NMR is indeed important, NMR can also probe the dynamic aspects of a molecule. In the earliest days of my career, I used NMR to measure the dynamic properties of DNA. While applied radio frequencies perturb the equilibrium state of the nuclei in a molecule, the perturbed state returns to equilibrium at a rate that depends on the dynamic state of the molecule. Measuring the nuclear relaxation properties of the hydrogen in DNA, I was the first to show that the DNA double helix was fairly flexible. Starting with those experiments, I have always been very interested in NMR nuclear relaxation and so became quite interested when several years ago I heard of a relaxation experiment that can measure molecular self-diffusion by NMR.
2. How does it work?
Think of putting a drop of food coloring in a perfectly still beaker of water. With time, the dye will spread out and uniformly fill the entire beaker. This process is caused by the molecules of water and dye randomly bumping into each other. It turns out that the rate of diffusion is directly related to the viscosity of the solvent (in this case, water) and the size of the dye molecule. Diffusion ordered spectroscopy, DOSY, is a simple NMR experiment where a series of 1D spectra are collected while applying a weak, variable strength magnetic gradient across the sample. Analysis of the signal attenuation as a function of the gradient strength can determine the diffusion rate.
3. Why do we use the instrument?
As an example, below is DOSY experiment run on a sample of four different molecules in a chloroform solution.
As seen in the DOSY spectrum above, each proton peak in a spectrum maps to the diffusion rate of the molecule to which it belongs. Not only does this give us a tool for molecular dynamics, it can be used to simplify spectra of complex mixtures. Here, there are two relatively small molecules (that can still be resolved from one-another), a medium “oligomer” and a high molecular weight polymer.
The Stokes-Einstein equation tells us that the diffusion is inversely proportional to the hydrodynamic radius of the molecule. With small molecules diffusing faster than big ones, it is possible to use the Stokes-Einstein equation to measure the molecular size and therefore the molecular weight of a polymer. With my colleagues here at GRC, I have been doing some experiments to determine just how large of a molecule DOSY can get useful information.
Below is the result of DOSY measurements of several precisely manufactured polyethylene glycols. To check the manufacturer, we determined the exact molecular weight by MALDI and this is plotted on the horizontal axis, with hydrodynamic radius as determined by DOSY. I’m really encouraged by these results. It means that we have a very wide range of systems we can study.
Using the Stokes-Einstein equation to convert a measured DOSY diffusion to a hydrodynamic radius, we plot this against the MALDI-validated molecular weights of several polyethylene glycols.
A great question to answer now is can DOSY be used to determine a distribution of sizes? We know that when hydrogen peaks can be resolved, different molecular diffusion can be measured for each component. But what about the case where the hydrogen NMR signals overlap? That, my friend, is a topic for another blog…