I've worked on developping a new type of treatment for DOSY experiments in
the laboratory of Professor Canet in Nancy, France. So what's I know about
DOSY :
The DOSY (Diffusion Ordered Spectroscopy ) experiments consists in measuring
self-diffusion coefficients using gradients pulses in NMR.
In DOSY experiments a set of one dimensional (1D) spectra is recorded in
which each spectrum has been encoded differently by molecular diffusion.
Such 2D-diagrams result from the analysis of series of 1D dimensional NMR
spectrum obtained via pulsed field gradient echo type experiment (PFGE) in
which the gradient strength (or length) has been increased . Each spectrum
differs from the other one only by amplitudes (integrals) of peaks which
depend mainly on the molecular diffusion coefficients and on the gradient
strength (or length) applied. Plotting the peak amplitude versus the square
of the effective gradient (strength or length) applied reveals an
exponential decay (possibly multiexponential) whose rates provide the
diffusion coefficients. It is therefore possible to reconstruct a two
dimensional spectra where chemical shifts are displayed along one dimension
and diffusion coefficients along the other one.
Concerning the treatment of experimental data, if the decay in the diffusion
dimension is a single exponential, a non-linear least-squares fitting can be
sufficient so as to evaluate the diffusion coefficient . However, in the
"multiexponential" case, this last method is not sufficiently efficient and,
in principle, an Inverse Laplace Transform (ILT) of the data should be
performed.
Provencher implemented this approach by developing the CONTIN algorithm
which enables the reconstruction of DOSY spectra of complex mixtures . Other
softwares such as SPLMOD or NLREG , based on the ILT analysis, have been
proposed. However, with all these approaches, separation of species with
similar molecular size is not always satisfactorily achieved. More recently,
another technique based on the maximum entropy analysis has been proposed by
Delsuc et al. . The algorithm MaxEnt implemented in the GIFA software
represents an further step for solving the ILT as it leads to a finer
resolution in the diffusion domain.
Some references :
Morris K. F., Johnson C. S. Jr., J. Am. Chem. Soc. 1992, 114, 3139-3141
Hinton D. P., Johnson C. S. Jr., J. Phys. Chem. 1993, 97, 9064-9072
Morris K. F., Stilbs, P., Johnson C. S. Jr., J. Anal. Chem. 1995, 66, 211-215
Barjat H., Morris G. A., Smart S., Swanson A. G., Williams S. C. R., J.
Magn. Reson. 1995, 108B, 170-172
___________________________________________________________________
CORINNE MOURO
Laboratoire de Résonance Magnétique en Biologie et Médecine (LRMBM)
Université de Rennes 1
Campus de Villejean
Avenue du Prof. Léon Bernard
35043 Rennes
France
Tel : 02-99-33-69-31
Fax : 02-99-33-68-92
e-mail : Corinne.Mouro_at_univ-rennes1.fr
___________________________________________________________________
Received on Mon Jan 31 2000 - 08:32:17 MST