FID

Liquid phase chromatography can use a supercritical fluid as an eluent. The solvent evaporates on leaving the column and allows detection by FID. At present, there are few instances in the petroleum industry using the supercritical fluid technique. 

This type of analysis requires several chromatographic columns and detectors. Hydrocarbons are measured with the aid of a flame ionization detector FID, while the other gases are analyzed using a katharometer. A large number of combinations of columns is possible considering the commutations between columns and, potentially, backflushing of the carrier gas. As an example, the hydrocarbons can be separated by a column packed with silicone or alumina while O2, N2 and CO will require a molecular sieve column. H2S is a special case because this gas is fixed irreversibly on a number of chromatographic supports. Its separation can be achieved on certain kinds of supports such as Porapak which are styrene-divinylbenzene copolymers. This type of phase is also used to analyze CO2 and water. 

The hydrocarbons are separated in another column and analyzed by a flame ionization detector, FID. As an example, Figure 3.13 shows the separation obtained for a propane analyzed according to the ISO 7941 standard. Note that certain separations are incomplete as in the case of ethane-ethylene. A better separation could be obtained using an alumina capillary column, for instance. 

Among the various detectors specific for nitrogen, the NPD (Nitrogen Phosphorus Thermionic Detector) we will consider, is based on the following concept the eluted components enter a conventional FID burner whose air and hydrogen flows are controlled to eliminate the response for hydrocarbons. 

It is evident from the figure that impurities can complicate the use of NMR integrals for quantitation. Further complications arise if the relevant spins are not at Boltzmaim equilibrium before the FID is acquired. This may occur either because the pulses are repeated too rapidly, or because some other energy input is present, such as decoupling. Both of these problems can be eliminated by careful timing of the energy inputs, if strictly accurate integrals are required. 

NOEs between nuclei are often exploited to demonstrate their spatial proximity, as described in the final section. It is possible to obtain decoupled NMR spectra without the complications of the NOE, by confining the decoupling irradiation to the period of the FID alone, and then waiting for 10 x before... 

Wlien is very short, which is almost always true with nuclei having/> 1/2, the dipolar contribution to relaxation will be negligible and, hence, there will be no contributions to the integral from either NOE or saturation. However, resonances more than about 1 kHz wide may lose intensify simply because part of the FID will be lost before it can be digitized, and resonances more than 10 kHz wide may be lost altogether. It is also hard to correct for minor baseline distortions when the peaks themselves are very broad. 

The remarkable stability and eontrollability of NMR speetrometers penults not only the preeise aeeiimulation of FIDs over several hours, but also the aequisition of long series of speetra differing only in some stepped variable sueh as an interpulse delay. A peak at any one ehemieal shift will typieally vary in intensity as this series is traversed. All the sinusoidal eomponents of this variation with time ean then be extraeted, by Fourier transfomiation of the variations. For example, suppose that the nomial ID NMR aequisition sequenee (relaxation delay, 90° pulse, eolleet FID) is replaeed by the 2D sequenee (relaxation delay, 90° pulse, delay i -90° pulse, eolleet FID) and that x is inereased linearly from a low value to ereate the seeond dimension. The polarization transfer proeess outlined in die previous seetion will then eause the peaks of one multiplet to be modulated in intensity, at the frequeneies of any other multiplet with whieh it shares a eoupling. 

In the linear approximation there is a direct Fourier relationship between the FID and the spectrum and, in the great majority of experunents, the spectrum is produced by Fourier transfonnation of the FID. It is a tacit assumption that everything behaves in a linear fashion with, for example, imifonn excitation (or effective RF field) across the spectrum. For many cases this situation is closely approximated but distortions may occur for some of the broad lines that may be encountered in solids. The power spectrum P(v) of a pulse applied at Vq is given by a smc fiinction 18]... 

Another problem in many NMR spectrometers is that the start of the FID is corrupted due to various instrumental deadtimes that lead to intensity problems in the spectrum. The spectrometer deadtime is made up of a number of sources that can be apportioned to either the probe or the electronics. The loss of the initial part of the FID is manifest in a spectrum as a rolling baseline and the preferential loss of broad components of... 

For quadnipolar nuclei, the dependence of the pulse response on Vq/v has led to the development of quadnipolar nutation, which is a two-dimensional (2D) NMR experiment. The principle of 2D experiments is that a series of FIDs are acquired as a fimction of a second time parameter (e.g. here the pulse lengdi applied). A double Fourier transfomiation can then be carried out to give a 2D data set (FI, F2). For quadnipolar nuclei while the pulse is on the experiment is effectively being carried out at low field with the spin states detemiined by the quadnipolar interaction. In the limits Vq v the pulse response lies at v and... 

The interval between the second and third pulse is called the mixing time, during which the spins evolve according to the multiple-spin version of equation B 1.13.2 and equation B 1.13.3 and the NOE builds up. The final pulse converts the longitudinal magnetizations, present at the end of the mixing time, into detectable transverse components. The detection of the FID is followed by a recycle delay, during which the equilibrium... 

FID does not die away before the deadtime has elapsed. In die case of inliomogeneously broadened EPR lines (as typical for free radicals in solids) the dephasing of the magnetizations of the individual spin packets (which all possess slightly different resonance frequencies) will be complete within the detection deadtime and, therefore, the FID signal will usually be undetectable. 

In electron-spin-echo-detected EPR spectroscopy, spectral infomiation may, in principle, be obtained from a Fourier transfomiation of the second half of the echo shape, since it represents the FID of the refocused magnetizations, however, now recorded with much reduced deadtime problems. For the inhomogeneously broadened EPR lines considered here, however, the FID and therefore also the spin echo, show little structure. For this reason, the amplitude of tire echo is used as the main source of infomiation in ESE experiments. Recording the intensity of the two-pulse or tliree-pulse echo amplitude as a function of the external magnetic field defines electron-spm-echo- (ESE-)... 

As was mentioned above, the observed signal is the imaginary part of the sum of and Mg, so equation (B2.4.17)) predicts that the observed signal will be tire sum of two exponentials, evolving at the complex frequencies and X2- This is the free induction decay (FID). In the limit of no exchange, the two frequencies are simply io3 and ici3g, as expected. When Ids non-zero, the situation is more complex. 

It is only during an evolution (perhaps between sampling points in an FID) that these totals need be divided amongst the various lines in the spectmni. Therefore, one of the factors in the transition probability represents the conversion from preparation to evolution the other factor represents the conversion back from evolution to detection. 

---