• View
• Edit
• History
• Print

BlockDiagram

< Experiments for students | Index | The probehead >

The figure shows a very sketchy block diagram of an NMR experiment. Let's suppose that the sample, inside the coil of the LC circuit on the left, has a non-zero nuclear magnetization, due either to a strong external static magnetic field or to an internal magnetic field. The experiment may be the simplest, a {$\pi/2$} pulse followed by detection of the FID.

The radio frequency is produced by a synthesizer on the top left. The next block, basically a fast switch, cuts a pulse of fixed time length from the continuous wave of the synthesizer. If the frequency corresponds to the nuclear Larmor frequency the time length and the power of this pulse determine the nutation angle of the spins. Good coherence is required, i.e. the initial phase of the sinusoidal signal must be very stable. We want the nutation angle to be {$\pi/2$}, but remember that the Fourier spectrum of a radio-frequency pulse of width {$\tau$} is broad and contains significant components up {$\omega\pm 1/\tau$}.

Since the NMR spectrum in solids is broad (nuclei resonate at different frequencies) we need to produce short pulses to irradiate a large fraction of them (the shorter the duration {$\tau$}, the broader the width of the irradiated spectrum {$1/\tau$}).

To keep the pulse short one requires a linear power amplifier. The amplification produces a large voltage, hence a large current in the coil, hence a large radio-frequency field {$B_1$} to compensate for the short duration; the linearity grants that all the Fourier components in the pulse are amplified by the same factor, to keep the envelope shape)

From here on the circuit looks like a T, where three paths meet: one is the transmitter (TX, top of the figure described above), the second is the probe circuit (left side, described in the next section) and the third is the receiver (RX, bottom of the figure). Ideally we would like it to act as an intelligent switch, by: i) opening the RX end of the circuit only during transmission of the strong {$\pi/2$} pulse and ii) opening instead the TX end when the much weaker NMR signal travels from the probe to the RX end.

Both conditions are accomplished by a simple passive circuit, a diode bridge made by two parallel diodes joined in reversed polarity with each other. This bridge is conductive when polarized above the diode threshold (typically 0.6V): the positive half of the wave passes (with a distortion around the node) through one diode and the negative one through the other.

Condition ii) is met by the first bridge in series on the path from TX to the probe. High power pulses pass through it, but for the weak NMR signal (<<0.6V) the bridge acts as an open circuit.

Condition i) is obtained by the bridge on the RX side, connecting the line to ground after a coaxial cable of length approximately equal to a quarter wave-length of the radio frequency. This works like the end of a rope tied to the wall: both impose a boundary condition ({$V=0$}, or zero oscillation) at one end. In quarter wave-length lines like these the power going in is reflected with a change of sign (to grant the local {$V=0$} boundary condition), so that the power coming back out at a quarter wave-length distance compensates exactly that going in. Summarizing, the circuit looks like an open one, as long as the diode bridge is conductive (high power pulses), but the weak NMR signal can travel to RX undisturbed.

Just after the RX diode bridge a broadband, fast pre-amplifier, with very good linearity properties multiplies the NMR signal by a gain factor. The signal is then transferred to audio-frequencies, in a detection scheme loosely analogous to that of a lock-in amplifier. It is then recorded in quadrature, a jargon expression to indicate that both phase and amplitude are preserved in the spectral transfer from a radio to audio frequencies. This is described with some more details in a forthcoming section.

< Experiments for students | Index | The probehead >