Fourier Transform NMR Spectroscopy Employing a Phase Modulated Rf Carrier

The present invention relates in general to Fourier transform NMR spectroscopy and more particularly to such spectroscopy employing a pulse modulated rf carrier to obtain the Fourier sidebands for exciting resonance of the sample.

Heretofore, Fourier transform NMR spectrometers have been proposed wherein an rf carrier is pulse modulated at low frequencies to produce a train of very short radio frequency pulses having Fourier sideband components of a sufficient bandwidth to cover the spectral lines of a sample to be analyzed for exciting the sample into nuclear magnetic resonance. For example, pulses of 100.mu. sec duration at a rate of 1 pulse per second was common. Resonance was picked up by suitable detector coils and was phase sensitive detected to produce a composite audio frequency resonance signal. This audio signal was fed to a computer for successive sampling of the detected resonance signal existing in the time frame between successive transmitter pulses and was used to derive a set of time displaced analog signal components which were converted to digital form and stored in respective channels of the memory of the computer for time averaging same. The time averaged resonance signal derived from the respective channels were read out of the time averager and Fourier transformed via a suitable Fourier transformer program into the frequency domain to obtain a time averaged spectrum of the sample under analysis. Such a spectrometer is disclosed in U.S. Pat. No. 3,475,680 issued Oct. 28, 1969 and assigned to the same assignee as the present invention.

One of the problems encountered in this prior art type of NMR spectrometer is that anomalous phase and intensity effects are encountered in the resultant spectral data. These anomalies occur when the interval between rf pulses is made short in comparison with the spin-spin relaxation times of the nuclei under analysis and they are associated with the establishment of a steady-state response where there are finite macroscopic transverse components of magnetization at the end of the pulse interval. The regular pulse sequence has the property of refocusing the isochromatic magnetization vectors that have been dispersed by field inhomogenity, giving rise to a spin echo at the time of the next pulse. In some cases, these echos, phase and intensity effects may be averaged out by short-term instabilities in the field/frequency ratio or by incoherence introduced by heteronuclear noise decoupling.

Pseudorandom excitation schemes are also known for providing wideband excitation of the sample as in U.S. Pat. No. 3,581,191 issued May 25, 1971 which utilize binary phase or intensity modulation. Such anomalies are described in an article titled "Phase and Intensity Anomalies in Fourier Transform NMR" by Ray Freeman and H. D. W. Hill appearing in the Journal of Magnetic Resonance, Vol. 4, pages 366-383, published in 1971.

In this article, as well as in copending U.S. patent application Ser. No. 367,598 filed June 4, 1973 and assigned to the same assignee as the present invention, a scheme is disclosed which effectively eliminates these anomalies by introducing a small random delay in the timing of the radio frequency pulses. Introducing this random delay in the timing of the radio frequency pulses is suitable for most spectrometer applications but there are cases where the timing between radio frequency pulses is relatively short and, thus, any delays introduced in the timing tend to greatly increase the spectrometer time required to obtain the desired time average spectral data.

It has also been proposed to eliminate the anomalous phase and intensity effects in Fourier transform NMR spectroscopy by periodically shifting the frequency of the carrier signal from one frequency to another of at least four different frequencies in accordance with a predetermined sequence. This change of the excitation frequency would serve to shift all the Fourier sideband lines of the radiation spectrum by the same amount. Thus, by using the frequency shift Fourier transform technique, an accurate adjustment of the pulse length is not necessary, nor is an adjustment of the phase of the apparatus necessary. The only thing to be changed is the computer program.

Also, the frequency shift spectrometer delivers four times the quantity of data as the conventional Fourier transform spectrometer as disclosed in the aforecited patent U.S. 3,475,680. Thus, the memory of the computer to evaluate the frequency shift Fourier transform measurement should be greater than that used for a normal Fourier transform spectrometer. For high resolution spectroscopy, the frequency shifting is extremely critical and dependent on the utilized pulse spacing. This frequency shift Fourier transform method is described by A Schwenk in the Journal of Magnetic Resonance, Vol. 5, pages 376-389, published in 1971.

The principal object of the present invention is the provision of an improved method and apparatus for pulse NMR spectroscopy for eliminating phase and intensity anomalies in the frequency spectrum utilizing a means for switching the phase of the rf excitation pulses in a predetermined manner.

In one feature of the present invention, the phase of the pulsed radio frequency carrier is periodically shifted from one phase to another of at least three different relative phases for eliminating undesired phase and intensity anomalies in the Fourier transformed spectrum of the sample under analysis.

In another feature of the present invention, the rf carrier is switched between four phases by means of switching in different delays of the pulsed carrier signal.

In another feature of the invention the phase of pulsed carrier signal is sequentially switched by deriving the carrier frequency by means of a divider which divides a higher frequency signal and variably delaying the high frequency signal to the divider to derive the variable phase shifts of the carrier signal.

Other features and advantages of the present invention will become apparent upon a perusal of the following specification taken in connection with the accompanying drawings wherein:

FIG. 1 is a schematic block diagram of an NMR spectrometer employing features of the present invention,

FIG. 2 is a schematic line diagram of a phase shifter employed in the circuit of FIG. 1,

FIG. 3 is a waveform diagram depicting four different phase sequences for the phase of the pulsed carrier utilized in the spectrometer of the present invention,

FIG. 4 is a schematic diagram of the control of the phase shifter of the present invention, and

FIG. 5 is a schematic block diagram of an alternative phase shifter of the present invention.

Referring not to FIG. 1 there is shown a Fourier transform NMR spectrometer employing features of the present invention. More particularly, a radio frequency transmitter 11 supplies a carrier signal at a suitable frequency displaced in frequency to one side of the nuclear magnetic resonance frequency of a sample under analysis which is disposed within a probe 12 and immersed in a dc polarized magnetic field of high homogeneity. The output of the rf transmitter 11 is fed to the input of a phase shifter 13 which serves to shift the phase of the rf carrier in 90.degree. increments according to the relative phase relationships indicated in FIG. 3. More particularly, the phase of the rf carrier is shifted according to the sequences one through four of FIG. 3 where phase A corresponds to an rf pulse with zero phase shift, phase B corresponds to an rf pulse with 90.degree. phase shift relative to A, C is an rf pulse with 180.degree. phase shift relative to A and D is an rf pulse with 270.degree. phase shift relative to A.

The first pulse train of sequencing pulses through A, B, C, D, phase shift has a duration corresponding to one quarter of the total time that the sample will be under analysis, e.g., total time usually in order of several minutes. And, similarly, the second pulse train follows the first pulse train and similarly extends for one quarter of the total observation time, as do the third and fourth trains.

The phase shifted carrier is passed by gate 14 in accordance with a duty cycle derived from a control 15. The pulse width and pulse period are selected by the control to derive Fourier side band components of sufficient bandwidth to simultaneously excite the various resonance spectral lines of the sample under analysis. The output of the gate 14 is fed to a power amplifier 16 and thence to the probe 12 for exciting resonance of the sample.

Resonance of the sample is detected in a detector coil, of conventional design, and fed to the input of a receiver 17 for amplification therein. The amplified output is thence fed to one input of a phase sensitive detector 18 for phase detection against a sample of the phase shifted carrier to derive an audio frequency composite resonance signal which is thence fed to the input of the signal averager 19 which output is transformed to the frequency spectrum in Fourier transformer 10 and displayed.

The control 15 preferably generates time share modulation commands so that the receiver 17 is deactivated as by an internal gate during the time that the transmitter pulse is applied to the probe 12 so that radio frequency energy is not directly coupled from the transmitter 11 through the probe 12 into the receiver 17 which would otherwise tend to overload the receiver. A time share modulation scheme forms the subject matter of U.S. Pat. No. 3,786,341 issued Jan. 15, 1974 and assigned to the same assignee as the present invention. Thus, the receiver 17 is gated on during the times in between successive rf pulses applied to the probe 12.

Each of the detected analog audio frequency composite, or a separate time averaging computer, resonance signals is fed to the input of the signal averager 19 wherein it is time scanned and sampled at a multiplicity of time displaced intervals. The sampled quantities are converted from analog-to-digital quantities and stored in respective channels of the memory of the computer. Each of the time scans of successive analog composite resonance signals is synchronized with the time of the ends of the rf pulses to preserve relative phase information of the composite resonance signals for separating the absorption mode resonance signal components from the dispersive components. This phase synchronization and time averaging scheme forms the subject matter of U.S. Pat. No. 3,581,191 issued May 25, 1971 and assigned to the same assignee as the present invention.

Successive composite resonance signals derived from the sample are time averaged by adding the sampled data from corresponding time points from successive scans. Upon the completion of the desired number of pulse resonances or scans, the time averaged data is read out of the respective channels and Fourier transformed in any manner from the time domain into the frequency domain by transformer 10. This could also be accomplished in a digital computer using conventional Fourier transform algorithms as taught in the aforecited patents, such as U.S. Pat. No. 3,475,680, to derive a time averaged spectrum of the sample under analysis which is then printed out of the computer by a conventional print out.

The deleterious anomalies can be eliminated by dividing the experiment into time frames in which different sequences of phases of the rf pulse excitation are applied, as long as three or more phases are available. In general, if there are Ni different phases, i>2 where sequence N.sub.1, N.sub.2 . . . Ni is designated as sequence 1, one may determine the phase sequences for each set by dividing the experiment time into i equal time sets, then

Furthermore, the sets may be used in any order. It is also believed that a single set having at least three or more phases in a random or pseudorandom sequence would also be effective.

An advantage of the spectrometer of FIG. 1 is that, the duration between successive pulses is constant as contrasted with the random pulse delay technique and accordingly avoids introducing problems where the delay may be significant with respect to the relaxation time of samples. As contrasted with the frequency shift method, the scheme of the present invention has the advantage that only one Fourier transform must be executed and no critical frequency adjustments are required. Also, the pulse spacing can be changed without any further change required. In addition the size of the memory of the computer can be reduced as contrasted with the memory requirements for the frequency shift method.

Referring now to FIG. 2 there is shown one embodiment of the phase shifter 13. In this embodiment, a pair of relay operated switches are provided in series with the transmission line between the rf transmitter 11 and the gate 14. The first relay operated switch has two positions. In a first position a zero degree phase shift or delay line 21 is switched into series with the transmission line 22 and the second relay operated switch similarly switches in another zero phase shift or zero delay line 24 in series with line 22. Thus, in this position zero phase shift is encountered between the transmitter 11 and the gate 14. The relay operated switches are preferably the type which always return to the first position unless energized. These relays could be semiconductor switches or polarity sensitive relays which operate in two directions based on the polarity of the driving pulses.

In a second position corresponding to phase shift B, the first relay operated switch is energized to switch into series with line 22 a 90.degree. phase shift or delay line 25 while leaving delay line 24 in series with the line 22 so that a total of 90.degree. phase shift is obtained. In a third position corresponding to phase shift C, the first relay switch is energized to switch into line 22 the zero degree phase shift or delay line 21 and the second relay operated switch is energized to switch into series with line 22 a 180.degree. phase shift or delay line 26 so that the total delay from the transmitter 11 to the gate 14 is 180.degree.. The fourth state of the phase shifter corresponds to phase shift D and the first relay operated switch is energized to connect the 90.degree. phase shift line 25 in series with line 22 and the second relay switch is energized with current from the control to switch into series with line 22 the 180.degree. delay line 26 such that the total delay from the transmitter 11 to the gate 14 is 270.degree..

Referring now to FIG. 4, there is shown one embodiment of the control 15. Oscillator 40, as of 1Hz, is fed to total pulse counter 42 which is provided with sufficient stages to count each pulse over the entire experiment time T without any spill over. When the designated last stage of counter 42 changes state it generates the reset command to the set generator 43 and sequence generators 44, 45, and 46. Assuming that the counter contains n stages, the output of the n-3 stage provides a wave train which has a negative going pulse at each increment of 1/4T, thereby providing a pulse train which enables division of the total experiment time into four equal sets. The output from the n-3 counter of counter 42 is connected to the set counter 43 which causes the set counter to provide an output to enable one of the gates 47, 48 or 49 depending upon the quarter of the experiment time. Outputs A, B, C, D of flip-flop 60 of set counter 43 are connected to AND gates 61 through 64 such that AND gate 61 (SET 1) provides an output when A.C is a logical 1, AND gate 62 (SET 2) provides an output when B.C is a logical 1, AND gate 63 (SET 3) provides an output when A.D is a logical 1, and AND gate 64 (SET 4) provides an output when B.D is a logical 1. The output of AND gate 64 is not connected to enable any gates since each pulse in SET 4 has a zero phase shift which is accomplished without energizing either relay 1 or relay 2. The output of oscillator 40 is connected to delay 54 and thence to gates 47, 48 and 49 for passage through the appropriate gate to their corresponding sequence generator. For example, during SET 1, the output of gate 47 pulses two stage flip-flop 70 of sequence one generator 44. The A, B, C, D outputs are connected to AND gates 71-74 such that AND gate 71 provides an output when A.C is a logical 1, AND gate 72 provides an output when B.C is a logical one, AND gate 73 provides an output when A.D is a logical one, and AND gate 74 provides an output when B.D is a logical one. It can be seen therefore that sequence one generator 44 provides an output from AND gates 71, 72, 73, and 74 in sequence upon the initiation of each pulse from gate 47. AND gate 71 is not connected to either relay coil 1 or relay coil 2, therefore when AND gate 71 is a logical 1 the pulse from phase shifter 13 to gate 14 will be at zero phase. AND gate 72 is connected through isolation diode D1 to relay coil 1 in phase shifter 13, so that when AND gate 72 is a logical one, the pulse from phase shifter 13 to gate 14 will be at 90.degree. phase. Similarly, AND gate 73 is connected through diode D2 to relay coil 2 to generate the 180.degree. pulse and AND gate 74 is connected through relay diodes D3 and D4 to both relay coils 1 and 2 in order to generate the 270.degree. pulse.

Sequence two generator 45 is less complicated since sequence 2 uses only two phases. Accordingly, gate 48 is connected to sequence generator 46 which comprises a single stage flip-flop having its B output connected through diode D5 to relay coil 2. Since the A output of generator 46 does not energize a relay coil, the phases switch back and forth between the zero phase and 180 phase on each pulse.

Sequence three generator 45 is identical to sequence one generator 44 except that the AND gates 82 and 84 are reversed in regard to which relay coil they energize. Accordingly, the sequence three pulse phase sequence is as shown in FIG. 3.

The output of delay 54 is connected to delay 55 before connecting to one shot multivibrator 50. Delay 55 is to give relay coils 1 and 2 an opportunity to switch and stop bouncing before the multivibrator 50 having the output waveform 52 energizes the gate 14 to enable the excitation pulse to excite the probe 12. The time share modulation pulse train 53 is also derived from the multivibrator 50 and fed to receiver 17 to cause the receiver to have low response during the time of the peaks 56 of the pulse train 52.

The repetition adjustment potentiometer 41 enables variation in the oscillator 40 frequency and pulse width potentiometer enables adjustment of the pulse width outputs of multivibrator 50. These adjustments enable the operator to adjust the density and shape of the Fourier components and envelope of the rf excitation energy.

Referring now to FIG. 5 there is shown an alternative rf transmitter phase shifter circuit 11 and 13. An rf oscillator, operating at four times the frequency of the carrier which is to be pulse modulated 31, feeds an output to one input of an AND gate 32. A second output of the oscillator 31 is fed to the input of a counter 33 which is under the control of an output derived from the control 15. The counter contains four different counting stages. In a first stage, the output of the oscillator 31 is fed directly without delay to the input of an OR gate 34. In the second stage of the counter the counter counts one complete period of the output frequency of oscillator 31 and then feeds an output to the OR gate 34. The third stage of the counter counts for two periods of the oscillator 31 and then feeds an output to the OR gate 34. The fourth stage of the counter counts for three periods and then feeds an output to the OR gate 34. Thus, the output of the counter 33, depending upon the particular stage that is energized by the output of the control 15, will provide delays of zero, one cycle, two cycles and three cycles of the oscillator frequency 4f.

The output of the OR gate 34 is fed in as a second input to the AND gate 32 for ANDING with the output of the oscillator 31. The output of the AND gate 32 is fed to a divider 35 for division by four to obtain the carrier frequency f. Thus, the carrier frequency f is phase shifted by 0.degree., 90.degree., 180.degree. and 270.degree. depending upon the output of the counter 33 as determined by the output of the computer 15. These phase shifts correspond to A, B, C, and D, respectively, as utilized in the diagram of FIG. 3.

While the invention has been described in terms of its preferred practice and in terms of preferred embodiments, departures may be made from the procedure and embodiments thus described and still remain within the scope of the claims.