The proposed method was experimentally tested on a laboratory setup (see Fig. 1), which included the tube with the flowing liquid, magnet-polarizer with magnetic induction BP, magnet-analyzer with magnetic induction BA, nutation coil, modulation coil, autodyne detector with detection coil and sound generator. As the flowing sample passes through the tube inside the magnet-polarizer, the polarization of protons occurs, which is detected in the magnet-analyzer by the magnetic resonance signal using the autodyne detector and the standard method of magnetic field scanning. The nutation coil is used to change the orientation of the nuclear magnetization vector.
Structural diagram of the NMR spectrometer with the flowing sample and inductive signal detection.
Examples of observed NMR signals from water (a) and ethyl alcohol (b) protons are presented in Fig. 2. These signals (wiggles), as in , have differences in the form of oscillations, depending on T1 and T2.
NMR signal from the flowing sample as a function of time from: (a) water and (b) ethyl alcohol.
The sequence of measurement to determine the longitudinal relaxation time of the flowing sample was carried out as follows:
◆ The flow rate of the flowing test sample, the amplitude of the radio-frequency field, as well as the frequency and amplitude of the modulation of the constant magnetic field, at which the amplitude of the NMR absorption signal reaches the maximum value, were measured;
◆ The range of the optimal flow rate within which the amplitude of the NMR absorption signal of the flowing test sample is reduced by a fixed value determined by the signal-to-noise ratio (for example, by 1% at the signal-to-noise ratio of 100, or by 10% at the signal-to-noise ratio of 10) was determined;
◆ The range of values of the longitudinal relaxation time corresponding to the measured range of the optimal flow rate of the flowing test sample was determined;
◆ The flowing test sample was replaced by a flowing reference sample with the known value of the longitudinal relaxation time, the value of which is within the range of values of the longitudinal relaxation time of the test sample;
◆ The flow rate of the flowing reference sample, the amplitude of the radio-frequency field, as well as the frequency and amplitude of the modulation of the constant magnetic field, at which the amplitude of the NMR absorption signal reached its maximum value, were measured;
◆ The flow rate of the flowing reference sample was reduced to a value corresponding to the amplitude of the NMR absorption signal, reduced by half compared with the maximum value;
◆ The calibration dependence of the relative amplitude of the NMR absorption signal from the flowing reference sample on the longitudinal relaxation time in the range of values corresponding to the measured range of the optimal flow rate of the flowing test sample was built;
◆ The flowing reference sample was replaced by the flowing test sample, and the test sample signal amplitude was measured relative to the maximum amplitude of the NMR absorption signal at the same values of the flow rate, the amplitude of the radio-frequency field and the frequency and amplitude of the modulation of the constant magnetic field;
◆ The relative amplitude of the NMR absorption signal from the flowing test sample was compared with the data of the calibration dependence and the deviation of the longitudinal relaxation time from the known longitudinal relaxation time of the flowing reference sample was determined.
In practice, a new method of the relaxation time T1 measurement consists of the following steps. On the first step, S(k), the dependencies of NMR signal on the saturation factor k were built (see Fig. 3). In this case, the saturation factor is equal to γH1T2 (γ is the gyromagnetic ratio and H1 is the intensity of the resonant radio-field (tunable parameter)). The dependences S(k) were taken for samples with the well-known relaxation time T1. On the second step the dependence of NMR signal on the relaxation time T1 was built (see Fig. 4). The dependence S(T1) (see Fig. 4) makes it possible to determine the relaxation time for varied sugar concentrations with high accuracy. It is important for this method to determine the amplitude of the NMR signals correctly. This problem is discussed in the next part of the paper.
NMR signal from the flowing sample as a function of saturation factor k: 1: Water; 2, 3, and 4: Water solution of sugar.
NMR signal from the flowing sample as a function of relaxation time T1.
Oxygen-rich water with the known longitudinal relaxation time of 0.65 s was used as a reference, and specially prepared decontaminated water obtained as a result of multiple boiling procedures was used as a test sample. In the experiment, in both cases, the transverse relaxation time was determined by the gradient of the working magnetic field of the magnet-analyzer and was 1 ms, while the longitudinal relaxation time of the test sample was equal to 3.25 s. Such a significant difference in T1 values is unambiguously explained by the influence of the electron paramagnetic (oxygen) on the relaxation rate in the reference sample, which should be taken into account in the course of measurement, as in .
This account assumes the selection of reference samples in which the corresponding procedure of environment decontamination to exclude the factor of similar influence is carried out. In this case, we emphasize the fundamental difference between the considered method of measuring the time of longitudinal relaxation of the liquid from similar methods for stationary samples. In particular, we are talking about the method of a quick passage of a constant magnetic field across the resonance, as in , according to which we measure the ratio of the NMR signals corresponding to two successive fast (in comparison with the rate of longitudinal relaxation) passages separated by a period of time τ. The signal observed at the first pass is proportional to the transverse macroscopic magnetization M0 at the optimal rate of passage and the corresponding intensity of the radio-frequency field, as in
where χ is the static nuclear susceptibility and H0 is the constant magnetic field. The signal of the second fast passage through the resonance in the opposite direction is proportional to the longitudinal component of the macroscopic magnetization MZ:
For the symmetric passage of a constant magnetic field, the NMR signal corresponds to the stationary value of the nuclear magnetization M, determined by the expression:
The longitudinal relaxation time T1 is determined by two measured amplitudes of the NMR signal with different values of τ. For the flowing sample, (1) is valid only if the time τ is significantly less than the time spent by the flow of sample in the analyzer volume θ, which is directly determined by the size of the volume and the rate of the liquid (otherwise, at θ<<τ, the sample will be changed by flowing in the analyzer volume, which leads to a loss of information about the dynamics of macroscopic nuclear magnetization in the period of the first passage through the resonance). As an example, let us take the volume of the analyzer of 1 cm3 at a flow rate of 50 cm3/s, (corresponding to the practice of NMR for the maximum of the signal-to-noise ratio). In this case, the time θ is equal to 0.02 s, which means the time τ should be at least several times less than this value, for example, 0.005 s. It is easy to calculate as in (1) that the change of the time τ twice from 0.005 s to 0.01 s corresponds to the relative change in the amplitude of the signals of only 0.2%, which is significantly lower than the limits of error of measuring the NMR signal amplitude in the flowing sample.
The alternative to the induction method of information retrieval in an NMR spectrometer with a flowing liquid is to use a highly sensitive magnetic sensor with optical pumping in the registration zone, as in . This method was first proposed in , where the NMR scheme of the spectrometer in combination with the quantum magnetometer on the Hanle effect was considered, as in . The practical implementation of such technique was developed much later with the advent of laser pumping sources. As in Fig. 5, the laser beam orients the Cs atoms placed in the absorption cell, which detect the rotation of the magnetization vector M. The radio-frequency coils (RF coils) produce the change in the orientation of the magnetization vector M in the external magnetic field (BEXT). Low-frequency modulated coils (LF) produce the modulation field B1. Unlike the resonance magnetometer with optical pumping, in which the measured magnetic field is detected by the frequency of resonance, in the Hanle magnetometers the signal absorption is measured. The amplitude of this signal is proportional to the value of measured magnetic field H0, the direction of which is oriented perpendicular to the circularly polarized light beam (see Fig. 5). Parallel to the measured field H0 an auxiliary modulating field H1cosωt is applied, the frequency of which satisfies the condition ω>>Γ, where Γ is the relaxation width of the resonance line. In weak fields, when the inequality γH0<<Γ (γ is the gyromagnetic ratio) is fulfilled, a slow precession of the magnetization vector M created by light will occur, the value of which decreases with the velocity Γ, while due to the fulfillment of this inequality, the average resulting magnetization vector will practically coincide with the direction of the pump light. At a non-zero magnetic field the resulting magnetization vector will be smaller in magnitude and directed at some angle to the direction of propagation of the pump light. When the inequality γH0>>Γ is fulfilled, the resulting magnetization vector becomes even smaller and, in the limit of an infinitely large magnetic field, tends to zero.
Structural diagram of the NMR spectrometer with the flowing sample and optical detection.
When the variable field H1cosωt is applied in the vicinity of the zero magnetic field, the vector M oscillates along the х-axis with the doubled frequency of 2ω. By increasing the magnetic field to a value comparable to the width of the resonance line, the oscillations of the vector increase in amplitude and occur at frequency ω. The presence of the field H1cosωt leads to the fact that the formation of the oscillating magnetization is realized not only in the vicinity of the zero-value of the field H0, but also for other values satisfying the resonance condition:
where n is an integer number. The widths of these resonances are the same for any value of n and are equal to 2Γγ−1 and the amplitude of the Hanle signal of the magnetometer is determined by the product of Bessel functions J1(γH1/ω) and J0(γH1/ω).
The sensitivity of the Hanle magnetometer at the parametric resonance (the minimum possible detectable variation) is equal or more than 10−9 Oe, which makes it possible to reliably detect the precession signal of the nuclear moments of atoms, located near the absorption cell of the magnetometer as in the proposed device. A similar signal in the form of a damped precession of the nuclear moments of helium-3 atoms was observed, as in , for 11 hours, which exceeds the time of transverse relaxation of helium-3 nuclear moments due to the gradients of the working magnetic field, polarized in the conditions of spin-exchange interactions with optically oriented vapors of alkali metals, as in . According to this work, in the case of the optical pumping of potassium vapor and inert gases (helium-3 and neon) mixture, the values of the inhomogeneous effective magnetic field generated by the atoms of electron and nuclear paramagnetic materials was equal to 10−6 Oe and 10−3 Oe, respectively, at the laser pumping power of 100 mW.
The operation of the Hanle magnetometer requires preliminary calibration, the degree of reliability of which depends on the orientation of the optical axis of the device relative to the vector of the measured field. In the actual operation of such a device, the transverse components of the field H0 differ from zero, and their magnitude and directions are random variables. The estimates show that for the compensation of transverse field components with the accuracy of 0.01 line width, the accuracy of magnetic field measurement approaches the sensitivity threshold determined by the relaxation width of the line Γ divided by the signal-to-noise ratio.
One of such ultra-miniature magnetic sensor (volume of several cubic millimeters), as in , provides accurate measurement of the transverse relaxation time by using a special volume built into the pipeline with the flowing sample and placed in a high-uniform magnetic field of 250 mG, almost excluding the influence of magnetic gradient on the measurement result. Such a technique, undoubtedly, can be in demand in the designs of NMR spectrometers with the flowing liquid, where the size of the working pipeline is orders of magnitude larger than the miniature analogue, as in . In this regard, it is interesting to compare the parameters of the useful signal of the Hanle magnetometer (in particular the signal-to-noise ratio) with the signal observed by the inductive method on the laboratory setup. The typical value of the Hanle parameter of magnetometers is equal to 3500 Hz−1/2, as in . For the inductive method applied to the flowing sample, as in , the signal-to-noise ratio in a uniform magnetic field BA can be estimated as follows:
where f is the noise factor taking into account the noise of the equipment, γ is the gyromagnetic ratio of protons, η is the fill factor, Δν is the bandwidth of the input amplifier, Q is the quality factor of the receiving radio frequency coil, β is the coefficient taking into account the reduction of the macroscopic magnetization during transportation of the flowing liquid through the connecting pipeline with the volume VT from the magnet-polarizer with induction BP to the magnet-analyzer with induction VA, χ is the nuclear magnetic susceptibility, α is the coefficient having an order of magnitude of 10−2 taking into account the weight contribution of the volume of the connecting pipeline in the analyzer volume, k is the Boltzmann constant, and T is the temperature. With reference to the laboratory setup (parameters: f≈2; η≈1; Q≈100; β≈0.1; BA≈500 G; BP≈5000 G; Δν≈1 Hz; αVT ≈1 cm3; χ≈3×10−10; Т≈300 K), the value is equal to 10000 Hz−1/2, which is in order of magnitude close to this parameter in the Hanle magnetometers. At the same time, an important advantage of using a quantum magnetometer with optical pumping is the absence of the magnet-analyzer and the ability of measuring both the longitudinal and transverse relaxation time with high accuracy comparable to the accuracy of measurement carried out by high-resolution NMR equipment.
The principal difference between the optical pumping modes in the cells with the anti-relaxation coating and with the buffer gas is the different temperature of the absorption cells, which in the case of using the buffer gas can significantly exceed the level of 100 °C. This creates a problem of effective isolation of the Hanle magnetometer sensor from the volume with the flowing sample and complicates the design of the entire device. In this regard, it is preferable to use absorption cells with anti-relaxation coating, the operability of which is provided at lower temperatures. In the case of cesium atoms, the operating temperature of the absorption cell with linear dimensions of 5 cm is close to room temperature and the quality factor of the observed signal is an order of magnitude higher than this parameter in the buffer gas cells. At the same time, the high variation of the sensitivity of magnetometers with optical pumping at the level of 10−14 T·Hz−1/2, as in , does not guarantee reliable detection of the NMR signal of the flowing sample in the case of ineffective suppression of variations of the external magnetic field. For example, to reduce the relative variation of the geomagnetic field of 10−5 in the area of the Hanle magnetometer sensor placement to the above level, it is required to use a magnetic shield with a dynamic shielding coefficient of 5×104.