APPLICATION OF NEUTRON-ACTIVATION ANALYSIS TO THE DETERMINATION OF FLUORITE IN SURFACE MINE WORKINGS

R. P. MESHCHERYAKOV, Yu. A. PANOV

(Presented by the seminar of the radioactivation analysis sector of the Research Institute of Nuclear Physics, Electronics, and Automation)

In the process of developing a deposit, one of the longest and most labor-intensive operations is contouring and estimating reserves, which is associated with repeated sampling. One of the methods that makes it possible to eliminate preliminary sampling is the radioactivation method, the essence of which is reduced to irradiating objects with a particle flux, followed by measurement over time of the radioactivity produced. Compared with the chemical method currently in use, this method has the following advantages: rapidity and a relatively simple analytical procedure, the possibility of determining very small amounts of individual elements in rocks, and the possibility of element-by-element analysis of the composition of rocks in their natural occurrence.

Initially, the method of neutron-activation analysis for fluorite was used by V. Ya. Bardovskii [1] in borehole logging. The analytical procedure under laboratory conditions is discussed in the literature [2, 3]. Its essence is reduced to irradiating rocks containing $\mathrm{CaF}_2$ with a neutron flux from a polonium-beryllium source, followed by measurement of the induced activity from the short-lived isotopes $\mathrm{N}^{16}$ and $\mathrm{F}^{20}$.

The most widespread nuclear reactions under neutron action are radiative-capture reactions $(n,\gamma)$, which occur with thermal neutrons. As the neutron energy increases, the probability of inelastic scattering $(n,n')$ begins at first to increase, and at energies $E = 0.5 \div 10$ MeV so do reactions with emission of charged particles $(n,p)$ and $(n,\alpha)$, and at still higher energies reactions of the $(n,2n)$ type. When neutrons of different energies interact with the nucleus of the stable isotope $\mathrm{F}^{19}$, the following reactions are possible: $\mathrm{F}^{19}(n,\gamma)\mathrm{F}^{20}$, $\mathrm{F}^{19}(n,p)\mathrm{O}^{19}$, $\mathrm{F}^{19}(n,\alpha)\mathrm{N}^{16}$, $\mathrm{F}^{19}(n,2n)\mathrm{F}^{18}$. Activation analysis of fluorite ores under conditions of their natural occurrence is based on the use of two reactions: $\mathrm{F}^{19}(n,\gamma)\mathrm{F}^{20}$, which proceeds with slow neutrons, and $\mathrm{F}^{19}(n,\alpha)\mathrm{N}^{16}$, with fast neutrons; moreover, the latter reaction is preferable, since the number of fast neutrons is significantly greater than that of thermal neutrons. The $(n,p)$ reaction is insignificant because of the rather high reaction threshold, and $(n,2n)$ because of the small activation cross section for neutrons with energies of $2 \div 5$ MeV when a polonium-beryllium irradiation source is used.

In fluorite ores the principal components are calcium fluoride and silica, whose content may vary from 0 to 98%. In addition, the ores contain a considerable amount of impurities: CaCO₃, Al₂O₃, MgO, S, etc., the contents of which vary over wide limits—from thousandths of a percent to tens of percent. An analysis of the nuclear-physical characteristics of the isotopes that affect the determination of fluorine shows that some are not interfering because of their insignificant Clarke abundances, while others are not interfering because the effective cross sections of the nuclear reactions that arise are 1.5–2 orders of magnitude lower than the effective cross section of the reaction \( {}_{9}\mathrm{F}^{19}(n,\alpha){}_{7}\mathrm{N}^{16} \). Consideration of questions of the activation of fluorite and impurities makes it possible to conclude that CaF₂ can be determined by measuring the activity of the isotopes \( \mathrm{N}^{16} \) and \( \mathrm{F}^{20} \). Owing to their sufficiently short half-lives, these isotopes do not require prolonged irradiation for their production, which makes it possible to avoid activation of accompanying impurities. The high energy of the radiation accompanying the decay of the radioactive nuclei \( \mathrm{N}^{16} \) and \( \mathrm{F}^{20} \) makes it possible practically to eliminate the influence on the analytical results of all isotopes whose radiation energy is below 1.3–1.5 MeV by introducing a high discrimination threshold. The principal interfering isotope is \( \mathrm{Al}^{28} \), formed by two reactions, \( \mathrm{Si}^{28}(n,p)\mathrm{Al}^{28} \) and \( \mathrm{Al}^{27}(n,\gamma)\mathrm{Al}^{28} \). It has a short half-life and emits high-energy \(\gamma\)-quanta. In addition to aluminum and silicon, the analytical results may be affected by \( \mathrm{O}^{16} \), from which, as a result of the \((n,p)\) reaction, the isotope being determined, \( \mathrm{N}^{16} \), is formed. But since the threshold of this reaction is 9.5 MeV, then, taking into account the insignificant number of neutrons of this energy in the source spectrum and the small cross section of this reaction near the threshold, this source of errors in the determination of fluorine by the \((n,\alpha)\) reaction may practically be neglected.

Let us consider the choice of the irradiation time and the measurement time of the induced activity. The relative magnitude of the registered activity, according to the law of radioactive decay, will be expressed as:

\[ \frac{N_{\mathrm{reg}}}{N_{\max}} = \left( 1 - e^{-\frac{0.693 t_1}{T_{1/2}}} \right) \left( e^{-\frac{0.693 t_2}{T_{1/2}}} - e^{-\frac{0.693 (t_2+t_3)}{T_{1/2}}} \right), \tag{1} \]

where

\[ \begin{aligned} t_1 &\text{— time of activation of the sample;}\\ t_2 &\text{— time required to transfer the sample from the activator channel}\\ &\text{\quad to the measuring device;}\\ t_3 &\text{— time of registration of the induced activity.} \end{aligned} \]

For example, the value of the relative magnitude of the registered activity under the time schedule \((t_1 = 15\ \mathrm{sec},\ t_2 = 5\ \mathrm{sec},\ t_3 = 15\ \mathrm{sec})\), due to the presence of fluorine in the analyzed sample, will be 36.4% of the maximum activity (here \(t_1\) and \(t_3\) are of the order of two half-lives of \( \mathrm{N}^{16} \)), while for isotopes having half-lives of more than 1 minute it will be 2.3%. Consideration of a whole series of time schedules \((15\ \mathrm{sec} \le t_1 \le 30\ \mathrm{sec},\ 5\ \mathrm{sec} \le t_2 \le 10\ \mathrm{sec},\ 15\ \mathrm{sec} \le t_3 \le 30\ \mathrm{sec})\) shows that the influence of interfering elements still remains significant. For its more complete elimination it is necessary to introduce into the time schedule a second count of the induced activity, equal in duration to the first and following immediately after it \((t_4)\). Using the difference between the first and second counts, the expression for the registered activity will have the form:

\[ \frac{N_{\mathrm{reg}}}{N_{\max}} = \left( 1 - e^{-\frac{0.693 t_1}{T_{1/2}}} \right) \left( e^{-\frac{0.693 t_1}{T_{1/2}}} - \right. \]

\[ -2e^{-\frac{0.693(t_2+t_3)}{T_{1/2}}} - e^{-\frac{0.693(t_2+t_3+t_4)}{T_{1/2}}} \biggr). \tag{2} \]

Calculation by equation (2) shows that the use of a differential measurement between the first and second counts makes it possible to reduce the contribution from isotopes with half-lives greater than \(T_{1/2}=60\) sec by a factor of \(2.0 \div 2.5\) for all the above-indicated values of \(t_1\), \(t_2\), and \(t_3\). In addition, application of the time scheme with a second count makes it possible, with a definite degree of accuracy, to exclude from the first count the background of the recording system and the activity due to the presence of natural radioactive isotopes in the sample.

It is evident from equation (2) that an indispensable condition for the correctness of determining \(\mathrm{CaF}_2\) from the short-lived activity of the radioisotope \(\mathrm{N}^{16}\) is the accurate maintenance of the time intervals, especially in the first count. Thus, with an irradiation time equal to 30 sec, an inaccuracy of 1 sec gives an error in the activity measurement of 0.75%, whereas a similar time inaccuracy in the first count gives 9.9%.

Thus, by introducing a discrimination threshold, selecting the irradiation time and the time for measuring activity, and taking the magnitude of the background into account, the influence of interfering elements in determining fluorine under field conditions can practically be eliminated.

On the basis of the data presented above, the authors developed and built an instrument for carrying out neutron-activation analysis, having the following technical characteristics:

1. The instrument is a \(\gamma\)-spectrometer and is adjusted by means of a discriminator to count \(\gamma\)-quanta with energy above 1.33 MeV.

2. The relative error of the counting instrument does not exceed 0.25% at 150 statistically distributed pulses per second at the input.

3. The instrument has an electronic timing device with the following operating time cycles: 10 sec—blocking, 20 sec—counting during activity measurement, 10 sec—blocking, 20 sec—counting in reversible mode.

4. The basic error of time counting by the timing device is not more than 2%.

5. A reversible-counter mode is provided, with the activity-measurement time set by a stopwatch.

6. The sensitivity of the instrument at the \(3\sigma\) level is approximately equal to 0.7%.

7. The accuracy of determination at a point for content classes of \(0 \div 5\)%, \(5 \div 50\)%, or \(50 \div 100\)% is, relative to the mean, 14%, 5.5%, or 1.5%, respectively. The accuracy of determining the average content of reserves in a deposit, as compared with the results of chemical analysis, is \(0.4 \div 2\)% relative and \(0.3 \div 0.6\)% absolute.

A block diagram of the instrument is given in Fig. 1. Let us consider the purpose of its individual units. The detector of the induced activity is a NaI crystal of size \(40 \times 40\) mm and an FEU-13. Pulses of negative polarity, taken from the anode load of the FEU through an emitter follower, pass through a connecting cable to an amplifier having two voltage-amplification stages and one current-amplification stage. The voltage-amplified pulses are shaped in a discriminator assembled on a tunnel diode of type AI-301A. The triggering threshold is set by adjusting the amplification. After amplitude shaping, the pulses pass to a high-speed binary cell, which makes it possible to reduce missed counts at their average arrival frequency of 150 Hz from 6.5% to 0.25%.

To decouple the counting device on dekatrons with a binary cell, a blocking generator operating in standby mode is installed. The circuit of the counting device has a reverse-start mode, which makes it possible to automatically register the induced activity with simultaneous subtraction of the background and long-lived activities. There is a timing

Fig. 1. Block diagram of the instrument

device, the purpose of which is determined by the formation of specified time intervals for time selection. The power-supply unit consists of a set of batteries and a voltage converter necessary for supplying the photomultiplier tube, dekatrons, and radio circuits.

Fig. 2. Calibration curve for converting the instrument readings into percent CaF₂ content

Calibration of the instrument is carried out before the beginning and after the end of work with it. It makes it possible to avoid errors associated with monotonic changes in the parameters of the installation, caused by aging of semiconductor elements and the photomultiplier tube, losses of light output by the scintillator, polymerization of the optical-contact material, etc. In addition, calibration takes into account the change in source activity due to continuous decay. For calibration, samples of fluorite ore with a content approximately equal to the maximum and average fluorite content in the deposit under study should be used. The dimensions of the calibration sample must be no less than \(250 \times 250 \times 100\) mm. After completion of work at the deposit, chemical analysis of these samples is performed. From the data on measurement of the induced activity of the standards and the CaF\(_2\) content, a calibration graph is constructed, by which the instrument readings are converted into percent CaF\(_2\) content at the given point (Fig. 2).

Field tests of the instrument [4] were carried out jointly with employees of the Chita Geological Administration. The results of determi-

measurements by the proposed method showed good agreement with chemical analysis, which once again confirmed the practical possibility of quantitatively determining fluorite by neutron activation with subsequent registration of \(\gamma\)-radiation. Let us consider some results in more detail.

The use of a discriminator with a threshold of \(1.33\) MeV makes it possible to avoid the influence of the radiation of most interfering elements—impurities present in fluorite ores. The only exception is \(Al^{28}\), which is formed from \(Al^{27}\) and \(Si^{28}\). The content of both elements, especially the latter, can reach considerable values. The energy of the \(\gamma\)-quanta of the isotope \(Al^{28}\) is \(1.8\) MeV; therefore the discriminator does not free the analysis from its influence. Elimination of this influence is achieved by selecting a time scheme for the analysis based on differences in the half-lives of \(F^{20}\) (\(T_{1/2}=11.3\) sec), \(N^{16}\) (\(T_{1/2}=7.3\) sec), and \(Al^{28}\) (\(T_{1/2}=2.31\) min). The following time scheme of analysis was selected: 40 sec—activation, 10 sec—holding time for the decay of short-lived isotopes, 20 sec—taking the first count (the total activity), 10 sec—taking the reading and resetting when the counting device operates without a reversible mode, 20 sec—measuring the background activity (natural, cosmic, induced).

This scheme ensures registration of activity mainly only from \(N^{16}\) and \(F^{20}\), since the contribution of the \(Al^{28}\) activity, even when its content is 100%, is less than 2%. In practice this conclusion was fully confirmed. In a series of experiments with a quartz hand specimen (\(0.8\% \ CaF_2\)), the magnitude of the induced fluorine activity, according to 35 measurements, did not exceed 10 pulses, whereas in hand specimens rich in fluorite ores (\(\sim 95\% \ CaF_2\)) it was \(\sim 550\) pulses. The reproducibility of the instrument readings was checked with the aid of a standard source of \(\gamma\)-radiation, \(Co^{60}\), which was placed at a definite distance from the detector.

The positions of the source and detector were strictly fixed; then readings were taken from the counting device with time selection switched on. The relative error of reproducibility of the readings over the course of a day after calculation is \(\pm 3.4\%\), which is close to the mean statistical value.

Analysis of the results of the field tests made it possible to solve the following questions: to evaluate the reproducibility of activation analysis, to establish its errors, and to evaluate the agreement between activation and chemical analyses. A comparison of the results of activation and chemical analyses, calculated for individual workings of one of the deposits, can be presented in Table 1.

Table 1

Overall, for this deposit, at an average content of \(36.2\% \ CaF_2\) (according to chemical analysis), the relative error is \(2.4\%\), and the most probable absolute error is \(0.87\%\).

Thus, the agreement of the results of activation and chemical analyses improves significantly in the transition from individual samples to workings and especially to the deposit as a whole. This indicates that the activation-analysis method is free of any systematic errors, while its errors are random and are leveled out when the results are averaged over workings and, especially, over the deposit as a whole.

This conclusion is confirmed by calculation of the results for the balance agreement of the average contents for the deposit and the correlation coefficient:

Table 2

Thus, the deviation of the balance of the results of activation analysis from the chemical analysis is 1.6%. The average content for the deposit, calculated by the two methods, differs very little (0.3–0.6% CaF$_2$). It must be borne in mind that it is practically impossible to achieve complete agreement between the results of chemical and activation analyses, since all our calculations for evaluating the method are based on comparison of activation-analysis data with chemical analysis, taken as the standard, although it itself may distort the results of determinations.

From all of the above it may be concluded that this sampling method is applicable under field conditions for determining fluorite in ores. This provides an economic benefit, since the costs of sampling, processing, and chemical analysis of samples are eliminated, and the promptness of estimating the reserves of the given deposit and putting it into operation is increased. In conclusion, the authors thank the staff members of the Chita Geological Administration, A. I. Kotov and I. I. Shirokii, for carrying out the field tests and for numerous comments on the procedure for working with the instrument.

REFERENCES

1. V. Ya. Bardovskii. Activation logging in prospecting for fluorspar deposits. In: Experience in the Application of Radioactive Methods in Prospecting and Exploration of Nonradioactive Ores. Gosgeoltekhizdat, p. 112, 1962.

2. E. M. Lobanov. Quantitative determination of fluorine in fluorite raw materials and products of its processing by the activation method using a polonium—beryllium neutron source. Izv. AN UzSSR, phys.-math. series, No. 3, p. 68, 1965.

3. V. I. Prokopchik, A. I. Subbotina. On an activation method for determining fluorite in samples of ores and beneficiation products. Atomic Energy, vol. 21, no. 5, p. 414, 1966.

4. L. I. Kotova, I. I. Shirokii. On the application of the neutron-activation analysis method in sampling fluorite veins in replacement mine workings. Questions of Geology of the Baikal and Transbaikal Regions, Issue 2(4), p. 322, Chita, 1967.