ON IMPROVING REPRODUCIBILITY IN ACTIVATION ANALYSIS

R. P. Meshcheryakov, N. A. Skuratov, Yu. M. Stolbov

(Presented by the scientific seminar of the Research Institute of Nuclear Physics)

Determination of the content of the isotope under study by the method of activation analysis usually reduces to obtaining a radioactive isotope and measuring its activity.

The induced activity is a function of the flux of incident particles. If the latter varies in space and time, then in practice it is very difficult to establish the relation between the magnitude of the induced activity and the content of the element of interest to us. Direct methods of flux registration are, as a rule, associated with approximate integration and do not take into account possible changes in particle energy; therefore the flux incident on the sample under study is most often judged from the activity of a reference sample. To avoid errors arising from differences in half-lives and reaction cross sections, the same element as the one being studied is used as the standard. If attenuation of the flux in the test and standard sample (“infinitely thin” samples) is not taken into account, then the irradiation conditions may be assumed to be identical. For samples of “infinite thickness” (the case of heavy charged particles), irradiation of both samples by one flux is excluded. In this case it is expedient to make use of the circumstance that, for a uniform flux, it is sufficient to require constancy of the area of the irradiated samples.

If the magnitude and anisotropy of the flux change sufficiently slowly, it is expedient to carry out averaging mechanically, for example by rotating the samples. At a constant rotation speed the process is analogous to approximate integration, and it can be shown that, with a static character of the flux oscillations, the ratio of the induced activities in samples identical in chemical composition will remain constant independently of the frequency and amplitude of the flux oscillations. Noticeable deviations will be observed only if the frequency of flux variations is a multiple of the rotation frequency of the samples.

For practical verification, a mechanism was tested by means of which the samples were rotated at a speed of about 4000 rpm in a flux of deuterons accelerated in a cyclotron.

The mechanism is shown in Fig. 1.

A shielding plate 2 mm thick has two identical openings strictly symmetrical with respect to the axis of rotation. The size of the openings was maintained with high accuracy. The diameter of the irradiated sampl—

holes of larger diameter; therefore, the ratio of the cross sections of the incident fluxes remains strictly constant.

Measurements of the β-activities obtained in this way showed that their ratio is reproduced with an accuracy of 0.05%.

Fig. 1

When measuring γ-activities with a scintillation γ-spectrometer, the main error is introduced by the unstable operation of its units: the photomultiplier tube, amplifier, and discriminator. This is explained by the fact that the position of the spectrum relative to the recording window shifts because of the unstable operation of the spectrometer. However, if the spectra of the activities being measured differ only in intensity, then the same window can be used for the measurement. This makes it possible to substantially reduce the influence of instability by measuring the samples successively, switching between them at high frequency. The greatest error in this method of measurement will occur when the intrinsic frequency of the instability is a multiple of the sample-switching frequency.

For an experimental check, we carried out measurements of the ratio of two γ-activities.

The measurement scheme is shown in Fig. 2.

Two samples were mounted on opposite sides of a steel disk 44 mm thick and 230 mm in diameter, which simultaneously served as a shield from the sample whose activity was not being measured at a given moment. An angle-of-rotation sensor was rigidly attached to the common shaft with the disk; its signals controlled the linear gate devices of the electronic commutator. The disk rotated at a speed of about 2700 rpm; the γ-quanta, through a lead collimator, fell on a NaJ (Tl) crystal measuring 100×100 mm. Signals from the FEU-49 photomultiplier were amplified by a USh-10 amplifier. The spectrum obtained at the amplifier output was analyzed by an ADD-1 single-channel discriminator. Pulses from the discriminator were fed to the inputs of two linear gate devices, each of which passed pulses only from its own sample. The outputs of the line-

throughput devices were connected to PS-10000 counting instruments.

The apparatus used had a stability of 3–4%. However, the measurements performed showed that the measured ratio of $\gamma$-activities is reproduced with an accuracy of 0.4%.

Fig. 2

Labels in the figure: sample 1; disk; sample 2; rotation-angle sensors; Pb; NaI(Tl); photomultiplier; preamplifier; amplifier; discriminator; electronic commutator; counting instruments 1, 2.

It should be noted that, when comparing two samples using one line of the $\gamma$-spectrum, exact reproduction of the position of the samples on the disk is necessary. When measuring the ratio between two lines of the $\gamma$-spectrum for two samples, exact reproduction of the position is not required.