Electron Defectoscopy

B. A. Kononov

(Presented by the scientific seminar of the Research Institute of EI)

It is quite obvious that there can be no universal method of nondestructive quality control of products and materials suitable for defectoscopy of the entire diverse range of industrial products. It is also obvious that the requirements imposed on the quality of products and, consequently, on inspection are continually increasing; therefore the urgency of the problem of improving inspection methods, refining the limits of their most effective application, and developing new methods of inspection is constantly growing.

In the present communication, the state and certain problems of one of the methods of radiation inspection—defectoscopy by fast electrons, or electron defectoscopy—are considered.

Electron defectoscopy (ED), like any other method of radiation inspection, is based on the dependence of the radiation field (i.e., its spatial and energy characteristics) behind a barrier on its thickness, atomic number, and absorber density, and on the presence of inhomogeneities in it. The region of rational use of ED and the sensitivity of the method are determined by the regularities of the interaction of electrons with matter and, first of all, by the range, scattering, and energy losses of the particles, the principal features of which we shall briefly recall [1–4].

Energy losses. Interacting with the atoms of a substance, an electron loses energy; moreover, the mean energy lost per unit path length of a particle, \(dE/dx\), is called the linear energy loss (LEL), or the stopping power of the medium, while the energy absorbed by a unit mass of the substance is identified with the absorbed dose. LEL can be calculated from the Bethe–Heitler theory or measured experimentally, and at present a sufficient amount of data is available [4–8]. In the range from 1 to 30 MeV, the LEL increases almost linearly with energy, from 1.5 to \(2.5 \div 5\) MeV·g·cm\(^{-1}\) for substances with atomic numbers from 13 to 82; moreover, at energies above 6 MeV, greater energy losses are observed in substances with large \(Z\). In a layer of finite thickness \(\Delta t\), the energy loss \(\Delta E\) will be greater than

\[ \int_{0}^{\Delta t}\left(\frac{dE}{dx}\right)\,dx \]

as a result of electron scattering. Scattering leads to an increase in the electron path in the barrier and to an increase in energy losses. Electron scattering increases with the \(Z\) of the substance and with decreasing electron energy. Therefore, if one considers energy losses in layers of finite thickness for a beam that is not too narrow—a problem of practical interest for defectoscopy—

they prove to be independent of \(E\) and weakly increasing with \(Z\). The medium offers, as it were, a constant resistance to electron penetration: at low energies (1–5 MeV) the LPE are small, but particle scattering is large. Conversely, as the energy increases, the LPE increase, but scattering decreases.

The independence of energy losses from the initial energy of the particle is confirmed by Fig. 1, which shows the relation between the penetration depth into matter of electrons with energy for substances with \(Z\) from 6 to 82. The experimental data used are ours and those of Harder [9]; the numbers next to the curves correspond to the ordinal numbers of the elements.

The range increases linearly with \(E\). From the curves it is not difficult to find the values \(\Delta E/\Delta t\) for different \(Z\), which can be used to calculate the mean energy of the electrons beyond the barrier.

The linearity of energy loss with increasing barrier thickness is illustrated in Fig. 2 according to the results of measurements made by us together with V. M. Antonov [10]. As the barrier thickness increases from 50 to 500 mg/cm\(^2\), the mean energy losses increase linearly. The energy losses depend only weakly on \(Z\), although the chemical composition and crystal structure may affect the slowing of particles within 10% (Fig. 2, curves for Al and KCl).

Fig. 1. Increase of range with increasing electron energy

Fig. 2. Mean energy losses of a narrow beam of electrons in barriers of 50–500 mg/cm\(^2\)

Let us note three consequences of the considered features of the interaction of electrons with matter. First, the energy loss does not depend on the electron energy (in the interval indicated above), and, consequently, electrons of different energies, passing through a given barrier, lose one and the same energy. Second, the energy lost is proportional to the barrier thickness, and therefore, as electrons penetrate into the absorber, their energy decreases linearly with depth. Third, the energy loss in the barrier depends only weakly on \(Z\), which makes it possible to use an electron beam to inspect variables in chemical composition or layered materials.

The thickness of materials \(t\) inspected by an electron beam with good sensitivity lies in the region \(t = (0.6 \div 0.9)R_{\mathrm{e}}\), where \(R_{\mathrm{e}}\) is the extrapolated range of electrons in the material. Using the relation proposed by us between the initial energy and the range of an electron, one can estimate the particle energy necessary for inspecting articles of a given thickness, or for a given maximum

electron energy, to estimate thicknesses accessible to electron flaw detection:

\[ t=(0.6\div 0.9)R_{\mathrm e}=(0.6\div 0.9)\cdot[0.53E-0.1-(2.5E+1)(Z-13)\cdot 10^{-3}], \]

where \(t\) and \(R_{\mathrm e}\) are in g/cm\(^2\), and \(E\) is in MeV.

The sensitivity of the intensity and dose of the beam and of the electron energy beyond the barrier to the thickness or the presence of defects in the object being inspected is substantial. Therefore, as in the case of gamma flaw detection, the following methods of electron flaw detection are possible: ionization (scintillation), spectrometric, and radiography.

The form of the dependences of the number of electrons, the dose, and the energy spectrum on the thickness and atomic number of the absorber is well known [1, 3, 8] and determines the sensitivity of the inspection method.

The sensitivity of the ionization inspection method can be represented as

\[ |\delta x|_{\min}=\xi\cdot\omega/a(Zx), \tag{1} \]

where \(|\delta x|_{\min}\) is the minimum size of the detected defect \(\Delta x\), referred to the thickness of the article \(x\); \(\omega\) is the sensitivity of the detector; \(\xi\) is the reliability coefficient, usually taken equal to 3; \(a(Z,x)\) is the ratio of the relative change in the radiation intensity behind the specimen to the relative change in the specimen thickness that causes it. Expression (1) does not take into account the ratio of the areas of the defect and the collimator (i.e., it is assumed that \(S_{\mathrm{def}}/S_{\mathrm{coll}}=1\)), which can be allowed for in the usual way.

For a monoenergetic electron beam with an energy of several MeV, for example, \(a(Z,x)\) can readily be obtained in the range from 20 to 30, which at \(\omega=0.05\) makes it possible to attain a sensitivity of the order of 0.5%.

The most important circumstance is that this high sensitivity can be achieved in inspecting layered materials, practically for any combination of thicknesses and atomic numbers of the substances of the layers, which is unattainable by other flaw-detection methods.

If X-ray film is used as the detector, inhomogeneities in the article can be documented [11]. The difference from radiography consists in the smaller range of thicknesses penetrated,

\[ t=(0.6\div 0.9)R_{\mathrm e}, \]

at which the correlation between film blackening and defect size is preserved for a given electron energy, and also in greater image blurring, which increases with distance from the surface from which the electrons emerge. At present, the limits of the appropriate application of electron radiography have not been determined.

A sensitivity several times higher has been achieved with the spectrometric method of electron flaw detection. The author, together with Yu. M. Stepanov, has analyzed in detail the possibilities of the spectrometric method as applied to the inspection of homogeneous and layered materials; this will be the subject of an additional communication. It is shown, in particular, that the sensitivity of a flaw detector is determined by the magnitude of the probable losses of electron energy in the article being inspected and by fluctuations of the beam energy. The magnitude of the minimum detectable defect does not depend on thickness and decreases only slightly with increasing atomic number of the target material, amounting to \(3\div 5\) mg/cm\(^2\), or 0.1% of the article thickness.

Each of the electron flaw-detection methods considered above, and especially the spectrometric method, requires high stability of the intensity and energy of the electron beam, as well as the possibility of smooth adjustment of the initial energy and good monoenergeticity of the electrons.

Which electron accelerators best satisfy the requirements

electron flaw detection? Let us consider some parameters of the accelerators presented in the table.

Table

Some parameters of electron accelerators important for flaw detection

Direct-action accelerators have high monoenergeticness and stability, but a low electron energy. Linear accelerators are poor because of the monoenergeticness of the beam, the complexity of control, and their high cost. The microtron satisfies all the requirements except for smooth regulation, which is complex. Its use is advisable if high inspection productivity is required. The betatron makes it possible to obtain, comparatively simply, a stabilized monoenergetic beam with easily regulated energy. However, in installations manufactured in series or in small batches, the repetition rate of radiation pulses is 50 Hz, and the intensity is of the order of \(10^{10}\) electrons/\(\text{cm}^2\) sec, which limits the productivity of inspection—for example, of welds—to a speed of up to 1 m/min. The area of expedient use of the electron beam of a microtron or betatron is subject to clarification and techno-economic analysis. Undoubtedly, at present there are sources of fast electrons that meet the requirements of electron flaw detection.

The situation is more complicated with detectors. At the Tomsk Polytechnic Institute a methodology has been developed for selecting a detector in order to achieve greater sensitivity. The methodology is based on creating conditions for the fullest possible use of that part of the transmitted electron beam which carries the greatest specific volume of information. In some cases the use of scintillation detectors is sufficient, but apparently a special detector must be created for electron flaw detection.

LITERATURE

1. A. A. Vorob’ev, B. A. Kononov. Passage of Electrons through Matter. Tomsk, TSU Press, 1966.
2. V. M. Antonov, B. A. Kononov, V. V. Evstigneev. Izv. AN SSSR, vol. XXVIII, No. 2, 400, 1964.
3. B. A. Kononov et al. Dosimetry of large doses. In: Collection of Reports. Tashkent, FAN Press, 1966.
4. B. A. Kononov et al. Izv. AN SSSR. Ser. fiz., 28, 400, 1964.
5. M. I. Berger, S. M. Seltzer, Publ. 1133, NAS-NRC, 1964.
6. H. E. Hall et al., Phys. Rev., 115, 633, 1959.
7. K. I. Vancamp, V. I. Vanhuse, Zeit. Phys., 211, 152, 1968.
8. B. A. Kononov, E. A. Gusev. Fizika. 1969, No. 6, 12.
9. D. Harder, Poschet, Phys. Let., 24 B, N 10, 519, 1967.
10. V. M. Antonov, B. A. Kononov. Abstracts of the VI Interuniversity Conference on Electron Accelerators. Tomsk, TSU Press, 1967.
11. B. A. Kononov et al. “Defektoskopiya,” 1967, No. 4, 81; 1968; No. 4,