The single energy X-ray photon device developed is composition of X-ray photon machine (A), shield bracket (B), fluorescence conversion target (C), intensity adjustment sheet (D), collimator (E), filter sheet (F) and monitoring detector, the schematic layout shown in Figure 1, the three process of X-ray energy modulation schematic is shown in Figure 2.

In order to modulate X-ray energy according to three stages as shown in Figure 2, the device structure designed was shown in Figure 1. We used simulation means of empirical formula and Monte Carlo to obtain the required modulation parameters, such as X-ray machine set voltage, absorbent material species, thickness and so on.

(1) Primary radiation. The primary radiation is produced from ionizing and radiation in the anode target with high speed electronics incident, which are composition with characteristic fluorescence (separate energy) and bremsstrahlung (continuous energy) [11] [12]. These spectrum and intensity of two kinds of X-ray are decided by the incident electron energy, angle and target material, the corresponding expression of cross sections see Formulas (1) and (2).

In the formula, ν is electronic speed, m/s. E is electron energy, MeV. x is an electronic walking distance, m. m is the electron mass, kg. z is electronic charge number, C. Z is the atomic number of the target substance. N is atomic density of the material in per-unit volume, m⁻³.

Based on the above empirical formulas, in low-energy region, the bremsstrahlung and fluorescence yield would increase with the increase of the incident photon energy, but the share of high energy in bremsstrahlung would increase, which also would affect monochromatic radiation of target photon. It is necessary to seek a balance between the fluorescence yield and scattering interference. In our work, under the condition of different setting voltages with electrons accelerated in 20 μm tungsten target and 250 μm beryllium target, using the simulation means of particle transport code MCNP [13], we simulated the primary X-ray spectrum, the simulation curve see in Figure 2, the results can provide data base for subsequent acceleration voltage setting.

(2) Fluorescence radiation. The primary X-rays incident in the target, which deposited energy in the target, and produced fluorescence photons, scattered photons and heat mainly through the photoelectric effect and Compton effect. According to the atomic energy level distribution and incident photon energy characteristics, in the low-energy region, the main fluorescence excited is k layer, and its cross section is determined by the atomic number and the incident photon, the relationship shown in Equation (3). Because of the soft X-ray energy is lower (), in this case, the basic component of Compton scattering composition come from Thomson scatter, whose scattering cross section can be simplified to Equation (4).

In the formula, is classical radius of the electron, about 2.8 × 10⁻¹³ cm.

From the Formula (3), the photoelectric cross section increase obviously with atomic number Z increases of target material, and reduce with incident photon energy increases. The Formula (4) shown the cross sections is independent with the photon energy, but is inversely proportional to the atomic number, it can clearly be seen that the energy of incident photon is to decisive parameters of fluorescence energy. In simulation, the X-ray fluorescence spectrum at different set voltage was shown in Figure 4, under the conditions of the initial radiation spectrum shown in Figure 3, structure and layout as described in Figure 1, the thick copper of fluorescent target C is 200 μm, the copper thickness of intensity control sheet D is 50 μm.

(3) Radiation filtered. As shown in Figure 1, when the X-ray of characteristic fluorescence distribution in (2) of Figure 4 penetrate fluorescence filter F, the photon attenuation will obey exponentially rule as shown in Equation 5. Due to the fluorescent to be modulated is composed by k_α and k_β (L fluorescence yield is very low, can be ignored), in the energy modulation, we select these characteristic target material as filter, which of excitation potential is slightly larger than k_α or k_β, meanwhile, the energy absorbing edge is near the target k_β in the middle of k_α and k_β energy. In order to make full use of material absorb coefficient mutation’s characteristic to strengthen photon absorption in region near k_β energy, and to adjust the fluorescence spectra of transmission, using iterative simulation method to determine the thickness d. The following describe of the Cu target as an example,

K_α1 energy is 8.047 keV, K_α2 energy is 8.027 keV, K_β energy is 8.940 keV, L fluorescent energy are 0.928 keV and 0.948 keV. Selecting Ni as the filter is more appropriate, the excitation potential of the corresponding energy is slightly higher than 8.940 keV, suitable for K_β energy copper absorption fluorescence. In our simulations by MCNP program, when Ni target is setting 50 um, the spectrum with absorption before and after obtained can see as Figure 5.

Figure 5 shows that the K_β energy of Cu fluorescence (8.940 keV) can absorb effectively with fluorescence modulated, and the interference components can been suppressed down nearly one orders of magnitude, the average energy of primary spectrum can modulated from 8.30 keV to 8.13 keV. Under these conditions, the total energy peak region is between 8 keV - 8.15 keV, which is obtained from spectrum modulated by the HPGe detector, the probability of peak count is up to 98% in total spectrum, and the energy monochromatic better than 98.2%, this can be very suitable for low-energy detector energy calibration. According to this principle and simulation methods within 4 keV - 15 keV energy range, in our work selection several monoenergetic photon of the 4.58 keV, 4.95 keV, 8.04 keV, 16.58 keV to design, the detailed parameter can see below summarizes.

The calibration of energy and efficiency is data foundation to ensure detection data reliable. For low-energy detector, the serious nonlinearity of detection efficiency and lack of multi-point energy calibration source, which

brought great difficulties for the absolute fluence of the radiation field measurements. In our work, according to the X-ray energy range, using the Monte Carlo method to get the relative detection efficiency curve, combined with absolute calibration method of Am-241 standard radioactive source, we gain the absolute detection efficiency curve with the relative efficiency’s absolute calibration, the efficiency curve of the detector in setting energy region was shown in Figure 6.

Using the measure system, we complete the X-ray radiation field measure under various operating conditions, the parameters of fluorescence spectrum and fluence rate were summarized in Table 1, the spectrum and fluence in the several typical condition of X source running measured are seen in Figure 7.

Shown in Table 1, using fluorescence converting target and energy modulators, we can provide favorable radiation field with different energy and fluence rate in 4 - 20 keV, to meet the different sensitivity calibration of single photon detector. Figure 7 shows the various conditions to produce monochromatic radiation field is good, 5.01 keV single photon obtained as target combinations of V-23 (vanadium) and Ti-22 (titanium), the energy resolving rate is better than 4%. 8.10 keV single photon obtained as target combinations of Ni-28 and Cu-29, the Energy half-width can reach 0.4 keV. To more than 10 keV photon, due to the fluorescence effect of X-ray machine’s anode and shielding material, some interference photon maybe influence spectrum, such as combination with Zr-40 and Nb-41, the main energy of emission photon is 16.58 keV, but also exist other energy interfere photon, that should be paid attention in application.