X-RAY SPECTROMETRY FROM A MICROCT SYSTEM

X-ray production is of fundamental importance for several applications, among these is the Microcomputed tomography technique (microCT). It is used to conduct studies in different areas of research such as industry, dentistry, archaeology and biomedicine. In this study, X-ray spectra were acquired with different filtering configurations in order to better understand the energy behavior of a X-ray beam used in a microCT system. Unfiltered spectra were acquired for voltages of 20 up to 80 kV and filtered spectra for voltages from 40 up to 120 kV. As results, the mean energy values of each filtered and unfiltered beams from 40 to 120 kV were obtained. These values were calculated based on the spectra evaluation, as well as the value of attenuation that each filtering systems generated in the spectra of 60 and 70 kV. To compose the filtration systems metallic physical filters of aluminum and copper were used. Spectra acquisition was done using a semiconductor detector, model XR-100CdTe, manufactured by Amptek.


INTRODUCTION
The capacity of X-rays to interact with matter makes usage of this type of radiation applicable in various situations. This capacity varies according to the density of the substance to be observed. In addition, the penetrating power of X-rays also depends on the energy spectrum of the beam.
Amongst its many applications are medical and industrial radiography. Computer tomography (CT) is widely used for medical purposes, and microcomputed tomography (microCT) are both of great importance to various fields, among them, but not limited to, dentistry, archeology and biomedicine [1].
In many situations, in both CT and microCT, filters are added to the system as a way of removing low energy photons to obtain better image results and/ or to reduce patient exposure to radiation. The detector in a microCT system has a limit of events it can process per time unit, limit that when reached causes saturation of said system. In order to avoid detector saturation under certain circumstances, metallic filters can be used to eliminate low energy photons from the energy beam. The use of metallic filters may also aid in the correction of what is known as the beam hardening artifact, which compromises the quality of the microtomography image [2]. This paper aims for a better understanding of the spectrum generated by an X-ray tube used in a microCT system and the way the spectrum is affected by the use of filters. A tungsten target, microfocused tube with maximum tension of 300 kV was used in this experiment.
The spectra were acquired in various filter and beam configurations, with different values of maximum energy. Among the tensions used are 60 and 70 kV, which are used in various works [3][4][5][6][7][8]. In [3] a 1 mm thick aluminum filter was used for dentistry tomography, while in [4] no filter was used in the tomography of rabbits. Reaching for a better result in image quality, the use or not of filters must be evaluated in each circumstance. These spectrums were obtained by a cadmium telluride (CdTe) semiconductor type detector, as in previous works before [9,10] and the attenuation values corresponding to each filtering system were calculated for certain energy intervals, as well as the average energy associated to each analyzed X-ray beam.

MATERIALS AND METHODS
The choice of a cadmium telluride (CdTe) semiconductor detector is justified by its high stopping power, a determining factor for the realization of this study due to its X-ray attenuation coefficient [11]. The term stopping power refers to the average loss of energy per unit length in a specific environment and it depends on the type of material and incident radiation [12]. The XR-100CdTe system, manufactured by Amptek, consists on a high performance X-ray detector, a preamplifier and a cadmium telluride cooler measuring 5 x 5 x 1 mm mounted on a thermoelectric, two stage cooler [11].
The X-ray tube used is a microfocused, reflection type, with 300 kV maximum tension and a tungsten target [13] that comprises the v|tome|x microtomography system, manufactured by General Electric (GE) [14].
For spectra acquisition the detector was installed within the tomographer in a stable manner, and it was kept aligned with the X-ray window throughout the entire experiment. To align the CdTe detector window and the X-ray tube, the detector system of the tomographer was utilized, which allowed the visualization of the entrance window of the detector and its alignment with the tube through the image shown on the screen of the tomography operation monitor. Figure 1 shows the tube-detector system used. A lead collimator aligned to the tube and the detector window was used to reduce scattered radiation at the entrance of the detector window as a way to lower incident beam intensity in the detector. This intensity reduction is necessary to reduce dead time during spectra acquisition which also avoids imprecisions on the resulting spectrum.
The calibration of the detector was made using a 241Am source (95mCi), with a known spectrum. The source emits alpha radiation as well as gamma radiation with energy peaks of 26.3 keV and 59.6 keV [15], so it was possible to relate each of the multichannel values to a specific energy value.
Aluminum and copper filters used were 0.5mm and 0.1mm thick, their areas larger than the tube window as to ensure the entire beam would pass through the applied filter. The filtering system was placed alongside the tube window, parallel to the detector and tube windows.
A computer containing the acquisition software ADMCA, from Amptek, was connected to the detection system, which made it possible to analyze and storage each of the acquired spectra.
To obtain spectra of satisfactory precision level a 300 second acquisition time was established as well as a dead time limit of 13%. The current inside the X-ray tube throughout the entire experiment was 5 . A threshold was fixated on channel 30 in order to discard any relatively small signal originated from electronic noise. If not discarded, noise signals of this nature would interfere with the study, for they are not originated inside the X-ray tube. After spectra acquisition the average energy values ( ) of each X-ray beam was calculated using weighted average (1): Where represents the energy at a certain point of the spectrum and is the number of counts at the same point. The attenuation value of the beam was also calculated for each filtering system.
Attenuation is given by (2): Where is the intensity of the filtered beam and 0 its intensity prior to its passage through the filtering system. Both average energy and attenuation values were calculated using output data generated by the spectra acquisition software.

RESULTS AND DISCUSSIONS
Results of this work comprise acquired and analyzed energy spectra from a X-ray tube using different tensions and filtering configurations.

Unfiltered acquired spectra
The presence of characteristic energy peaks around 10 keV was observed in all spectra, as well as the appearance of new characteristic energy peaks in the 80 kV spectrum. In Figure 2 it is possible to observe, by inspection of the characteristic energy peak values, that the tube under analysis has a tungsten target [16]. Possible characteristic peak values variations from the ones established in known related literature [17,18] are due to experimental imprecisions.

Energy spectra with application of filters
Aluminum and copper were the materials chosen as filtering systems. Both are of great importance to the microCT technique, for they are largely used in microtomography [19].
The acquisition of spectra with voltages above 80 kV were possible using filtering systems, with no reliability loss, which is justified by the attenuation caused by the filter and the non-saturation of the detector under these conditions [12]. Using filters thicker than 0.5 mm made studying greater potential differences possible, 120 kV being the maximum one, as shown in Figure 4. Spectra with maximum energy up to 50 keV were almost entirely attenuated. In these cases, the maximum count was around 2.5% of the maximum for potential difference of 120 kV, as shown in  For aluminum only filters the maximum value of applied tension obtained was 90 kV, with the purpose of not exceeding dead time limit. The potential limit was slightly higher than the one in the unfiltered case. The difference between limits of voltage reached in copper and aluminum is due to higher attenuation coefficient of copper when compared to aluminum [20]. Figure 6 presents the 40 to 90 kV spectra acquired using a 1.0 mm aluminum filter.  To improve visualization of the spectra in Figure 7, where copper is used in the filtering system, Figure 8 shows the magnification of the area that contains it.

Energy spectra at 70 kV tube voltage
In Figure 10, as in Figure 7, the spectra for each filtering system are presented, however, in this case, at tube voltage of 70 kV.

Average energies
Average energy values from each acquired spectra were also calculated. These values are presented in Figure 12 by filtering system and the X-ray tube voltage. The application of filters caused a similar effect in terms of beam energy average in all analyzed tensions ( Figure 12), which can also be observed in Figure 13.