Bismuth Sulfide Nanoflowers for Detection of X-rays in the Mammographic Energy Range

The increased use of diagnostic x-rays, especially in the field of medical radiology, has necessitated a significant demand for high resolution, real-time radiation detectors. In this regard, the photoresponse of bismuth sulfide (Bi2S3), an n-type semiconducting metal chalcogenide, to low energy x-rays has been investigated in this study. In recent years, several types of nanomaterials of Bi2S3 have been widely studied for optoelectronic and thermoelectric applications. However, photoresponse of Bi2S3 nanomaterials for dosimetric applications has not yet been reported. The photosensitivity of Bi2S3 with nanoscale “flower-like” structures was characterized under x-ray tube-potentials typically used in mammographic procedures. Both dark current and photocurrent were measured under varying x-ray doses, field sizes, and bias voltages for each of the tube potentials – 20, 23, 26 and 30 kV. Results show that the Bi2S3 nanoflowers instantaneously responded to even minor changes in the dose delivered. The photoresponse was found to be relatively high (few nA) at bias voltage as low as +1 V, and fairly repeatable for both short and long exposures to mammographic x-rays with minimal or no loss in sensitivity. The overall dose-sensitivity of the Bi2S3 nanoflowers was found to be similar to that of a micro-ionization chamber.


Energy dispersive spectrometry
The chemical composition of the nanoflowers was confirmed with energy dispersive spectrometer (Oxford Instruments Microanalysis System INCA Energy 350) as part of the scanning electron microscope (SEM, JOEL JSM-6460). The sample was dispersed in ethanol and dropcasted on aluminum foil attached to a silicon substrate. Results of the elemental analysis are presented in this section (Figures S1, S2, and Table S1).   All results in weight% Powder x-ray diffraction analysis Figure S3 shows the powder x-ray diffraction (XRD) pattern of the nanoflowers recorded on a Bruker D8 Advanced X-ray diffractometer using the parameters listed in Table S2. The XRD pattern shows polycrystalline nature and all the peaks in the XRD pattern can be indexed to orthorhombic Bi2S3 (JCPDS 17-0320) with no indication of impurities ( Figure S3).   Figure S4 shows the spectra for x-ray tube-potentials 20, 23, 26 and 30 kV using the SpekCalc simulation software. The "air thickness" parameter of 154 mm was used to account for the FSD of 150 mm (from the x-ray setup), and an additional 4 mm from the distance between the tip of the 1 cm cone and the surface of the test device. The mean x-ray energy for each of the tubepotential was also obtained from the simulations.

WinXCom simulation
The mass attenuation coefficients (cm 2 /g) of Bi2S3 for the mean x-ray energies: 9.78, 10.6, 11.4, and 12.4 keV (output from SpekCalc simulations) were obtained using the WinXCom program as shown in Figure S5. S-5

Auto-Zeff simulation
Energy-weighted effective atomic number (Zeff) of Bi2S3 nanoflowers was estimated using Auto-Zeff software ( Figure S6). The Zeff was found to be in the range of 45 to 46.5 for the x-ray energies used in this study (9.78, 10.6, 11.4, and 12.4 keV).

Figure S6
: Energy-weighted Zeff of Bi2S3 for x-ray energies from 10 to 100 keV as simulated in Auto-Zeff program.

Energy bandgap calculation
The bandgap of Bi2S3 nanoflowers was calculated from the diffuse reflectance spectra (%R versus wavelength) measured using UV-Vis-NIR spectrophotometer (Shimadzu UV-2501PC).
The absorbance (F(R)) was calculated from the diffuse reflectance spectrum using the Kubelka-Munk function 1 , and the energy bandgap (Eg) was then estimated by substituting F(R) in the Tauc equation.

S-6
Kubelka-Munk function is given by: where, R is the measured diffuse reflectance.
For the UV-Vis-NIR range, the optical absorption coefficient (α) can be determined by the Tauc equation: where, hν is the photon energy, Eg is the bandgap, and A and n are constants. 2 Since Bi2S3 is a direct bandgap material, n=1/2. By substituting α with F(R) in equation (2), the Tauc plot ((F(R) * hν) 2 vs. hν) was obtained and the energy bandgap (Eg) was then estimated by extrapolating the linear portion of the plot to the energy axis as shown in Figure S7. The bandgap was found to be 1.33 eV.  x-rays. The rapid loss of photocurrent was evident for all exposures, however, the overall photoresponse was fairly repeatable for both short-and long-term exposures ( Figure S10). Figure S9: Photoresponse of the device with commercial Bi2S3 powder exposed to different xray tube potentials 20 to 30 kV.
S-9 Dose dependence and sensitivity of the commercial Bi2S3 powder was also evaluated under 20 and 30 kV x-rays ( Figure S11). As expected, commercial Bi2S3 was also found to be appreciably sensitive to dose changes at the tube potentials as low as 20 and 30 kV. However, in addition to the rapid loss in x-ray induced signal over time, the overall sensitivity (percentage change in photocurrent at maximum dose with respect to the minimum) was found to be relatively lower at both at 20 kV (168.5% for the powder and 241% for the nanoflowers) and 30 kV (217% for the powder and 241% for the nanoflowers) in comparison with the nanoflowers at the same tube potentials and doses. S-10