Large-area X-ray imaging is one of the most widely used imaging modalities that spans several scientific and technological fields. Currently, the direct X-ray conversion materials that are being commercially used for large-area (> 8 cm × 4 cm without tiling) flat panel applications, such as amorphous selenium (a-Se), have usable sensitivities of up to only 30 keV. Although there have been many promising candidates (such as polycrystalline HgI₂ and CdTe), none of the semiconductors were able to assuage the requirement for high energy (> 40 keV) large-area X-ray imaging applications due to inadequate cost, manufacturability, and long-term performance metrics. In this study, we successfully demonstrate the potential of the hybrid Methylammonium lead iodide (MAPbI₃) perovskite-based semiconductor detectors in satisfying all the requirements for its successful commercialization in synchrotron and medical imaging. This new generation of hybrid detectors demonstrates low dark current under electric fields needed for high sensitivity X-ray imaging applications. The detectors have a linear response to X-ray energy and applied bias, no polarization effects at a moderate bias, and signal stability over long usage durations. Also, these detectors have demonstrated a stable detection response under BNL’s National Synchrotron Light Source II (NSLS-II) 70 keV monochromatic synchrotron beamline.

Introduction

X-ray based imaging techniques are widely used in a wide range of applications. In particular, flat panel X-ray imagers (FPXIs) are now widely used in digital radiography, fluoroscopy, digital tomosynthesis, image-guided radiation therapy, and cone beam computed tomography, and applications in non-medical fields such as non-destructive, cultural heritage investigations, metrology, materials science research, geophysics, and homeland security. In recent days, for the diagnosis of COVID-19, FPXIs are widely used for radiographic chest scans. In this paper, we will discuss the prospective applications of FPXIs, based on hybrid Methylammonium lead iodide (CH₃NH₃PbI₃, or MAPbI₃) perovskite semiconductor, for medical X-ray radiography and synchrotron imaging.

In the medical arena, it is well recognized that more and more patients are subjected to a higher level of accumulated radiation exposure and a concomitant increase in cancer risk²⁵. Indirect and direct X-ray detection principles are the mechanisms by which the X-rays absorbed in the detector are converted to an electrical signal for data processing. Indirect conversion detectors are based on scintillating films where the absorbed X-ray produces photons in the scintillator, and the photons are subsequently detected by a photosensor. Direct conversion detectors are based on semiconductors films where the absorbed X-rays produce electron and hole charge carriers in the detector, which then drift towards the device electrodes under an imposed bias. The most prominent indirect conversion detector materials are CsI and Gd₂O₂S. Amorphous selenium (a-Se) is the only direction conversion material used in commercial FPXIs. Indirect FPXIs with scintillating layers (such as microcolumnar CsI and Gd₂O₂S have high detective quantum efficiency (DQE) and are the detectors of choice for all hard X-ray imaging applications. However, their relatively low image contrast, higher quantum noise, X-ray scattering (low modulation transfer function (MTF), and cone beam artifacts do not allow for X-ray dose reduction beyond current clinical levels. Direct FPXIs, however, have a higher DQE, MTF, and higher contrast-to-noise (CNR) ratio that makes them suitable for imaging fine anatomic structures and, in principle, lowering of X-ray dosage relative to indirect FPXIs. Figure 1 shows the mass attenuation coefficient of CsI and a-Se versus X-ray energy used in indirect and direct FPXIs, respectively. At energies up to 40 keV, a-Se has a higher attenuation coefficient than CsI. Beyond 40 keV, however, the attenuation coefficient of a-Se is lower than that of CsI by close to one order of magnitude, rendering a-Se direct conversion FPXIs impractical for chest and torso imaging as well as in any type of tomography applications that operate at energies > 40 keV. Mammography is the biggest market for current direct FPXIs, yet these FPXIs lose their effectiveness for dense breast tissues. Figure 2 shows a plot of MTF versus DQE for direct and indirect commercial FPXIs. Clearly, the indirect FPXIs have a much smaller MTF than direct FPXIs based on a-Se. The highlighted region denotes our best estimate on the expected performance of next-generation X-ray detectors that combine the best properties of both indirect and direct FPXIs.




Materials engineering and discovery have significantly benefitted from synchrotron-based imaging techniques in understanding the interplay between atomic-scale structures and macroscopic properties. Modern X-ray imaging techniques, such as coherent X-ray diffraction imaging and phase-contrast imaging, require high beam coherence and highly penetrating X-rays. In the current third-generation synchrotrons, such coherence is only attainable for the lower end of the hard X-ray spectrum. The new generation of synchrotrons, however, achieves enhanced coherence at beam energies higher than 40 keV. This increase in coherence benefits a number of scientific research areas: (1) intensely focused and coherent X-ray beams will dramatically advance nanoscale imaging, nano-spectroscopy, and nano-diffraction; (2) coherent diffractive imaging and ptychography will reach spatial resolution approaching atomic length-scales and permit time-resolved studies; and (3) the statistical information derived from photon correlation spectroscopy will open up previously inaccessible timescales and length-scales. These new prospects for non-destructive imaging require high spatial resolution and efficient detectors, which again can only be successfully fulfilled using high sensitivity FPXIs.

A typical three-dimensional lead halide perovskite structure consists of eight octahedra of lead halide with Pb at the center. When the halide ions are changed, the displacement of the octahedra directly adjusts the bandgap of this class of material. The size of the cation (such as CH₃NH₃⁺) is an important factor for the formation of the cubo-octahedral perovskite structure by coordination with the anions. MAPbI₃ has emerged as a new generation of photovoltaic materials that achieved high power conversion efficiencies of around 23% within four years of their development. It is also an excellent candidate for radiation detection due to the presence of high-Z elements Pb. MAPbI₃ has a very high attenuation coefficient, especially at high energies, similar to HgI₂. MAPbI₃ can, however, be produced at much higher thicknesses than HgI₂. The strong spin-orbital coupling of MAPbI₃ perovskites due to the presence of heavy elements and inversion asymmetry in the crystal structure results in Rashba effects. Due to this, MAPbI₃ single crystals demonstrate high diffusion lengths of up to 175 µm and carrier lifetimes of 15 µs. Moreover, the recombination of charge carriers also minimizes in perovskite single crystals and polycrystalline thin films. Studies of Rashba-type effects in perovskites are still in their early stages, and further understanding of the phenomena and impact on charge-carrier recombination and transport is needed. In our study, the mobility-lifetime (µτ) product of holes in 200 µm-thick polycrystalline MAPbI₃ measured using the Hecht relation was found to be of the order of 10^–4 cm²/V, similar to 250 µm thick polycrystalline HgI₂ sensor and three orders of magnitude better than polycrystalline CZT. The µτ product of a single-crystalline MAPbI₃ device is of the order of 10^–2 cm²/V, comparable to the state-of-the-art CZT device. Also, perovskites have a very low density of defects and traps within the bandgap.

Although there has been a lot of progress in recent years on the development of MAPbI₃ based X-ray detectors [e.g.], none of these studies have demonstrated the repeatability and reliability of these devices. In addition, most of these studies are performed using thin sensor layers that significantly limits the X-ray stopping power or have used detector configurations that are incompatible with the widely used large-area high spatial resolution flat panel imaging. There also has not been any stability data, which shows that these detectors can work for a considerable duration of time under X-ray irradiation. Reliability and repeatability have been the main problem with other direct halide semiconductor materials (such as HgI₂). These detectors only demonstrate its excellent properties (such as the low dark current) for a few days under electric bias and then becomes unstable. As a result of this, even after a few decades of research, these are not commercially available for FPXI applications. In the following sections, we will discuss the fabrication techniques and properties of the large-area MAPbI₃-based X-ray detectors that are prime candidates for a new generation of direct FPXIs. The main target of this study was to develop X-ray detectors that can be reliably and reproducibly used for the applications mentioned above.

Results and discussion

Along with the efficiency and sensitivity of the X-ray detectors, the dark current density is a fundamentally important factor for the proper functioning of the readout matrix on which the halide sensor layer is being deposited. Nearly all detectors based on MAPbI₃ have been reported to have a very high leakage current due to the relatively lower MAPbI₃ bandgap. In order to tackle this problem, we focused on minimizing the dark current of the MAPbI₃-based detectors with repeatable results while maintaining high X-ray sensitivity. Table 1 lists the performance of the MAPbI₃-based detector configurations. Table S1 schematically shows the detector configurations. Figure S1 shows the legend for the various layers that are included in each configuration. Two types of polymers (polymers A and B in Figure S1) were used to fabricate the charge transporting layers. MAPbI₃ detectors with different sensor thicknesses were fabricated in this study. The thickness range varied from 200 to 1400 µm. An SEM image of a typical MAPbI₃ layer is shown in Supplementary Figure S2. Figures S3 and S4 show the X-ray characterization setups used in this study. More details on these setups are given in “Methods” section. The characterization experiments were started using a MAPbI₃-based sensor with no additional charge manipulating layers, i.e., the MAPbI₃ layer was biased directly from both sides. In subsequent steps, the charge controlling layers were added between the MAPbI₃ sensor and the electrical contacts.