Introduction
Among ionizing radiation, including alpha (heavy charged particles), beta (fast electrons), gamma (electromagnetic wave), and neutron (uncharged particle), monoenergetic gamma rays that undergo state-to-state transitions (due to the well-defined energies of the nuclear states) pass through matter from several millimeters to several centimeters without any change in radiation properties, allowing the identification of radionuclides (called gamma spectroscopy) [1]. Over the past few decades, gamma-ray detectors have been developed and widely used for radioactive contamination monitoring, spent fuel management, geological formation inspection, non-destructive evaluation, and nuclear safety. After the Fukushima Daiichi nuclear accident, there has been increasing interest in gamma-ray imaging systems that can visualize the distribution of radioactive sources and provide real-time scanning of radioactive material contamination in a wide area by estimating the angular direction of gamma rays incident on the system using image reconstruction algorithms [2–5]. Furthermore, the non-destructive examination of spent nuclear fuel (e.g., residual heat and axial burnup distribution) can be implemented by measuring the net count rates of fission products and their isotopic ratios (e.g., 154Eu/137Cs and 134Cs/137Cs), which have different half-lives such as 30.1 years for 137Cs, 2.1 years for 134Cs, and 8.6 years for 154Eu [6–8]. In addition, nuclear non-proliferation can also be implemented by detecting specific energy of gamma rays from special nuclear materials [9], including 235U (185 keV) and 239Pu (129 keV and 413 keV).
Commercial gamma-ray detectors have been developed using a single crystal, including high-purity Ge (HPGe) [10]. HPGe has a low bandgap (0.67 eV), which makes it highly susceptible to thermal noise and requires liquid nitrogen cooling for operation. Therefore, there has been increasing interest in developing gamma-ray detectors operable in room temperature, such as single-crystal CdZnTe (CZT), CdTe, and TlBr [11, 12]. Typical methods for growing such single crystals include the Czochralski and Bridgman methods [13]. However, commercial single-crystal growth methods require high-temperature processes and high costs, and scaling up the size of single crystals usually leads to a growth bottleneck due to poor crystallization quality (or randomly distributed polycrystalline growth patterns), low repeatability, high cost (long growth time), and cracking. Therefore, it is fundamental to develop novel room temperature operable gamma-ray detectors that are easy to fabricate and capable of high-performance gamma spectroscopy.
As an alternative to single-crystal growth, attempts to develop gamma-ray detectors using inorganic nanoparticles (NPs), including nanocrystals (NCs), with high atomic number (Z) are increasing worldwide due to their tunable optoelectronic properties and many different fabrication methods [14–16]. Note that the NP is often used for metal and wide-gap nano-sized particles with either an amorphous or a crystalline structure, and NC is for nano-sized semiconductors with a crystalline structure (if the size of the particle is smaller than the Bohr radius of its exciton, it is called a quantum dot [QD]) [17]. The two terms agree on the idea of nano-sized particles, with sizes typically less than 100 nm, hereafter referred to as NPs. Several studies have reported that the optoelectronic properties of NPs, such as optical band gap and luminescence wavelength or intensity and charge transport properties, can be tuned by the NP size, the capping ligands [18, 19], the solvent or polymer matrix [20, 21], and the core-shell structure [22–24]. In addition, a solution-based synthesis of NPs is a low-temperature process [25–27], and various NP compounds, such as CdS, CdSe, PbSe, InP, and CsPbBr3, have been used in the development of photovoltaic and light-emitting diode (LED) devices [28–30]. The devices have been developed at a low cost, high performance, and large scale in a flexible and bendable form using various device fabrication techniques such as doctor blade coating, spin coating, dip coating, and inkjet printing [31–33].
It is worth noting that NCs can exhibit highly efficient carrier multiplication (CM) by absorbing a single high-energy photon greater than at least two times the bandgap (Eg). The CM results from the lack of a well-defined translational momentum (or weak charge-phonon coupling) of the quantum-confined energy levels, including the strong Coulomb interaction between carriers. Subsequently, multiple excitons are generated in an impact-ionization-like process, improving the charge carrier generation statistics [28, 34–36]. The relaxation of the translational momentum conservation contributed to the reduced CM threshold in NCs. Specifically, the CM onset in colloidal PbSe QDs is 3Eg–4Eg [34] compared to >6.5 Eg in bulk PbSe [37]. Nevertheless, considering the limited energy of solar light, the CM phenomenon could not be effectively exploited to enhance the efficiency of photovoltaic devices due to the high CM onset energy (e.g., 2.9Eg for 4.1 nm PbTe NC [38] and 3.7Eg to 4.7Eg for 4.6–6.6 nm PbSe NC [39, 40]). In contrast, ionizing radiation, such as gamma rays, has enough energy to remove an electron from an atom, producing energetic secondary electrons. These radiation-induced secondary electrons have an energy much larger than the bandgap, producing excitons by cascade-like relaxation along their trajectory. A gradual increase in CM efficiency was observed with higher exciton energy [41].
Although previous studies for the application of a photon (from a few keV to tens of keV X-rays) counting detector based on NPs are rich and extensive, these devices have not found significant use in the radionuclide analysis, which is the recognition of its monoenergetic and relatively high energy of photons (from tens of keV to few MeV gamma rays). These previous studies are not reviewed here; one can refer to the literature, which gives an excellent perspective on this field [42–44]. Likewise, although many recent studies have been reported that failed to implement gamma spectroscopy [45–48], for brevity, we focus primarily on uniform and thick (few hundred microns to millimeters) nanocomposite detectors with a high effective atomic number (Zeff), which in the last decade have proven gamma spectroscopy with improved energy resolution by considering an efficient exciton energy transfer mechanism.
This review describes the technological trends in the fabrication of room temperature nanocomposite gamma-ray detectors and presents representative results of gamma-ray spectroscopy realized by indirect (by radiation-induced luminescence or scintillation) and direct (by radiation-induced current) detection mechanisms. Furthermore, we would like to discuss in detail the scintillator based on luminescent NCs and non-luminescent (or weakly luminescent) NPs, the band-alignment engineering for efficient exciton transfer, and the optical properties and gamma-ray measurement performance of the scintillator as a function of the concentration of the contained NPs and organic dyes. In addition, we review fabrication recipes of direct detectors for nanocomposite assembly with fewer cracks and thicknesses ranging from a few tens to hundreds of microns, self-assembled NC colloidal solids with millimeter scale, NC packing with tailored shapes and sizes, and techniques for long-term stability at ambient conditions.
Nanocomposite Indirect Detectors
Developing composite arrays of NPs into scintillators has been relatively more dominant modality being investigated by the radiation detection community, utilizing their accelerated superfluous light emission due to the suppression of non-radiative loss processes. Development of nanocomposite-based scintillators is often conducted by embedding scintillating NPs into optically transparent light propagating media, such as organic scintillators. The main challenge in this respect is the optical self-absorption by the nanoparticles, the matrix, and the innumerable interfaces as the scintillation photons meander their way to the readout surface [14].
As it is certainly desired to observe the full-energy absorption events of gamma rays by a nanocomposite scintillator, the following aspects of the scintillator are mainly required:
(1) Improving the effective atomic number (Zeff) of nanocomposite scintillators; This includes increasing the weight percent of inorganic NPs and exploring new NP materials with high Z numbers.
(2) Engineering energy band alignment to facilitate exciton energy transfer; The hot electrons following the photoelectric effect will excite π-electrons of adjacent molecules, such as organic matrix, and the appropriate band alignment between donors (organic matrix) and acceptors (NPs or fluorescent dyes) is required to transfer exciton energy via Förster resonance energy transfer (FRET).
(3) Mitigating the aggregation of NPs to increase transmittance; NPs tend to aggregate at higher concentrations and during the thermal conditioning to obtain a polymerized nanocomposite; larger agglomerate size leads to low transmittance due to enhanced Rayleigh scattering.
(4) A large volume of scintillator; A nanocomposite scintillator with a high weight percent of NPs has a thickness limit due to NP aggregation and transmittance loss. Therefore, a large volume scintillator with a moderate weight percentage of NPs may be an alternative.
1. Fabrication of Nanocomposite Scintillator
Since the stability of NPs determines the scintillator’s performance, protection from moisture, oxygen, and heating is critical in the fabrication of nanocomposite scintillators. Continuous exposure of photoexcited NPs to O2 leads to irreversible degradation, such as NP coalescence followed by surface ligand detachment, surface trap formation, core dissociation, and undesirable changes in optoelectronic properties such as spectrum broadening [49]. In particular, the perovskite structure such as CsPbBr3 is susceptible to moisture because the formation of lead hydroxide (Pb(OH)2) is induced and then decomposed to form lead oxide (PbO) and water (H2O) molecules, leading to further hydration [50]. Thermal energy is another factor in ligand dissociation due to the weak binding energy between capping ligands and NP surfaces.
NP stability can be improved by ligand design [51, 52], protective coating [53, 54], core/shell structure [55, 56], ligand crosslinking [57, 58], and polymer-NP composites [59]. Kim et al. [51] demonstrated improved thermal stability by replacing an X-type ligand (e.g., oleic acid [OA]) with a Y-type ligand (e.g., dodecanethiol), which has strong covalent bonds with atoms on an NC surface. Jo et al. [60] observed that oxide overcoating (i.e., In2O3) on InP/ZnS NC improved the photostability. Lim et al. [23] reported that increasing the shell thickness can improve the quantum yield of CdSe/CdZnS NC film samples. Sun et al. [61] reported that the use of cross-linkable 4-vinylbenzyl-dimethyloctadecylammonium chloride (V18) ligands can crosslink between MAPbBr3 NCs via thermal treatment with a radical initiator, thereby improving the photoluminescence quantum yield (PLQY) by passivating surface traps. Direct crosslinking between NPs and polymers is relatively favorable because a short intermolecular distance (due to strong covalent bonding) can result in lower moisture and oxygen permeability [49].
The fabrication of nanocomposite scintillators requires not only stability, but also high transparency and high loading of well-dispersed NPs. Such high loading is prone to agglomeration of NPs due to strong interparticle attraction. Uniform dispersion of NPs in monomers followed by polymerization has been reported to obtain transparent nanocomposites with high loading [62]. Since simple mixing of as-synthesized NPs with most polymers results in non-uniform particle distribution or aggregation, surface modification, such as replacement of the capping ligands, is required for favorable binding to the polymer matrix, which will be discussed in detail in the following section.
1) Polymerization for stability enhancement and high-loading NPs
Jin et al. [63] observed that YbF3 with an OA capping ligand can be dissolved in nonpolar solvents such as toluene, chloroform, and hexane, to form clear solutions as shown in Fig. 1A. However, after polymerization with the monomer vinyl toluene (VT), the YbF3 NPs were segregated at the bottom of the vial, and cracks were formed. Note that the NPs with surfactants, including phosphonic acid, fatty acid, thiols, and amines, can also be dispersed by the weak force in the monomer, such as VT [64]. The polarity of the capping ligands, such as OA, commonly used in the synthesis of NPs, is nonpolar, but that of the polymer matrix, such as polyvinyl toluene (PVT), is polar. With increasing NP loading, the large discrepancy in polarity can lead to the exclusion of NPs from the polymer [65]. Thus, two separate domains (polymer and agglomerated NPs) are created due to the strong van der Waals force of attraction between polymer chains and the strong interparticle interaction of NPs [66, 67].
Vaidya et al. [68] demonstrated that the use of a crosslinker (e.g., divinylbenzene [DVB]) could prevent the segregation of NPs during polymerization by crosslinking the linearly growing polymer chain, as shown in Fig. 1B. The figure compares the luminescence image of polymer/NCs bead composites with and without DVB obtained by confocal laser scanning microscopy. The emission wavelengths of 520 nm and 620 nm are from CdSe/ZnS with core sizes of 2.6 nm and 5.2 nm (and a fixed shell thickness of 2–3 nm). The absence of the crosslinker resulted in many bright spots where the aggregation of NCs, and a red shift of the emission peak occurred. Fig. 1C and 1D support this as shown in the transmission electron microscopy (TEM) images corresponding to the nanocomposites with and without the crosslinker. The research findings concluded that a uniform dispersion of the NCs in the monomer solution can be maintained during the polymerization because DVB brings the fast kinetics of crosslinking to the terminal groups of the polymer chains.
Covalent bonding between NPs and monomers has been reported not only to achieve high packing density of films, but also to disperse NPs uniformly without aggregation. Sahi et al. [69] demonstrated highly transparent and high loading (30 wt%) of LaxCe1-xF3 NPs-polystyrene nanocomposite scintillator with surface modification of LaxCe1-xF3 NPs. Capping LaxCe1-xF3 NPs with hexadecyl-p-vinylbenzyl(dimethyl)ammonium chloride (HVDAC) was designed because it provides a long hydrocarbon chain to allow homogeneous dispersion in styrene and copolymerizes NCs with styrene by the thermally-initiated polymerization process. In addition, Jin et al. [63] proposed partial ligand exchange to prevent NP aggregation during polymerization in which OA ligands are partially replaced with bis[2-(methacryloyloxy)ethyl] phosphate (BMEP). YbF3 NPs were dissolved in chloroform, and BMEP in chloroform was added dropwise for 3 hours. The mixture was left in an ice basin and then dried to obtain the BMEP-exchanged NPs. Fourier transform infrared (FTIR) spectra was then used to confirm the ligand exchange, as shown in Fig. 2A. The TEM image given in Fig. 2B confirmed a fairly homogeneous dispersion of the 45 wt% BMEP-exchanged NPs in the PVT matrix. A 1 mm thick nanocomposite scintillator containing 63 wt% YbF3 NPs and organic fluorescent dyes was as transparent as a neat PVT monolith, as shown in Fig. 2C. Therefore, the remaining OA ligands provide good solubility of the NPs in the monomer solution via a weak force, while BMEP allows the P–OH head group to anchor on the NC surface, and the tail end has a high copolymerization reactivity.
For scintillator applications, it is important for nanocomposite scintillators to maintain high transparency at the final emission wavelength. When the transmittance of the 1 mm thick nanocomposite scintillator was measured by ultraviolet (UV) visible spectroscopy, the transmittance (68.4%) in the scintillator with 6 nm YbF3 NPs (63 wt%) was lower than that of the neat PVT monolith (more than 80%) at the final light-emitting region of the scintillator (416 nm). Cai et al. [70] also reported that a nanocomposite scintillator with a high loading of 9 nm Gd2O3 NPs (31 wt%) and an organic fluorescent dye uniformly dispersed in a millimeter-sized PVT matrix had a lower transmittance (69.6% at 550 nm) than the pure PVT monolith (76% at 550 nm). In addition, Zhao et al. [71] demonstrated that the transmittance gradually decreased with the loading of 5 nm HfO2 NPs and then sharply decreased at the highest loading of NC (50 wt%) in a toluene solution, as shown in Fig. 2D, when the concentration of fluorescent dyes was maintained. Considering that the band gap of both the matrix (polymer or solvent) and the NPs (YbF3, Gd2O3, or HfO2) is much larger (4–6 eV) than that of the fluorescent dyes (2–3 eV), the transmittance loss is caused by the following factors: (1) the absorption of photons by PVT and NPs in the UV region (<400 nm); (2) the refractive index mismatch between the NP and the matrix; and (3) the scattering effect by impurities and NPs. When the particle size is 10 times smaller than the incident wavelength, the Rayleigh scattering theory is used to explain the transmittance (T) loss of nanocomposites as follows in Equation (1) [72]:
where λ is the wavelength of the incident light, Vp is the volume fraction of the particles, t is the thickness of the nanocomposite, r is the radius of the particles, and nm and np are the refractive indices of the organic matrix and the inorganic NPs, respectively. The refractive index of inorganic NPs is approximately in the range of 1.49–1.59 for YbF3, 2.15–2.28 for Gd2O3, and 1.89–2.12 for HfO2, while that of the organic matrix is roughly 1.49–1.58 for PVT and toluene. The transmittance decreases as the difference in refractive index between the particle and the matrix increases. It has been shown that the apparent light scattering loss due to the refractive index mismatch can be mitigated if the particle size is 10 times smaller (below 40 nm) than the wavelength of visible light [73, 74]. Apart from the fact that the thicker sample has a low transmittance, the transmittance decreases with a higher particle loading (Vp) and a larger particle size (>100 nm) [75]. Due to the insufficient amount of (polymeric) matrix composites, higher loading of NPs would lead to aggregation of NPs (which represents an increase in domain size) [65]. Thus, good NP dispersion is essential to achieve high transmittance for nanocomposite scintillators.
Thermal curing is a common polymerization method in which free radicals are generated from a thermal initiator to obtain a transparent nanocomposite scintillator. Liu et al. [76] successfully fabricated a nanocomposite scintillator based on core-shell structure NCs using a polymerization process. As shown in Fig. 3A, after the partial BMEP modification, CdxZn1-xS/ZnS NCs were dispersed in VT monomer solution with a fluorescent dye, the DVB, and Luperox 231 (Sigma- Aldrich Chemical Co.) as a thermal polymerization initiator. Chen [77] demonstrated that the thermal curing method can easily scale up the volume of nanocomposite scintillators from several hundred micrometers to a few millimeters, with different shapes using glass molds. The glass molds were simply cleaned in advance with a wet chemical cleaning process, such as immersion in hydrogen peroxide solution for 30 minutes. The cleaning process is described in detail in [77]. Cai et al. [70] demonstrated the fabrication of a transparent nanocomposite with a high loading of Gd2O3 NPs (31 wt%) with a diameter of about 9 nm uniformly dispersed in millimeter-sized samples, as shown in Fig. 3B. For example, a transparent monomer (VT) solution was prepared by dispersing Gd2O3 NPs with BMEP ligand exchange, thermal initiators (Luperox 231), and organic fluorescent dyes of 4,7-bis{2′-9′,9′-bis[(2″-ethylhexyl)fluorenyl]}-2,1,3-benzothiad (FBtF) in the pretreated glass molds. It was cured in the vial at 95 °C for 48 hours. The scintillator (⌀14 mm×3 mmt) removed from the mold was transparent after polishing, as shown in the inset of Fig. 3B.
Meanwhile, due to the poor thermal stability of perovskite NCs, Xin et al. [78] proposed the fabrication process of a nanocomposite scintillator using UV curing. Fig. 3C illustrates UV- and thermally polymerized perovskite (CsPbBr3) nanocomposites using photo- and thermal initiators with the crosslinker (DVB) in a monomer solution (butyl methacrylate). The PLQY was significantly increased by UV curing (62.2%) than by thermal polymerization (9.2%), and time-resolved photoluminescence (PL) decay curves in Fig. 3D show that the thermally polymerized sample has the shortest decay time due to decomposition of NCs and increased surface defects.
2) Other manufacturing methods: liquid scintillator, porous matrix
Zhao [79] reported the limitation of fabrication using UV curing for nanocomposite scintillators with a fluorescent dye, in which CsPbBr3 NCs were mixed with VT monomer, crosslinker (DVB), photo-initiator (Irgacure 810; Ciba Specialty Chemicals), and many different fluorescent dyes, such as 2-tert-butyl-4-(dicyanomethylene)-6-[2-(1,1,7,7-tetramethyljulolidin-9-yl)vinyl]-4H-pyran (DCJTB), pyrromethene 580 (PM-580), and pyrromethene 597 (PM-597). The bulk solutions remained liquid and became almost colorless after 2 days of UV irradiation, as shown in Fig. 4A. The bleaching effect was caused by the fact that organic fluorescent dyes were decomposed by radicals generated from photo-initiators under UV irradiation. Furthermore, the UV curing caused insufficient polymerization in the inner part of the bulk scintillator due to the limited transmission depth of UV.
As an alternative, Yu et al. [80] designed a liquid scintillator containing perovskite NCs, which could be easily obtained without additional ligand exchange, as there is no need for polymerization between NCs and polymers. Another advantage of the liquid scintillator is that the detector volume can be easily scaled up by increasing the volume of the quartz cuvette. Furthermore, a HfO2 NP-loaded liquid scintillator was designed by mixing the toluene as the primary solvent and naphthalene as the cosolvent in a cylindrical cuvette, as shown in Fig. 4B [71]. The cosolvent, which has a smaller band gap than the primary solvent, was added to enhance the exciton energy transfer. On the other hand, Letant and Wang [81] attempted a nano-porous glass plate to impregnate CdSe/ZnS NCs within an array of interconnected pores (~⌀10 nm). The dry “thirsty” porous host matrix absorbed the NCs in the stock solution of NCs for 2 days with continuous stirring to promote homogeneous filling of the NCs into the pores. After evaporation of the solvent, the air-dried sample with an NC density of 10 mg/cm3 is shown in Fig. 4C.
2. Luminescent Nanocrystals
Luminescent NCs have been in the spotlight as a light emitting source in scintillators since the core/shell structure NCs and perovskite have been in vogue recently. The luminescence mechanism of NCs has traditionally been based on band-edge and trap-state emission. The shell structure not only improves the photochemical/physical stability but also enhances the radiative recombination with the type I structure, which spatially confines electrons and holes in the core due to the band alignment of the core and shell materials [22, 82–84].
Fig. 5A illustrates how energy deposited from gamma rays is converted into visible photons via three main FRET pathways in the luminescent NC-based nanocomposite scintillator:
(1) Electromagnetic radiation, such as gamma rays, undergoes mostly multiple Compton scattering with the organic matrix or NCs, or scarcely photoelectric effect with NCs, and deposits its full energy to the nanocomposite.
(2) The subsequent emitted hot electrons due to the gamma-ray interaction excite π electrons of adjacent molecules (organic matrix, fluorescent dye) by Coulomb scattering.
(3) The exciton energy of about a few eV is transferred from the organic matrix to the NCs by FRET or fluorescence.
(4) Because the majority of luminescent NCs suffer from their small Stokes shift during the band-edge emission (i.e., the small difference between the fluorescence emission peak and the first exciton absorption peak), a third component, a fluorescent dye, is commonly required to mitigate the strong self-absorption of the NCs. When the NCs are de-excited to a ground state by radiative recombination, fluorescent dyes absorb the light and re-emit visible photons with a larger Stokes shift, followed by photon detection using a photosensor (e.g., photomultiplier tube [PMT]).
Maximizing the efficiency of exciton energy transfer is very important in designing nanocomposite scintillators (because the deposited gamma-ray energy should be proportional to the number of visible photons produced for gamma spectroscopy). Liu and Bazan [85] noted that FRET requires a resonance condition where the highest occupied molecular orbital and the lowest unoccupied molecular orbital energy levels of the acceptor are within the energy levels of the donor. If the exciton energy in the donor (e.g., polymer molecule) is to be transferred to the acceptor (e.g., fluorescent dye), a transmitter molecule (e.g., NCs) facilitates the FRET process between the donor and acceptor molecules. In addition, the distance between the matrix molecule and NCs should be within 1–10 nm since FRET occurs between an excited molecule and a ground-state molecule via non-radiative dipole-dipole coupling. Considering that the volume of the nanocomposite scintillator mostly consists of organic polymers, the concentration of the NCs should be high enough to reduce the average distance between the donor and the transmitter. Note that if a donor (or transmitter) is highly luminescent, the energy can also be transferred to the transmitter (or acceptor) by radiative energy transfer.
Despite the bright light emission property of the NCs, a fluorescent dye is required to prevent strong self-absorption in luminescent NCs due to its small Stokes shift. For example, Zhao [79] reported that a nanocomposite scintillator containing 20 wt% of CsPbBr3 NCs without organic dyes, where the scintillator’s overall effective Z is calculated to be 31, showed only a Compton edge in the 137Cs gamma spectrum implicating that the gamma-ray interaction was mostly to be the Compton scattering with organic molecules in the composite, rather than the photoelectric absorption by the NCs. Another speculation on the result is the poor light harvesting from the nanocomposite due to the strong self-absorption of NCs (small Stokes shift), which is also well-expected. Liu et al. [76] also reported that the scintillation decay time of NCs could not be measured by time-correlated single-photon counting for CdxZn1-xS/ZnS core/shell in PVT (without dyes) because the signal was unreadable.
Liu et al. [76] fabricated transparent and high-loading nanocomposite scintillators, as shown in Fig. 5B, and observed that the light yield gradually increased as the concentration of CdxZn1-xS/ZnS core/shell NCs increased from 0 wt% to 60 wt% with 2 wt% of FBtF dye in the PVT monolith. For example, Fig. 5C shows that the Compton edge shifts to a higher channel number due to the closer distance between NC and PVT molecules (increased FRET efficiency) as the NC concentration increases. The light yield of the scintillator is, in principle, anticipated to be proportional to the channel number at Compton edge. When measured with the 137Cs gamma source, the nanocomposite scintillator with 60 wt% of CdxZn1-xS/ZnS core/shell NCs exhibited an energy resolution of 9.8% at the X-ray (Cd Kα of 16 keV) escape peak (646 keV). Note that the X-ray escape peak is more intense than the photopeak from which it is derived because the low Zeff and density of the scintillator allow the 16 keV X-ray to escape the solid with high efficiency. In particular, X-ray escape peaks can stand out more as the active volume of the detector becomes smaller than the typical penetration depth of the proceeding particles, which makes it one of primary concerns in the design of thin-film-based detectors, such as organic-material-based or nanomaterial-based detectors [86, 87].
The energy resolution is relatively inferior to that of a commercial single-crystal scintillator, such as NaI(Tl), and was ascribed to the wavelength mismatch between the NC emission (emission peak at about 530 nm) and the photocathode’s (or PMT’s) sensitivity (maximum response at 420 nm). In addition, the peak-to-Compton ratio was relatively low because the Zeff of the scintillator itself was not high enough. In case of typical cadmium-based NCs for optoelectronic applications, the highest Z-number of the NC elements is 48 (cadmium, which is composed of the core), and the shell, which occupies a larger portion of the NC volume, was composed of ZnS.
Note that since most of the nanocomposite volume consists of the polymer matrix such as PVT, gamma rays can mostly undergo Compton scattering with the polymer matrix and escape from the scintillator. Possible improvements could then be achieved by high-density packing of NCs and less volume of the organic matrix, thus preventing events (full energy absorption at NCs) from being shrouded in the undesirable Compton scattering events at PVT molecules. Compton rejection by anticoincidence technique [1] can also be an alternative, but the fundamental material property should be improved a priori.
Yu et al. [80] attempted a liquid-type nanocomposite scintillator containing perovskite (CsPbBr3) NCs to avoid the polymerization process. The CsPbBr3 NCs were dispersed in a solvent (1,2,4-trimethylbenzene [TMB]) with the fluorescent dye (PM-580). As the NC loading increased from 0 wt% to 40 wt%, at a fixed dye concentration of 0.75 wt%, neither the full-energy peak nor the escape peak of 137Cs gamma rays was visible by liquid scintillators containing 0–30 wt% of NC. At 40 wt% NC loading, the X-ray escape peak at 601 keV shown in Fig. 6A exhibited an abnormally poor energy resolution of 27%, even considering the final emission wavelength (576 nm) being not matched to the PMT (maximum sensitivity at 420 nm). The X-ray escape peak turned out to be a combination of Pb Kα and Cs Kα escape peaks with an average centroid value of two escape peaks at 587 keV and 631 keV. Note that the channel number was calibrated to energy using the fitted Compton edge (478 keV), assuming a linear relationship exists between the channel number and energy. The Compton edge was determined as the inflection point in the slope. This escape peak was attributed to the small volume of the scintillator (⌀22 mm×2 mmt). In addition, the loading of 50 wt% CsPbBr3 NC was limited due to the difficulty in dissolving the NC and the strong gelation behavior.
Sahi et al. [69] reported results from a nanocomposite scintillator containing 30 wt% of the blue-emitting La0.6Ce0.4F3 NCs. The 5d→4f transition of the Ce3+ ion in the La0.6Ce0.4F3 NCs allows the luminescence with the PL peak at 318 nm, and the use of organic fluorescent dyes, such as 2,5-diphenyloxazole (PPO) and 1,4-bis(5-phenyloxazol-2-yl) benzene (POPOP), shifts the peak wavelength to 410 nm to have a wavelength matching with a PMT (maximum sensitivity at 420 nm). The blue-emitting nanocomposite scintillator (⌀20 mm×5 mmt) shows a prominent photopeak by relatively low energy gamma rays (122 keV) by measuring 57Co radionuclide, as shown in Fig. 6B. However, due to its low density (1.2 g/cm3), the 137Cs gamma spectrum in Fig. 6C only shows an X-ray escape peak at about 622 keV, allowing La and Ce Kα X-rays (~34 keV) to escape, followed by photoelectric effects. Nevertheless, the scintillator was demonstrated to be capable of gamma spectroscopy and to outperform a commercial EJ-200 plastic scintillator (Eljen Technology). Meanwhile, Letant and Wang [81] presented the 59 keV gamma photopeak of the radionuclide 241Am obtained by the nano-porous glass impregnated with CdSe/ZnS NCs, as shown in Fig. 6D. Although the measurement time was in the order of days, a prominent peak was observed with an energy resolution of 15%.
3. Non-Luminescent (or Weakly Luminescent) Nanoparticles
The mechanism of gamma-ray detection by a non-luminescent (or weakly luminescent) NP is similar to those by the luminescent NCs above, while nanocomposite scintillators have NPs with a wider bandgap than an organic matrix. Cai et al. [70] proposed a scintillator based on Gd2O3 NPs, which can essentially eliminate the problem of small Stokes shift of NCs. Fig. 7A shows the possible FRET paths within the energy band alignment of the components during the scintillation process in the Gd2O3 NPs/FBtF dye/PVT system. Because the Gd2O3 NPs are non-luminescent (having very weak luminescence due to the surface defects) the exciton energy should be transferred down from Gd2O3 NPs to the PVT and, then, from the PVT to the fluorescent dye via FRET. The radiative recombination of the exciton at the dyes would finally result in the emission of visible photons with the emission peak at 520 nm. As given in Fig. 7B, the nanocomposite scintillator containing 31 wt% of Gd2O3NPs shows an escape peak at around 600 keV (Here, authors claimed that it is due to the escape of Bi K X-rays; however, it obviously seems to be an incorrect analysis, as this work did not involve any usage of Bi element.) with an energy resolution of 11.4% when measuring 137Cs gamma rays. The escape peak was attributed to the low density (1.2 g/cm3) of the nanocomposite scintillator. Moreover, increasing the amount of NP leads to decreased light yield (27,000 and 22,000 photos/MeV for 31 wt% and 40 wt% NPs, respectively) due to the mismatch in refractive index between Gd2O3 (1.80) and PVT (1.55).
Chen et al. [88] obtained a highly transparent YbF3 NPs-PVT nanocomposite scintillator (⌀10 mm×2 mmt) by matching the refractive index between the NPs (1.52) and the polymer matrix (1.55). Fluorescent dyes, such as 2-(4-biphenylyl)-5-phenyl-1,3,4-oxadiazole (PBD) and POPOP, were incorporated into the scintillator to have a wavelength match with the PMT (maximum sensitivity at 420 nm). In Fig. 7C, when 137Cs pulse height spectra were acquired by varying the content of YbF3 NPs, the Compton edge channel was shifted downward with increasing NP content, representing a linear decrease in light yield. Note that the light yield decreased 2-fold when the NP content was three times higher (7,380 photons/MeV and 3,750 photons/MeV for 15.8 wt% and 45.8 wt% NPs, respectively). In contrast, a small decrease in light yield was observed in Fig. 7D when the scintillator’s thickness was increased by three times (7,380 photons/MeV and 6,380 photons/MeV for 2 mm and 6 mm thick samples with 15.8 wt% NPs, respectively). The transmission loss due to the longer optical path length in the thicker scintillator is likely to cause light yield degradation.
Liu et al. [89] reported fabrication of a nanocomposite scintillator containing HfO2 NPs in a PVT matrix with fluorescent dyes (PBD and POPOP for wavelength matching to the PMT). Because the Hf atom has a higher Z-value (72) than Gd (64) or Yb (70), an enhanced photoelectric effect (approximately proportional to Z4–Z5) is expected [1]. Still, it shows a relatively low peak-to-Compton ratio, and the presence of an X-ray escape peak obscured the photopeak. Zhao et al. [71] recently demonstrated the efficient transfer of exciton energy in a toluene-based liquid scintillator containing high concentrations of HfO2 NPs. A cosolvent (naphthalene or 9,9-dimethylfluoroene) with a smaller bandgap than a primary solvent (toluene) was added to enhance the exciton energy transfer, and the two fluorescent dyes (PBD and POPOP) were used. The cosolvent plays an important role as a transmitter between the donor (primary solvent) and the acceptor (primary fluorescent dye) in the FRET process. In Fig. 8A, the replacement of 25 wt% toluene by naphthalene or 9,9-dimethylfluorene shows a substantial increase in the Compton edge channel (when the nanocomposite contains 20 wt% of NPs).
It was also suggested that the liquid scintillator can be readily scaled up by using a larger quartz cell, as given in Fig. 8B, which essentially avoids the creation of two separate domains (polymer and agglomerated NPs) occurred by the strong van der Waals force of attraction between polymer chains and the strong interparticle interaction of NPs. Increasing the thickness of the liquid scintillator from 2 mm to 20 mm resulted in significant suppression of the X-ray escape peak, as shown in Fig. 8C, where the toluene and naphthalene were used with a volume ratio of 5.2:1, including 40 wt% of NPs, 1 wt% of PBD, and 0.1 wt% of POPOP. An enhanced discernibility of photopeak can be achieved by complicated fitting process, however, the energy resolution of the photopeak was still rather poor of 14.8% at 662 keV. This was attributed to the loss of transmittance (44% at 430 nm) with a light path of 20 mm, resulting in poor energy resolution when compared to the transmittance in the 2 mm thick scintillator (90%).
McKigney et al. [90] reported the incorporation of cerium-doped NPs into a polymer matrix. Cerium is one of the popular activators used for inorganic scintillators, acting as a luminescence center by creating energy states within the forbidden band. It has been reported that Y2SiO5:Ce NPs shows a higher intensity of radioluminescence than bulk Y2SiO5:Ce particles. This enhanced intensity is likely due to the nanoscale property of the particles, as opposed to particles larger than 100 nm. Such large particles result in strong optical photon scattering in the visible region, causing turbidity of the material [75]. Photopeaks of 241Am and 57Co radionuclides could be observed from the LaF3:Ce NPs nanocomposite scintillator as shown in Fig. 8D [90].
Nanocomposite Direct Detectors
In principle, detection of incident quanta through the direct conversion of deposited energy into electric signal could be preferred and more promising in terms of the information carrier conversion efficiency. The losses of information carriers associated with converting the charge into scintillation light and back again to the electric charge at the readout media would be less favorable for the energy resolution in spectroscopy. In addition, efficient charge- and multi-exciton generation phenomena in the NC assembly shown by the cathodoluminescence (CL) experiment buttress the promises of developing an efficient and high-resolution gamma-ray detector based on the direct-conversion approach. As one wants to observe the full-energy absorption of gamma rays by a nanocomposite direct detector, the following aspects of the nanocomposite assembly are crucial [14, 15].
(1) Improvement of the electronic coupling; Radiation-induced charge carriers should be collected by electrodes and may face millions of NCs and interfaces. The transport of charge carriers over a shorter distance can be promoted by creating a closed-packed NC and using short-chain ligands. Furthermore, inducing oriented attachment can enhance the electronic coupling of NCs, such as higher charge carrier mobility, in which the NCs are atomically bonded or self-assembled with high crystallographic phase purity and a strongly preferred growth direction [91, 92].
(2) Preventing crack formations; The drying of the colloidal solution of NCs is accompanied by the formation of cracks. The internal cracks act as a potential barrier to charge carrier transport and provide trapping sites when extracting charge carriers. Blending polymer with NC or growing macroscopic clusters from self-assembled NC colloidal solids can be an alternative measure.
(3) Uniformity and stability of the nanocomposite solids; When the NC aggregation occurs, a locally formed cluster has a higher capacitance, decreasing the induced current and a non-uniform signal-to-noise ratio over the local domain. This poor dispersion should also be avoided to prevent the space charge effect (which distorts the internal electric field over time).
(4) Low level of leakage current; The level of leakage current depends on the NC characteristics (NC size, ligand chain length, dispersion state) and the applied voltage across the detector. A sufficiently high bias voltage is required to generate an electric field large enough to achieve an efficient charge separation from the NC assembly. Subsequently, high resistivity (typically greater than 108 Ω cm) is essential to minimize the leakage current under the bias. Note that there is a trade-off between the charge collection efficiency and the material resistivity.
1. Assembly of Nanocomposite Direct Detectors
There are relatively few studies of nanocomposite direct detectors that have achieved gamma spectroscopy compared to nanocomposite scintillators. The development of the direct detector has been hampered by cracks that form during the drying of the colloidal NCs dispersed in a solvent. Kim [14] observed that the solids of the PbSe NC assemblies suffered from cracks at the macroscopic scale, as shown in Fig. 9A taken by a scanning electron microscope (SEM). When a droplet of the NC solution was dropped on a metal film and dried a hundred times each, the crack formation was attributed to the vigorous evaporation of the highly volatile solvent (e.g., chloroform), followed by the propagation of the stacking fault through the layers [93]. Fig. 9B illustrates that the internal cracks act as a potential barrier to charge carrier transport and provide trapping sites when extracting charge carriers created by radiation from the nanocomposite solids. Furthermore, the drop-casting method resulted in a rough surface morphology and a non-uniformly dense nanocomposite, which would degrade the electrical contact between the interfaces of the nanocomposite layer and the metal contact (as metal atoms can diffuse into the device through the voids), resulting in increased leakage current [94–96].
1) Polymer blending
Kim [14] and Hammig [16] reported the fabrication of a mixed NC/polymer composite assembly, where hybrid composite structure approaches for NC assemblies have already been extensively explored in LED and solar cell applications [97–99]. In this approach, a conductive (semiconducting) polymer, such as poly[(p-phenylenevinylene)-alt-(2-methoxy-5-(2-ethylhexyloxy)-p-phenylenevinylene)], referred to as para-MEH-PPV, was used. The para-MEH-PPV is known to be a hole-transporting agent (due to the higher mobility of holes than electrons) and has an ionization potential of ~5.1 eV, similar to the bulk PbSe (~4.75 eV). The conductive polymer plays an important role not only in protecting the NCs from moisture and oxygen but also in transporting radiation-induced charge carriers between the NCs and the polymer. Furthermore, the spin-coating method was chosen to achieve uniform deposition of NCs with smoother surface morphology onto the substrate and faster drying than the drop-casting method [31]. The typical spin-coating process, however, usually results in a very low material retention efficiency (little hold-up on the substrate) and, thus, only a thin-film layer with tens of nanometer thickness. In this regard, it was desired to develop a much thicker active region for gamma-ray detection.
Therefore, the idea was to create a deep groove (a few tens of microns) on a substrate that (1) holds the NC solution until it dries without leaking the solution; (2) provides easy access to the top and bottom electrode metals; and (3) separates these electrodes by insulating walls. The fabrication of the NC assembly in the deep groove of a silicon wafer is illustrated in Fig. 9C [100, 101]. The growth of tens of microns of oxide on a silicon wafer is achieved by plasma-enhanced chemical vapor deposition, where the plasma decomposes silicon source gases, such as SiH4, into silicon radicals and allows them to react with oxygen radicals derived from oxygen source gases, such as N2O. The photolithography technique (e.g., photoresist application, photomask alignment, exposure, development, and stripping) was used to create a pattern. The patterned oxide layer was etched in buffered HF solution for 20–30 minutes to obtain sunken area. A gold metal film was then deposited using a thermal evaporator, and the existing photoresist on the silicon dioxide was stripped using a photoresist etchant. The blended NC/polymer composite was then grown on the prepared substrate, and a second metal film was deposited as the top electrode [100].
The fabrication of a thick NC assembly with high uniformity is certainly necessary for ionizing radiation detectors as opposed to other optoelectronic devices, such as solar cells, LEDs, and photodetectors, which are mostly thin film-based (a few hundred nanometers thick). Fig. 9D illustrates that the thickness of the NC assembly can be controlled by repeated procedures of spin coating the NC dispersion on the prepared silicon wafer-based substrate, where the spin-coater was used to achieve uniform coverage compared to drop-casting. Depending on the rotation speed and the concentration of NC in the monolayer drop, the number of castings determines the thickness, which ranges from tens of nanometers to tens of micrometers, the depth of the etched area on the substrate.
As the thickness of the NC assembly had to be limited by the thickness of the silicon dioxide wall, Kim [14] designed a deep wet etching process to create a trench with an etch depth of a few hundred microns on a glass wafer, as illustrated in Fig. 10A [102–104]. A glass wafer was first cleaned with a PIRANHA solution (a mixture of sulfuric acid and hydrogen peroxide). A chromium (Cr) metal film (50 nm thick) was then deposited on the glass wafer, followed by an Au film (500 nm thick) deposition. The Cr film was used as an adhesion layer between the glass wafer and the Au film. The patterned photoresist layer was created by photolithography. The Cr/Au film was then etched with Cr and Au etchants, and another Cr/Au layer was developed on the backside of the glass wafer to protect the backside from the subsequent deep wet etching. The exposed area was etched in a buffered oxide etchant for 1.5– 2 hours to create a trench several hundred microns deep. After removing the Cr/Au layer, a second photolithography was applied to place the metal electrode on the bottom surface of the trench, followed by the deposition of the NC assembly and the top metal electrode. Fig. 10B and 10C show the cross-section of the prepared glass-based substrate examined by the SEM and the final fabricated device, respectively. Plastic (polytetrafluoroethylene) substrates can also be used to realize nanocomposite layers with a thickness of a few millimeters and a diameter of 15 mm, as shown in Fig. 10D [14].
2) Other fabrication methods: self-assembled cluster, NC pellet
Davis and Hammig [105] designed a millimeter-sized macroscopic cluster of self-assembled tris(diethylamino)phosphine (TDP)-PbSe NC colloidal solids presenting less crack formation. This method, called oriented attachment, facilitates the nucleation and growth of colloidal solids in stock solution without the use of destabilizing agents. When the density of the NCs is high enough, the freedom of movement of the NCs is restricted, and the inter-crystallite attachment between NCs is aggressively promoted. Fig. 11A is indicative of the macroscopic loosely bound clusters of PbSe NC, resulting in a 1.1 mm×1.1 mm×0.7 mm PbSe colloidal solid after solvent drying, as given in Fig. 11B. The obtained solid has a density of 3.8 g/cm3, which is lower than that of single-crystal PbSe (8.8 g/cm3).
McCrea et al. [106] presented a simple and scalable method to produce a pressed pellet that produces a nanocomposite solid with a stereotypical shape, up to several hundred microns thick, for PbSe NCs with passivation by phosphorus- oxygen (P–O) moieties, which will be further discussed in section ‘Surface modification for long-term stability.’ A PbSe NC powder was obtained by precipitating PbSe NCs from a stock solution using a poor solvent or taken after a ligand exchange process. The powder was completely dried in a vacuum desiccator to remove any solvent present and loaded into a cylindrical die with a diameter of 4 mm. The pure PbSe NCs were then manually pressed using a pellet press (manufactured by International Crystal Laboratories, S/N 082). The pellet has a thickness of 852 μm and a density of 3.32 g/cm3.
Minimizing the leakage current of a direct detector under bias is critical because the radiation-induced current is typically in the sub-pA to nA range. Thus, typical devices have high resistivity (≥108 Ω cm) since the leakage current is inversely proportional to the resistivity of the device if the device is to be used at room temperature. McCrea et al. [106] also observed the change in the resistivity of the pellet as the NC diameter and ligand type were changed. The resistivity was measured from I–V data based on Pouillet’s law, in which its resistance was obtained by linear regression analysis on the I–V curve ranging from –10 V to 10 V, and its thickness and interfacial area were known. A schematic of the I–V curve measurement setup and a representative I–V curve acquired by a PbSe NC pellet are presented in Fig. 11C and 11D, respectively. Then, each device combination (NC diameter and ligand type) was tested by changing the capping ligand of OA and TDP with a short-chain ligand (e.g., NH4SCN) or a medium-length ligand (e.g., 1-octanethiol). As the NC size becomes smaller (4.8 nm in diameter) with NH4SCN ligands, the bandgap increases due to the stronger quantum confinement effect, resulting in an increase in the resistivity of the pellets to 106–107 Ω cm, which is two orders of magnitude greater than those of the large NC (8.1 nm in diameter). Longer capping ligands may facilitate higher resistivity owing to the insulating property of the organic ligands; however, Kuo et al. [19] revealed that such long-chain ligands would cause, in turn, the reduction packing density of the NC assembly leading to an inefficient charge transport behavior.
3) Surface modification for long-term stability
One would also desire improved stability in ambient air without using polymers in the NC assembly. Sykora et al. [107] reported that the Pb and Se atoms on the NC surface are highly susceptible to oxidation, where such oxides, including PbO, PbSeO3, or SeO2, can be formed. In this regard, Woo et al. [108] reported NC colloidal solids that demonstrate long-term stability through surface modification. When a Se precursor of trioctylphosphine (TOP) was replaced by TDP, it was observed that the existing OA ligands on the NC surface were released by the diethylamine, resulting in the formation of a diethylamine monolayer on the PbSe NC surface. This monolayer was removed due to the high temperature (~170 °C) during the PbSe NC synthesis process, leaving a phosphorus-based derivative of TDP on the surface and forming P–O moieties by minor surface oxidation or a union of OA remnants, as illustrated in Fig. 12A [105].
According to Fig. 12B, the presence of these unique P–O moieties was demonstrated by 31P{1H} nuclear magnetic resonance spectra with a large resonance in the TDP-PbSe spectrum. In addition, in Fig. 12C, the FTIR spectrometer reveals that the peak of the P–O stretching bond is present only in the TDP-coordinated NC spectrum compared to the TOP-coordinated one. When both TOP- and TDP-PbSe NC samples were exposed to the air, Davis and Hammig [105] also observed that the features of the absorption spectrum, such as the peak position and shape, remained stable over time (~585 days) for the TDP-PbSe but not for the TOP-PbSe (which showed blue shifts and a dramatic decrease in luminescence intensity after 4 days). As can be seen from Fig. 12D, the X-ray photoelectron spectroscopy confirmed that TDP-PbSe NC exposed to air for 117 days revealed similar results with the 5-day-old TOP-PbSe NC. The figure also revealed the oxidation of the 55 days old TOP-PbSe NC in terms of the valence state of Se atoms at the surface.
2. Nanocomposite Direct Detector as a High-Resolution Gamma-Ray Detector
The application of nanocomposite to the gamma-ray spectroscopy by the direct detector approach is largely based upon the radiation-induced charge carriers, which are desired to diffuse and drift efficiently between NCs until they are collected at metal contacts, where the drifting charges lead to the induced current. Kim [14] and Hammig [16] reported the experimental setup to extract the induced current from the blended para-MEH-PPV/PbSe NC assembly, as shown in Fig. 13A, where the two black lines on the assembly indicate the physical access to the bottom and top electrodes. Fig. 13B shows the energy spectrum obtained when gamma rays from a 133Ba source were irradiated onto the 22 μm thick nanocomposite assembly with a metallized area of 1 cm×1 cm. The energy spectrum shows the full-energy absorption peaks at 356 keV and 384 keV, including the Pb and Se X-ray escape peaks due to the insufficient thickness of the detector [86, 109].
Furthermore, the performance of the detector sample was compared with three typical commercial direct (i.e., semiconductor) detectors, such as silicon, CZT, and HPGe detectors, as given in Fig. 13C. Here, the nanocomposite assembly was fabricated with the blended structure. A narrower peak width (or better energy resolution) represents the ability to discriminate adjacent gamma-ray peaks, allowing accurate radionuclide identification when multiple peaks overlap. The assembly was fully depleted at a bias voltage of +100 V but biased at +320 V to produce a more uniform electric field across the detector, called the ‘over-depleted’ condition. All measurements were performed at room temperature except for the coaxial HPGe, which was cooled with liquid nitrogen and biased at +3,800 V. The CZT detector spectrum was obtained by a 20 mm×20 mm×15 mm CZT detector with 11× 11 pixelated anodes on one side and a large-area planar cathode on the other side to utilize single-polarity charge sensing (because the electron has faster mobility than the hole) and depth-dependent correction (to compensate for signal loss due to electron trapping as a function of the travel distance) methods. The close-up of the peak at 356 keV, as can be seen in Fig. 13D, shows that the blended para-MEH-PPV/PbSe NC assembly exhibits superior performance (full-width at half maximum [FWHM] of 1.5 keV) to the silicon and CZT detectors (FWHM of 3.4 keV), which is comparable to the HPGe (FWHM of 1.4 keV). A higher energy response to 662 keV gamma rays emitted from a 137Cs source was also observed as shown in Fig. 13E, corroborating that the nanocomposite direct detector can provide successful gamma spectroscopy with a remarkable energy resolution of 0.32% [14, 16].
Meanwhile, as shown in Fig. 13F, the shift of the full-energy absorption peak by time was observed from the blended para-MEH-PPV/PbSe NC assembly as measuring gamma rays from the 133Ba source. The X-ray escape peak was obscured due to the peak shift with time. This peak shift is likely attributed to the polarization effect caused by the formation of NC clusters, as the direct polymer mixing process is subject to aggregation of NCs due to the incompatibility between the polymer and NCs [64, 110]. Wolcott et al. [111] corroborated that decreasing the inter-NC distance leads to a stronger polarization effect because the spatial distribution of the electron or hole wave function of PbSe NCs is closer to the interface with shorter ligands, where the dielectric environment is changed, and electronic coupling between NCs occurs. Bhardwaj et al. [112] reported that a larger capacitance was observed with a larger size cluster (from 9 nm to 72 nm) of PbSe NCs, confirming that space charge polarization occurs strongly at low frequencies. Chen et al. [113] reported that aggregation of Si3N4 NPs in a polymer matrix can increase the density of charge traps, contributing to space charge polarization under electric field. This polarization effect would distort the internal electric field distribution and affect the energy resolution of the radiation detector or change the peak position with time [114]. Fig. 13G is an indicative of the different electronic responses, such as the noise level of the output signal, for each domain resulting from the uneven distribution of NCs due to cracks. If NC clusters are formed locally, the capacitance of a nanocomposite will increase locally due to the increased amount of PbSe NCs [112], which has a much higher dielectric constant (~250 for bulk PbSe and ~100 for 12 nm PbSe NC) [115] than the MEH-PPV polymer (2.5) [116]. This increased capacitance would reduce the induced current and cause non-uniform signal-to-noise ratio over the entire region. It is noteworthy that the uneven packing distribution via the physical mixing process results from the incompatibility between the polymer and the NCs, leading to the formation of an interfacial void [110]. When the voids are found, partial discharges or electrical treeing are induced, which is susceptible to a dielectric breakdown in the blended assembly [117].
Davis and Hammig [105] tested the crack-free macroscopic cluster of self-assembled TDP-PbSe NC colloidal solids with dimensions of 1.11 mm×1.11 mm×0.68 mm, and the experimental setup is illustrated in Fig. 14A. The bottom contact was realized by physically attaching the solids to an aluminum-based plate and an aluminum probe as the top contact. When the applied bias voltage was 300 V, the leakage current was measured to be 1.0 nA. The induced current was passed through the charge sensitive preamplifier into the pulse shaping amplifier, and the energy spectrum was acquired by a multichannel analyzer. The energy spectrum obtained from the TDP-PbSe NC cluster solid was then compared with commercial detectors such as the single-crystal CdTe and HPGe given in Fig. 14B. The close-up of the photopeak at 81 keV, as can be seen in Fig. 14C, shows that the colloidal solid has a narrower FWHM of 0.65 keV than the CdTe detector but worse than HPGe. Furthermore, the absence of X-ray escape peaks was demonstrated by comparing the energy spectrum obtained from a thinner sample, operated at 100 V with a leakage current of 1.1 nA. In Fig. 14D, the thinner sample showed the prominent X-ray escape peak at 69 keV (due to Se Kα, Pb Lα, and Pb Lβ X-ray escapes following the photoelectric effect), the backscatter peak at 62 keV (81 keV gamma rays backscattered from the aluminum-based plate and returned to the detector), and the 53 keV gamma-ray peak (which is not clear due to the increased noise at around 50 keV). The experimentally obtained energy spectrum, including the backscatter peak and the X-ray escape peaks, was corroborated by simulation results using a Monte Carlo N-particle transport code [86, 87]. This thinner sample had worse energy resolution (FWHM of 1.9 keV) because the increased detector capacitance (inversely proportional to the detector thickness) resulted in the attenuation of the induced current. A linear fit between the radiation energy and the amplitude of the pulse output, as shown in Fig. 14E, demonstrated energy linearity.
McCrea et al. [106] reported that the stereotypically shaped pressed pellet (using PbSe NC with the OA ligands and the P–O moieties) is also capable of gamma spectroscopy, demonstrating a full-energy absorption peak at 81 keV as measured by the 133Ba source, as can be seen in Fig. 14F. However, this approach resulted in degraded energy resolution (~3.8 keV FWHM at the 81 keV photopeak) compared to the self-assembled TDP-PbSe NC cluster solid. Still, the fabrication time for a millimeter-sized direct detector was drastically reduced, and energy linearity was demonstrated with the linear fit as shown in the inset of the figure.
Summary and Outlook
This review has comprehensively outlined the concepts of gamma-ray detection by nanocomposite materials, the current status of the fabrication of NC-based gamma-ray detectors, and the performance of gamma spectroscopy by nanocomposite scintillator and NC assembly of direct detectors, as summarized in Table 1. Many studies have reported challenges and different approaches for: (1) nanocomposite scintillators by considering band-alignment engineering for efficient exciton transfer, and those made of various inorganic luminescent NCs and non-luminescent NPs, and (2) direct detectors of nanocomposite assembly to achieve fewer cracks and a high packing density of NC with a thickness of a few tens to hundreds of microns depth, and self-assembled NC colloidal solids with a millimeter scale and long-term stability in an ambient condition, and simple and scalable fabrication for a stereotypically shaped assembly. They share a common interest in fabricating a relatively thick (a few millimeters to centimeters) medium considering the mean free path for a 662 keV photon interacting with a bulk semiconducting media (e.g., >1 cm for PbS or PbSe, >2 cm for CdTe or CZT single crystals) [14]. Besides, the reproducibility of NC and(or) NP is required because their optoelectrical properties—such as bandgap, emission spectrum, and resistivity—depend on fine details, including the size, shape, size distribution, and surface functionality [118].
Although each approach has its own set of advantages for gamma-ray detection, it is important to fully understand the underlying mechanisms along with their shortcomings. Since the NPs’ stability determines the scintillator’s performance, the direct linkage between the NPs and the polymer was designed to protect against moisture and oxygen. This approach then requires the partial ligand exchange due to the polarity mismatch between the capping ligand and the polymer matrix and sometimes requires a crosslinker such as the DVB to prevent the aggregation of the NPs. For this reason, Lee [119] and Liu et al. [76] reported that a slight loss of NPs is inevitable during the ligand exchange process, and some NPs would suffer from ligand detachment (dangling bonds are formed), which would result in the degradation of NCs.
The typical thermal polymerization approach was limited to heat-sensitive NPs, such as lead halide perovskite NCs, and UV polymerization would lead to insufficient polymerization in the inner part of the scintillator due to the limited UV transmission depth. Thus, the liquid nanocomposite scintillator discussed herein serves as a prime example. Fluorescent NCs also suffer from a small Stokes shift, resulting in strong self-absorption at high NC loading. Adding another fluorescent dye to counteract the small Stokes shift is not always helpful; it might result in a redshift of the emission wavelength that is no longer sensitive to a photosensor, leading to a lower energy resolution [120]. Therefore, the application of luminescent NC-based plastic scintillators using a silicon photomultiplier (SiPM) is under development [121].
Non-luminescent (or weakly luminescent) NPs have been proposed, which essentially eliminates the problem of the small Stokes shift of NCs. This approach still suffers from the 137Cs X-ray escape peak at about 600 keV, predominating over the full-energy absorption peak at 662 keV. Complication by X-ray escape peaks is not a trivial matter in spectroscopy, especially when the active volume of the detection media is comparable to or smaller than typical range of primary electrons such as thin-film detectors, imposing challenges of creating detector samples of certain size [86, 87]. Although the nanocomposite liquid scintillator can be easily scaled up from 2 mm to 20 mm in thickness, this approach was still not free from the occurrence of an X-ray escape peak and resulted in a relatively inferior energy resolution of 14.8% at 662 keV. In fact, the wider bandgap of non-luminescent NPs (Eg>5 eV) has limitations in charge-creation statistics than luminescent NCs (Eg≃2–3 eV), resulting in a low energy resolution.
The fundamental limitations regarding the scintillation yield are: (1) the requirement of sufficient thickness (~1 mm) and (2) the inclusion of many different molecules. NCs that exhibit near-unity PLQY cannot guarantee their performance at high concentration, providing that they are subject to a small Stokes shift. In addition, when other components, such as polymers, are used to make a 1 mm thick nanocomposite scintillator, many different molecules result in complicated paths for the gamma rays to deposit their full energy. When Compton scattering or photoelectric absorption occurs in the nanocomposite scintillator, the emitted hot electrons will excite various kinds of molecules—solvent, cosolvent, NPs, fluorescent dyes—each of which has a different ionization potential. The number of excitons produced would then vary up to the paths for a given deposited energy of gamma-ray.
On the other hand, the main challenges addressed for the gamma-ray spectroscopy by nanocomposite arrays via direct conversion approach were: (1) charge loss during the slowing-down of the charged-particle in the inactive matrix; (2) charge trapping during subsequent electron and hole transport; and (3) small detector volumes, which are mostly risen by the challenge of non-uniformity and instability of the nanocomposite and constituent NCs.
The direct detectors of the blended NC/polymer assembly discussed herein demonstrated an excellent energy resolution of 0.32% at 662 keV, which is better than a state-of-the-art detector such as the CZT (0.48%) and comparable to the HPGe detector (0.27%), which has one of the best energy resolutions among gamma detectors. The blended NC/polymer composite assembly tends to alleviate the crack formation, and the spin-coating method was chosen to achieve uniform deposition of NCs with smoother surface morphology on the substrate, which can facilitate good electrical contact between the interfaces of the nanocomposite layer and the metal contact. Nevertheless, the stability of NCs as well as the assembly itself should be improved to avoid the internal gain shift over time, and the entire NC region is encouraged to have uniform charge transport behavior and noise level. A recent study has shown that the optical and chemical stability in an ambient environment can be improved by modifying the ligand organization on the NC [105]. The macroscopic cluster of self-assembled TDP-PbSe NC colloidal solids also resulted in a millimeter-sized pure NC detector without cracks. This self-assembled cluster, however, was quite fragile (it broke apart during handling), and each produced cluster has a different dimension and shape. Therefore, a manually pressed NC pellet was designed and produced in a stereotypical shape with a millimeter-scale diameter and a thickness of 852 μm. The obtained NC pellet showed degraded energy resolution compared to the self-assembled TDP-PbSe NC cluster solid [106].
To the best of our knowledge, several research works on NP-based gamma-ray detectors have demonstrated the capability of gamma-ray spectroscopy in the last decade. In principle, nanocomposite-based detectors are well-anticipated to revolutionize the capabilities of radiation instruments by yielding a high-performance detector via a low-cost solution-based fabrication modality, if the main challenges addressed in this review could be overcome.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Print 
