Novel design strategies in radiation detection technology are recently moving towards thin, high-sensitivity and flexible solid-state devices especially characterized by low manufacturing costs and easy-processing. In this respect, the class of new and versatile semiconducting materials called perovskites is progressively experiencing an ever increasing scientific focus worldwide [1]. Perovskites include any material with the same type of crystal structure, ABX₃, with A, B cations and X anion. In particular, metal-halide perovskites, mainly cesium lead halide (CsPbX₃) and methylammonium lead halide (CH₃NH₃PbX₃) attracted recently huge attention due to their improved optical and electrical properties. Unique physical properties of halide perovskites range from high-absorption coefficients, long-range ambipolar charge transport, high exciton-binding energy, high dielectric constant, all very attractive optoelectronic characteristics [2]. Growth from low-cost processes allows to manufacture them both as layered thin films, deposited on almost any kind of substrates and as single crystals of tailored size. Due to their strong stopping power, defect-tolerance, large mobility-lifetime product, high sensitivity per unit volume, halide perovskites were recently tested in several areas where stringent photodetection and dosimetric requirements are needed, as e.g. for medical imaging [3] radiation detectors [4, 5] and dosimeter [6–8]. Moreover, by opportunely varying the stoichiometry of the anion, starting from CsPbI₃ towards CsPbBr₃ and finally CsPbCl₃, it is possible to finely tune the material bandgap from about 1 eV to 3 eV [9]. This feature provides access to a variety of different applications in photodetection and dosimetry. In particular, high-bandgap energy is an interesting feature for semiconductors in devices subject to harsh radiation environments, as the radiation induced lattice damage has no influence on the dark current of the detector [10].

This paper focuses on active devices used as real-time dosimeters for proton therapy. Their principle of operation is based on the continuous read-out, during irradiation, of the current, I, flowing between two electrodes supplied with a constant bias [11]. For a proper operation the current signal should be linearly dependent on the proton fluxes. To provide real-time beam monitoring in-vivo, the device should be operated in transmission during the dose delivery. In this condition, the thickness of the device should be minimized. To avoid a detrimental reduction of the sensitivity of the device, the active material should be characterized by a high sensitivity per unit volume. Recent results on inorganic halide perovskite detectors made of CsPbBr₃ and CsPbCl₃ [5, 7] evidenced sensitivity per unit volume for these materials of the same order or even higher than that of silicon, diamond and silicon carbide [11, 12], among the most semiconductors used for radiation detection purposes. With respect to CsPbBr₃, CsPbCl₃ seems to be characterized by a current response as a function of time during exposure to radiation less prone to instabilities due to defects or polarization effects [13–15]. Moreover, its high bandgap (3.0 eV) should provide less dark current making it an intrinsically radiation harder material, as in the case of SiC [16]. Moreover, there is an increasing interest in the scientific community for a flexible film dosimeter that can be attached to the irregular surfaces of a patient for in vivo dosimetry during radiotherapy [17]. In this respect, lead halide perovksite appear attractive, as they may be grown directly on any kind surfaces by low temperature magnetron sputtering techniques.

This paper shows first measurements carried out with CsPbCl₃ thin film devices deposited on flexible substrates carrying interdigitated electrodes (IDE) when exposed to proton beams with energy in the range 100–228 MeV and extraction currents in the range 1–10 nA, of interest for proton therapy applications.

Cesium lead chloride (CsPbCl₃) films have been grown by Radio Frequency (RF) magnetron sputtering on plastic substrates equipped with an array of Cu electrodes using a Korvus HEX (Korvus Technology Ltd.) sputter equipped with an RF source working at 13.56 MHz [18–20]. The precursor salts (CsCl and PbCl₂) are grinded in equal molar ratio in a mixer mill (Retsch model MM400) to produce the powder used to manufacture the sputtering target: a 5 cm diameter disk realized by pressing the perovskite powder by means of a pneumatic press (11.5 MPa working pressure) for 24 h at 150°C [21]. The deposition is carried out at room temperature with 20 W RF power and with a 20 sccm argon gas flow. Dynamic working pressure is 2 × 10⁻⁶ atm, deposition rates are in the range 5–7 × 10⁻² nm/s. The sample is nominally undoped even if unwanted p-doping is due to Cs vacancies [22]. The film thickness is monitored during growth by means of a quartz crystal microbalance. During deposition the sample holder rotates to ensure thickness homogeneity across the substrate.

A detailed electro-optical characterization of CsPbCl₃ thin films, deposited with the same technique on glass, has been reported elsewhere [13, 20]. Room temperature transmittance spectrum provided a value of the bandgap of 3.0 eV. Room temperature XRD spectra indicate the presence of the correct crystalline phases CsPbCl₃ (ICDD PDF-4+ 18-0366), orthorhombic and tetragonal crystalline phases [13]. SEM analyses reported in [20] evidences a film morphology characterized by a continuous network of several tens of nanometer size crystals with a random distribution of bigger crystals with typical micrometer/sub-micrometer size. I-V characteristics performed with samples of different thickness on glass substrates, from 100 nm to 1.5 μm indicate a good homogeneity of the semiconductor material independently of the deposition time [21]. Thermally stimulated currents after exposure with a 365 nm LED priming measured from room temperature up to 370 K showed no evidence of trapping effects due to intragap states on the electrical transport properties at room temperature of the films. As a consequence, measured photocurrents at room temperature appear to be stable in time with negligible priming effects.

With the purpose of prototyping first flexible thin film real-time detectors based on inorganic perovskite, we have deposited with the same technique a set of CsPbCl₃ films on PET substrates equipped with interdigitated electrodes (IDEs). Figure 1A shows the sample, named #7, tested in this study under proton beam, carrying a CsPbCl₃ film 1 μm-thick deposited on the whole interdigitated electrode array of a flexible plastic substrate (PW-IDEPD100 by DropSens). The two interdigitated electrodes have 100 µm bands/gaps pitches and 5 × 7 mm² total area and nominal 175 μm thickness. The prototype is especially designed to get high electric signals starting from rather low thickness of material on a decentralized assay in planar electrode configurations. In the whole, the device is equipped with 35 pairs of contacts, 5 mm-long each: active area is 5.0 × 3.5 mm². To perform electrical measurements, the two connection tracks of the IDEs are directly connected to coaxial cables via sliding contacts. A pristine IDE substrate, #14, without the CsPbCl₃ thin film deposition, has been also prepared in same conditions, to be used as a reference during proton beam tests. The electrode termination are connected to Keithley 6517A/B high resistance source/electrometers, used both to supply an external voltage and to read-out the current, typically in the range 100nA-0.1 pA. The internal circuit of the instrument allows to monitor the current in real-time with varying integration times in the range 3–100 ms depending on the specifically investigated current ranges.

The response of the two devices during irradiation has been tested at the proton therapy center in Trento, Italy, which is part of the Trentino Healthcare Agency (Azienda Provinciale per i Servizi Sanitari—APSS, Italy). A cyclotron (IBA, Proteus 235) serves two medical treatment rooms both equipped with rotating gantries. Our test has been carried out at the experimental area, whose activity is managed and supervised by Trento Institute for Fundamental Physics and Applications (TIFPA) of the National Institute for Nuclear Physics (INFN). The total energy range of the proton therapy facility is 70–228 MeV, typical treatments fluxes are in the range 10⁶–10⁹ p/s [23].

After a preliminary optical and electrical characterization of the devices, we have measured in real-time configuration the current response of our devices to proton beams with energy in the range 100–228 MeV and extraction currents in the range 1–10 nA. During tests under the proton beam, the prototypes, including its electrode terminations, has been sealed into an insulating envelope to avoid unwanted current contributions due to the ionized air close to electrodes. Our configuration still allows for sample flexibility. Picture in Figure 1B shows the sample under test, #7 within its insulating envelope, bending under a tiny mechanical stress. The devices have been encapsulated in a 10 × 10 cm² PolyMethylMethAcrylate (PMMA) slab 1 cm thick (see Figure 1C) within a cylindric hole of the same diameter of the sliding contacts connector (1.3 cm), ensuring sample flexibility. The PMMA phantom has been placed at the isocenter of the beam at the TIFPA experimental line (see Figure 1D). During the proton test beam the flexible sample has been always kept in planar configuration.

The proton flux per unit extraction current (1 nA) impinging on the detector area is plotted as a function of the proton beam energy in Figure 2, together with an exponential best-fit to data shown as a guide-line. The fraction of protons reaching the detector has been estimated by integrating the beam profile, essentially Gaussian, on the detector area: f = 1 2 π σ E 2 ∫ − a 2 a 2 ∫ − b 2 b 2 ⁡ exp − x 2 + y 2 2   σ E 2 d x   d y , (1) where a and b are the lateral dimensions of the detector and σ E is the spatial standard deviation (i.e. the width) of the beam Gaussian profile for that particular energy E [23]. The maximum positioning error of the sample with the precision alignment laser system can lead to an error of 5% in the case of narrowest beam.

This paper shows first lead halide CsPbCl₃ thin film deposition on flexible PET substrates carrying interdigitated electrodes (IDE) for sensor developments. The CsPbCl₃ thin films have been deposited by magnetron sputtering at room temperature by means of a one-step process on IDEs made with Pd on plastic substrates (CsPbCl₃/Pd IDE). First measurements of these prototype sensors in real-time configuration are reported under test beam with proton energy in the range 100–228 MeV and extraction currents in the range 1–10 nA, of interest for proton therapy applications. Experimental results evidence the good performances of the CsPbCl₃/Pd IDE in terms of real-time monitoring, linearity with current extraction and proton fluxes in almost two orders of magnitude range, 4x10⁷—2 × 10⁹ p/s. These promising results, when coupled with easyness of fabrication, low processing costs and high versatility of electrode configurations, all features characterizing our manufacturing process, put into evidence lead halide perovskite, CsPbCl₃, as a promising candidate for future radiation monitoring in protontherapy.