*Communication* **Structural and Chemical Peculiarities of Nitrogen-Doped Graphene Grown Using Direct Microwave Plasma-Enhanced Chemical Vapor Deposition**

**Šarunas Meškinis , Rimantas Gudaitis, Mindaugas Andruleviˇ ¯ cius and Algirdas Lazauskas \***

Institute of Materials Science, Kaunas University of Technology, K. Baršausko 59, LT-51423 Kaunas, Lithuania; sarunas.meskinis@ktu.lt (Š.M.); rimantas.gudaitis@ktu.lt (R.G.); mindaugas.andrulevicius@ktu.lt (M.A.) **\*** Correspondence: algirdas.lazauskas@ktu.edu; Tel.: +370-671-73375

**Abstract:** Chemical vapor deposition (CVD) is an attractive technique which allows graphene with simultaneous heteroatom doping to be synthesized. In most cases, graphene is grown on a catalyst, followed by the subsequent transfer process. The latter is responsible for the degradation of the carrier mobility and conductivity of graphene due to the presence of the absorbants and transfer-related defects. Here, we report the catalyst-less and transfer-less synthesis of graphene with simultaneous nitrogen doping in a single step at a reduced temperature (700 ◦C) via the use of direct microwave plasma-enhanced CVD. By varying nitrogen flow rate, we explored the resultant structural and chemical properties of nitrogen-doped graphene. Atomic force microscopy revealed a more distorted growth process of graphene structure with the introduction of nitrogen gas—the root mean square roughness increased from 0.49 ± 0.2 nm to 2.32 ± 0.2 nm. Raman spectroscopy indicated that nitrogendoped, multilayer graphene structures were produced using this method. X-ray photoelectron spectroscopy showed the incorporation of pure pyridinic N dopants into the graphene structure with a nitrogen concentration up to 2.08 at.%.

**Keywords:** microwave; plasma-enhanced; CVD; nitrogen-doped; graphene; catalyst-less; transfer-less; synthesis

#### **1. Introduction**

Graphene belongs to a class of two-dimensional (2D) materials, which are widely known for their unique structures and outstanding physical, chemical, and mechanical properties [1–7]. It was demonstrated in many studies and reviews that the properties of 2D materials can be drastically altered, enhanced, or tuned via molecular and atomic doping [8–11]. For instance, the carrier concentration and type of carrier can be easily changed through the substitution of dopant atoms on the sulfur site in titanium trisulfide without having any impact on the band extrema [12]. After doping carbon nanotubes with N or B atoms, they become n-type or p-type atoms, respectively [13]. Doped graphene offers unique properties such as ferromagnetism [14], superconductivity [15], etc. Specifically, graphene heteroatom doping methods can be categorized [16] into post-treatment approaches, e.g., wet chemical methods [17], thermal annealing, arc-discharge [18], plasma treatment [19], hydrothermal treatment [20], and gamma irradiation [21], and in situ approaches, e.g., chemical vapor deposition (CVD) [22], bottom-up synthesis [23], and ball milling [24]. Methods which fall into the latter category are more favorable, as graphene synthesis can be achieved with simultaneous heteroatom doping [16]. Depending on the choice of heteroatoms for doping, e.g., B [25], N [26], P [27], S [28], F [29], Cl [30], Br [31], I [32], etc., new or improved properties of graphene materials may arise and could be useful for a number of applications, including supercapacitors [33], fuel cells [34], lithium ion batteries [35], solar cells [36], and sensors [37].

**Citation:** Meškinis, Š.; Gudaitis, R.; Andruleviˇcius, M.; Lazauskas, A. Structural and Chemical Peculiarities of Nitrogen-Doped Graphene Grown Using Direct Microwave Plasma-Enhanced Chemical Vapor Deposition. *Coatings* **2022**, *12*, 572. https://doi.org/10.3390/ coatings12050572

Academic Editor: Luca Valentini

Received: 29 March 2022 Accepted: 20 April 2022 Published: 22 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The CVD technique can be considered as the most effective approach used for graphene doping that does not affect its crystalline nature [38]. Generally, graphene is grown via CVD on a catalyst substrate, e.g., Ni or Cu [39,40], and afterwards, a complicated transfer process is required for the wet chemical removal of the metallic catalyst and the transfer of the graphene onto the required surface based on its intended functionality and application [41]. Importantly, the transfer process is known to be responsible for the generation of defects and the degradation of the carrier mobility and conductivity of graphene, since chemical alterations (i.e., during the wet chemical removal of catalyst) and mechanical damage (i.e., during the transfer process) are almost unavoidable during this process [42]. Additionally, the graphene transfer process is known to introduce unwanted metallic impurities which alter the electrochemical properties of graphene [43]. The demand for monocrystalline Si(1 0 0) continues to rise, as it is a major substrate used in semiconductor device fabrication and optoelectronics. The catalyst-less and transfer-less synthesis of graphene on monocrystalline Si(1 0 0) is meaningful in this context.

Herein, we focused our efforts on demonstrating that the catalyst-less and transfer-less synthesis of graphene can be achieved with simultaneous nitrogen doping in a single step via the use of a direct microwave plasma-enhanced CVD. Importantly, nitrogen-doped graphene synthesis was achieved at a considerably low temperature of 700 ◦C. By varying nitrogen flow rate, we explored the resultant structural and chemical properties of nitrogendoped graphene through atomic force microscopy (AFM), Raman spectroscopy, and X-ray photoelectron spectroscopy (XPS).

#### **2. Materials and Methods**

#### *2.1. Microwave Plasma-Enhanced CVD of Nitrogen-Doped Graphene*

The direct transfer-less synthesis of nitrogen-doped graphene was performed by employing the microwave plasma-enhanced CVD system Cyrannus (Innovative Plasma Systems (Iplas) GmbH, Troisdorf, Germany). Monocrystalline Si(1 0 0) (UniversityWafer Inc., South Boston, MA, USA) was used as a substrate. Plasma cleaning (power 1.7 kW, operating pressure 22 mbar, temperature 700 ◦C, hydrogen flow rate 200 sccm, and process duration 10 min) of the substrate was performed until the heater reached the target temperature. A methane, hydrogen, and nitrogen gas mixture was used for the direct synthesis of nitrogen-doped graphene. Afterwards, methane and nitrogen gas were introduced into the chamber. The growth process was performed in a single step. A steel enclosure on the substrate was used for the elimination of the unwanted direct plasma effects. The technological parameters of the plasma (i.e., power 0.7 kW, operating pressure 20 mbar, temperature 700 ◦C, and process duration 60 min) and the flow rate of hydrogen (75 sccm) and methane (35 sccm) gases were kept constant, while the flow rate of nitrogen gas was varied in the range of 0–110 sccm. Samples were denoted depending on the flow rate of nitrogen used in the process: N0 (N2, 0 sccm), N35 (N2, 35 sccm), N75 (N2, 75 sccm), or N110 (N2, 110 sccm).

#### *2.2. Characterization*

A NanoWizardIII atomic force microscope (JPK Instruments, Bruker Nano GmbH, Berlin, Germany) was used to conduct experiments at room temperature, with data analyzed using a SurfaceXplorer and JPKSPM Data Processing software (Version spm-4.3.13, JPK Instruments, Bruker Nano GmbH). An ACTA probe (Applied NanoStructures, Inc., Mountain View, CA, USA, specification: cantilever shape—pyramidal; radius of curvature < 10.0 nm and cone angle—20◦; calibrated spring constant—54.2 N/m; reflex side coating—Al with thickness of 50 nm ± 5 nm) was used to acquire AFM topographical images in contact mode.

Raman spectroscopy was performed using an inVia Raman spectrometer (Renishaw, Wotton-under-Edge, UK) equipped with a semiconductor green laser (wavelength of 532 nm), 2400 lines/mm grating, confocal microscope (50× objective), and CCD camera. The 5% laser output power was used for spectra recording (10 × 10 s accumulation time) in order to avoid sample damage. Band deconvolution into separate components was performed with OriginPro 8.0 software (OriginLab, Northampton, MA, USA).

Chemical state changes were investigated by employing a Thermo Scientific ESCALAB 250Xi spectrometer with monochromatized Al Kα radiation (hν = 1486.6 eV): X-ray spot of 0.9 mm in diameter; base pressure better than 3 × <sup>10</sup>−<sup>9</sup> Torr; 40 eV pass energy in transmission mode; energy scale calibration according to Au 4f7/2, Ag 3d5/2 and Cu 2p3/2. The peak fitting procedure was performed using the original ESCALAB 250Xi Avantage software.

#### **3. Results and Discussion**

The AFM analysis of pristine graphene (Figure 1a) was conducted over a 1.1 × 1.1 <sup>μ</sup>m2 area for quantitative morphological evaluation. The pristine graphene surface had a random distribution of surface features (mean height, *Z*mean of 0.76 ± 0.2 nm) with varying angle orientation to each other, without a preferred direction. The root mean square roughness (*R*q) was determined to be 0.16 ± 0.2 nm. The skewness (*R*sk) was determined to be 0.19 ± 0.2, indicating that the surface peaks dominate over the valleys. The pristine graphene surface exhibited a leptokurtoic distribution of surface features with a kurtosis (*R*ku) value of 3.32 ± 0.2.

**Figure 1.** AFM surface topography of pristine graphene (**a**), N35 (**b**), N75 (**c**), and N110 (**d**).

Some significant changes in surface topography (Figure 1b–d) as well as morphology (Table 1) can be observed for nitrogen-doped graphene: with the increase in nitrogen flow rate, the grain-like surface feature size increases in dimensions, with deeper valleys formed in between the boundaries (Figure S1). In general, the height of the surface structures and the roughness increase with the nitrogen flow rate. This change can be explained by the collision between nitrogen and carbon atoms as well as inner atom substitutional placement in the sp<sup>2</sup> hybridized lattice that results in the formation of various nitrogen-induced defect complexes [44], which evidently leads to a more distorted growth process of the graphene structure. Furthermore, graphene synthesis is performed in hydrogen plasma, which is also known to be responsible for defect formation [45,46]. Figure S2 shows the AFM step profiles for the corresponding samples. The thickness of the synthesized graphene films was found to be in the range of 9–13 ± 1 nm with no clear dependence on the nitrogen gas flow rate.


**Table 1.** Surface morphological parameters.

Figure 2 shows Raman spectra of pristine graphene grown directly on monocrystalline Si(1 0 0) substrate, as well as nitrogen-doped graphene samples. Raman spectra are presented in two actual wavenumber ranges of 1100–1800 cm−<sup>1</sup> and 2450–3150 cm−1, respectively. Four deconvoluted bands were determined using Gaussian components consisting of D and G bands at 1347 cm−<sup>1</sup> and 1599 cm−<sup>1</sup> and 2D and G + D bands at 2687 cm−<sup>1</sup> and 2937 cm−<sup>1</sup> for pristine graphene. The high ID/IG ratio of 1.26 indicates highly defective graphene. The low I2D/IG (0.15) and the broad 2D band in the Raman spectrum indicates the multilayer structure of graphene. The nitrogen doping of graphene is confirmed by the increased ID/IG ratio of samples N35 and N75 (Figure 2b,c), as well as the blue shift of the G band position at 1596 and 1595 cm−<sup>1</sup> for N75 and N110 (Figure 2c,d), respectively. Additionally, a red shift of the 2D position is observed for N110. Previously, Raman studies of nitrogen-doped graphene were conducted in depth to explain the red and blue shift behavior of 2D and G peak positions, corresponding to n-type doping and compressive/tensile strain in graphene [38,44,47]. S. Zheng et al. investigated the metalcatalyst-free growth of graphene on insulating substrates via ammonia-assisted microwave plasma-enhanced CVD [48]. They also observed the blue shift of the G peak position for nitrogen-doped graphene samples.

XPS was employed for the investigation of chemical state changes in the graphene structure resulting from the nitrogen doping. Figure S3 shows high-resolution XPS spectra in C 1s region for pristine graphene, N35 and N110, respectively. The deconvolution of spectra for these samples showed similar results: the strongest component at 284.2 eV, a less intense component at 284.8 eV, and a low-intensity component at 285.7 eV. A highly asymmetric component at 284.2 eV was assigned to C–C (sp2) bonds [22,49]. The second component at 284.6 eV was assigned to C–C (sp3) bonds [22,50]. The low-intensity component at 285.7 eV was assigned to C–O–C bonds [51,52]. It was found that the sp2/sp3 ratio decreased with nitrogen gas flow rate (Figure S3). Figure 3a does not indicate nitrogen-related peaks in the high-resolution N 1s spectrum, confirming the pristine nature of graphene. Figure 3b,c show the high-resolution XPS N 1s spectra with deconvoluted components of N35 and N110. The deconvoluted components in the high-resolution XPS N 1s spectra were attributed to pyridinic N (398.9 eV and 398.6 eV) [53,54], sp<sup>2</sup> C–N bonds (398.0 eV and 397.8 eV) [55,56], and N–Si bonds (397.2 eV and 397.0 eV) [57,58]. The latter comes from the contribution of the substrate as nitrogen-doped graphene was grown directly on monocrystalline Si(1 0 0). It is also important to note that the sp2 C–N bonding configuration is similar to pyridine. On the basis of these results, it was determined that the direct microwave plasma-enhanced CVD process produced graphene doped with pure pyridinic N. It was previously reported that the pyridinic N dopant efficiently changes the structure of the graphene valence band, including increasing the density of π states near the Fermi level, as well as reducing work function [59]. Lower work function can dramatically enhance the emitting current in graphene-based electronic devices [60]. It was also demonstrated in [61] that pyridinic nitrogen configuration in graphene contributed to the high catalytic performance. D. Wei et al. investigated the low-temperature critical growth of nitrogen-doped graphene on dielectrics via plasma-enhanced CVD [62]. They also found that nitrogen atoms in graphene are mainly bonded in the pyridinic N form (i.e., a nitrogen atom with two carbon atom neighbors assembling a hexagonal ring). The results of the atomic concentration calculation (Table 2) showed an increase in nitrogen from 1.04 at.% to 2.08 at.% for N35 and N110, respectively, indicating an increased level of pyridinic N doping with an increase in nitrogen gas flow rate. A very similar nitrogen concentration (i.e., 2.0 at.%) was reported for nitrogen-doped graphene synthesized using the N2:CH4 ratio of 2:1 via the direct microwave plasma-enhanced CVD process [44]. A nitrogen concentration of 2.0 at.% was also reported for nitrogen-doped graphene synthesized using a camphor:melamine ratio of 1:3 via the atmospheric pressure CVD process [38].

**Figure 2.** Raman spectra of pristine graphene (**a**), N35 (**b**), N75 (**c**), and N110 (**d**) recorded at 532 nm excitation wavelength and presented in two actual wavenumber ranges.

**Figure 3.** Deconvoluted high-resolution XPS spectra in N 1s region of pristine graphene (**a**), N35 (**b**), and N110 (**c**).


**Table 2.** Atomic concentration calculation results.

#### **4. Conclusions**

We demonstrated the catalyst-less and transfer-less synthesis of nitrogen-doped graphene in a single step at a reduced temperature of 700 ◦C by using direct microwave plasma-enhanced CVD. The nitrogen flow rate was varied to explore the structural and chemical peculiarities of nitrogen-doped graphene. AFM analysis showed that the height of the graphene surface structures and the roughness increase with the nitrogen flow rate due to nitrogen-induced defect complexes, which evidently lead to a more distorted growth process of the planar graphene. These structural changes were also quantified via Raman spectroscopy. Furthermore, it was determined that nitrogen-doped multilayer graphene structures were produced using this method. XPS analysis showed that pure pyridinic N was incorporated into the graphene structure during the simultaneous doping process. The level of pyridinic N doping increased with the nitrogen gas flow rate. Nitrogen-doped graphene could have potential applications in optoelectronics.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/coatings12050572/s1. Figure S1: Characteristic height profiles for the corresponding lines drawn in AFM images of pristine graphene (a), N35 (b), N75 (c), and N110 (d), Figure S2: Characteristic step profiles for the corresponding lines drawn in AFM images of pristine graphene (a), N35 (b), N75 (c), and N110 (d), Figure S3: Deconvoluted high-resolution XPS spectra in C 1s region of N0, N35, and N110.

**Author Contributions:** Conceptualization, R.G., Š.M. and A.L.; investigation, R.G., A.L., Š.M. and M.A.; writing—original draft preparation, Š.M. and A.L.; writing—review and editing, A.L. and Š.M.; visualization, A.L.; project administration, Š.M.; funding acquisition, Š.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was (and is) funded by the European Social Fund under the No. 09.3.3-LMT-K-712-01 "Improvement of researchers' qualification by implementing world-class R&D projects" measure. Grant No. 09.3.3-LMT-K-712-01-0183.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** A special thanks go to A. Vasiliauskas, A. Guobiene, K. Šlapikas, V. Stankus, D. ˙ Peckus, E. Rajackaite, T. Tamuleviˇ ˙ cius, A. Jurkeviˇciut¯ e, Š. Jankauskas, and F. Kalyk for ˙ technical assistance.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Electrical Method for In Vivo Testing of Exhalation Sensors Based on Natural Clinoptilolite**

**Gianfranco Carotenuto 1,\* and Luigi Nicolais <sup>2</sup>**


**Abstract:** Natural substances with a complex chemical structure can be advantageously used for functional applications. Such functional materials can be found both in the mineral and biological worlds. Owing to the presence of ionic charge carriers (i.e., extra-framework cations) in their crystal lattice, whose mobility is strictly depending on parameters of the external environment (e.g., temperature, humidity, presence of small gaseous polar molecules, etc.), zeolites can be industrially exploited as a novel functional material class with great potentialities in sensors and electric/electronic field. For fast-responding chemical-sensing applications, ionic transport at the zeolite surface is much more useful than bulk-transport, since molecular transport in the channel network takes place by a very slow diffusion mechanism. The environmental dependence of electrical conductivity of common natural zeolites characterized by an aluminous nature (e.g., chabasite, clinoptilolite, etc.) can be conveniently exploited to fabricate impedimetric water-vapor sensors for apnea syndrome monitoring. The high mechanical, thermal, and chemical stability of geomorphic clinoptilolite (the most widely spread natural zeolite type) makes this type of zeolite the most adequate mineral substance to fabricate self-supporting impedimetric water-vapor sensors. In the development of devices for medical monitoring (e.g., apnea-syndrome monitors), it is very important to combine these inexpensive nature-made sensors with a low-weight simplified electronic circuitry that can be easily integrated in wearable items (e.g., garments, wristwatch, etc.). Very low power square-wave voltage sources (micro-Watt voltage sources) show significant voltage drops under only a minimal electric load, and this property of the ac generator can be advantageously exploited for detecting the small impedimetric change observed in clinoptilolite sensors during their exposition to water vapor coming from the human respiratory exhalation. Owing to the ionic conduction mechanism (single-charge carrier) characterizing the zeolite slab surface, the sensor biasing by an ac signal is strictly required. Cheap handheld multimeters frequently include a very low power square-wave (or sinusoidal) voltage source of different frequency (typically 50 Hz or 1 kHz) that is used as a signal injector (signal tracer) to test audio amplifiers (low-frequency amplifies), tone control (equalizer), radios, etc. Such multimeter outputs can be connected in parallel with a true-RMS (Root-Mean-Square) ac voltmeter to detect the response of the clinoptilolite-based impedimetric sensors as voltage drop. The frequency of exhalation during breathing can be measured, and the exhalation behavior can be visualized, too, by using the voltmeter readings. Many handheld multimeters also include a data-logging possibility, which is extremely useful to record the voltage reading over time, thus giving a time-resolved voltage measurement that contains all information concerning the breathing test. Based on the same principle (i.e., voltage drop under minimal resistive load) a devoted electronic circuitry can also be made.

**Keywords:** clinoptilolite; impedimetric sensor; surface conductivity; apnea syndrome monitoring; voltage drop

**Citation:** Carotenuto, G.; Nicolais, L. Electrical Method for In Vivo Testing of Exhalation Sensors Based on Natural Clinoptilolite. *Coatings* **2022**, *12*, 377. https://doi.org/10.3390/ coatings12030377

Academic Editor: María Dolores Fernández Ramos

Received: 21 February 2022 Accepted: 10 March 2022 Published: 13 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

Materials with a unique combination of properties (multifunctional materials) can be found among natural substances characterized by significant structural complexity [1]. When a possible application for such nature-made materials has been identified, the development of a device with performance comparable to that of systems based on artificial substances can follow [2]. There are also examples in the literature in which similar technological results cannot be reached with comparable performance by using manmade materials [3,4]. As a consequence, the capability to identify the applicative potentiality hidden in the most common natural substances represents the key factor to come in the technological exploitation of nature-made materials, thus promoting sustainability. Among the different available natural substances, the most powerful potentialities have been found in complex types of solids, since their artificial replication is quite difficult [5,6]. Numerous natural systems have extreme complexity, because complexity is the intrinsic characteristic of the whole natural world [7]. For example, there are a number of biological tissues, mineral substances, and biominerals with specific physical properties, such as piezoelectricity, electrical conductivity, magnetism, etc. (e.g., onion skin, diatomite, eggshell, etc.) [1,2], that are advantageously exploitable for different technological applications.

Natural zeolites represent an interesting class of multifunctional materials, and a number of potentially useful physical and chemical characteristics of these mineral substances have been industrially exploited [8–14]. Usually, natural zeolites have a quite low Si/Al atomic ratio, and for such a reason they are named "aluminous zeolites", in contrast with the "high-siliceous zeolites" that are mainly originated by synthesis. Owing to the aluminum atoms contained in these chemical compounds, the presence of as many extra-lattice cations required to balance the negative charge localized on the aluminum atom follows. Natural zeolites contain different types of cations (typically Na+, K+, Ca2+, Mg2+, and Fe3+), and since these cations may have single or multiple charge, there is a significant fraction of the aluminum atoms in the lattice with unbalanced negative charge. Consequently, electrical transport in zeolite lattice becomes possible by cation hopping among the neighbor negatively charged sites [15,16]. Natural zeolites are therefore solid-state ionic conductors with single-charge carrier (usually corresponding to the largest monovalent cations present in crystal lattice, since they have the lowest charge density [17]). However, these crystalline solids may conduct electricity only at high temperature (since in this case a larger fraction of charge carriers have enough energy exceeding the activation barrier required for jumping between the nearest negative sites) or by hydration/solvation (because in this case the Coulomb interaction between cations and the negative sites decreases) [18–21]. Owing to the presence of a regular array of few angstrom-sized channels (1–13 Å) crossing the zeolite crystal lattice, all cations in the lattice can be solvated only by small polar molecules, that is, molecules with an average size lower than the channel diameter; these include, for example, molecules of water, methanol, formaldehyde, etc. Consequently, natural zeolite crystals can be technologically exploited as both thermal and chemical sensors [22]. In their use as chemical sensor, these materials are able to promptly modify their external surface electrical conductivity by exposition to water vapor. The surface electrical conductivity of natural zeolites with Si/Al close to 5 (e.g., clinoptilolite) may change significantly by exposition to water vapor, and in addition the water molecules spontaneously desorb from these slightly hydrophobic materials when these last are removed from the humid environment. Therefore, zeolite chemical sensors show completely reversible electrical behavior with fast return to the original electrical conductivity value. Owing to the involved ionic conduction mechanism, impedance measurements are required for zeolitic sensors. Sinusoidal or square-wave signals with frequency lower than 1 kHz can be conveniently used. A slab of natural zeolite with a convenient value of the Si/Al ratio can be used as a gas-phase water sensor, and, according to the literature [23], geomorphic clinoptilolite has shown to be very adequate for such an application.

Natural clinoptilolite is the most widely spread zeolite type of natural origin [24]. This crystalline substance is characterized by low cost and excellent mechanical, thermal, and chemical resistance. A slab of natural clinoptilolite has a polycrystalline nature made of highly compacted few micron-sized single lamellar crystals [25,26]. Geomorphic clinoptilolite can be readily used to fabricate water-vapor sensors by applying two parallel electrodes to the surface of a small and perfectly flat slab. The electrodes can be simply obtained by painting two very close rectangular areas on the slab surface by using silver paint. Electrodes must be located on the slab surface in order to exclude from the electrical transport mechanism those cations belonging to internal channel surface. Thus, preventing the slow diffusive mass transport mechanism involved in the solvation of inner cations, the maximum response-fastness possible for this ceramic sensor can be achieved. Owing to the excellent mechanical resistance of clinoptilolite, very robust self-supporting sensors can be fabricated. For the high thermal stability of clinoptilolite, these water-vapor sensors can be exploited for technological applications also at temperatures higher than room temperature.

Recently, the use of naturally available nanostructured mineral substances (e.g., halloysite nanotubes, sepiolite nanofibers, attapulgite, etc.) in the fabrication of resistive humidity sensors to be used for different body-related-humidity-detection applications (e.g., respiratory behavior, speech recognition, skin moisture, non-contact switch, diaper monitoring, etc. [27]) has been proposed in the literature [28–30]. These nature-made sensitive materials have shown good humidity-sensing performances at different relative humidity (RH) conditions, and, in addition, their use represents a facile, low-cost, and environmentally friendly strategy to achieve high-performance sensing devices, without requiring complex manufacturing approaches frequently adopted in this field [31–33].

Inexpensive fast-response ceramic moisture sensors can be conveniently used as devices for apnea syndrome control (i.e., human exhalation sensors) [34–36] by using a miniaturized wearable electric revelation technique. Here, the capability of geomorphic clinoptilolite sensor to promptly and reversibly detect the humidity exhaled during breathing was evaluated by using the "voltage-drop" technique, an original instrumental approach that requires very simple electronic circuitry. Device miniaturizing needs extremely reduced electronic circuitry, which is essential for the wearability of these health monitors (e.g., apnea-syndrome monitor).

#### **2. Materials and Methods**

A commercial sample of natural clinoptilolite (TIP—Technische Industrie Produkte, GmbH, Waibstadt, Germany) was used in the as-received form for fabricating prototypes of the impedimetric sensor. In particular, the sensors were obtained by cutting the raw zeolite piece in the form of rectangular slabs of small thickness (5.0 mm × 10.0 mm × 3.0 mm) by using a mini electric cutting machine (electric mini-grinder, VUM-40, Vigor, Fossano (CN), Italy, equipped with a diamond cutting wheel). Two electrical contacts (1 mm spaced) were painted on the slab surface by using a curable silver paste (EN-06B8, ENSON, Shenzhen, China), and wires were cold-welded to the electrodes by the same paint. According to silver-paste specifications, the obtained device was left to dry in air for 2 days and then baked in an oven at 140 ◦C for 30 min. Two-lead measurements were used to test the moisture sensor; this type of measurement involves a small contact resistance due to the silver–zeolite interface that slightly reduces the device sensitivity. Since the silver electrodes are located at the zeolite surface, only hydrated monovalent cations (i.e., KOH2+) belonging to external surface can participate in the electrical transport.

The clinoptilolite sample morphology was investigated by scanning electron microscopy (SEM, Quanta 200 FEG microscope, FEI, Hillsboro, OR, USA) after its powdering by a hammer. The characteristic Si/Al molar ratio and the type of extra-framework cations present in the mineral were determined by using energy-dispersive X-ray spectroscopy (EDS, Inca 250, Oxford Instruments, Oxford, UK). The type of crystalline solid phases present inside the mineral were identified by using large-angle powder diffraction (XRD, X'Pert PRO, PANalytical, Oxford, UK).

A handheld digital multimeter (DMM, UT-71D, Uni-Trend, Dongguan, China) was used as a true-RMS voltmeter for measuring the voltage across the ac generator output, when it was connected in parallel to the zeolite-based humidity sensor. The DMM included an optically insulated high-speed data-logging system, which allowed the multimeter to record voltage measurements at a speed of 8–9 Sa/s (sampling per second). Such high-speed voltage recording required setting the multimeter in low-resolution mode (4000 counts). In particular, DMM was connected to a PC by a cable, and measurements were recorded by using a dedicated software (UTC/D/E Interface Program, version 3.00, 2017, Uni-Trend Technology, Dongguan, China). Square-wave voltage sources embedded into different handheld DMMs (i.e., DT830D, DT-830B, DT832, ANENG AN8206, ANENG AN8008, KONIG KDM-100, and UT20B) were tested. The best results, in terms of lowest power source, were found with the square-wave source embedded in the DT830D digital multimeter (cheap entry-level DMM, available under different trademarks). The squarewave trace was obtained by an analog oscilloscope (AO-610-2 10 MHz, Voltcraft, distributed by Conrad Electronic SE, Hirschau, Germany), and the harmonic composition of the squarewave signal was obtained by a Fast Fourier Transform (FFT) analysis based on a 2-channel 10 MHz USB oscilloscope (PicoScope 2204A-D2, Pico Technology, Cambridgeshire, UK). A resistance decade box (3280, PeakTech GmbH, Ahrensburg, Germany) was also used to determine the I–V characteristics of the square-wave sources.

#### **3. Results and Discussion**

Natural clinoptilolite usually consists of a combination of different crystalline solid phases (zeolites, quartz, etc.), and the clinoptilolite phase is only the principal component of such a natural composite. As a consequence, the exact composition of the mineral needs to be experimentally determined by using the X-ray diffraction technique (XRD). According to the diffractogram shown in Figure 1, the clinoptilolite sample contained the following main crystalline solid phases: clinoptilolite (48.4 wt.%), anorthite (42.0 wt.%), quartz (8.9 wt.%), and stilbite (0.7 wt.%); indeed, the most intensive peaks belonging to the diffraction patterns of these minerals can be clearly detected in the XRD of Figure 1. According to the literature, the found crystalline phase composition is quite typical for clinoptilolite of natural origin [24].

**Figure 1.** XRD of the natural zeolite used for fabricating the impedimetric water sensor with indication of peaks of the clinoptilolite crystalline phase.

Geomorphic clinoptilolite has a polycrystalline structure. Owing to the extremely high pressure applied to the stone for millions of years, such polycrystalline structure results highly consolidated and characterized by a density (2.15 g/cm3) close to that of a zeolite single-crystal (absence of macro-porosity) [26]. The clinoptilolite phase is made of perfectly staked single-lamellar crystals whose morphology can be easily visualized by Scanning Electron Microscopy (SEM) after having delaminated the mineral by applying a strong compressive stress (e.g., hammer blow). The lamellar morphology of the clinoptilolite single crystals is shown in Figure 2a. As visible, all single-lamellar crystals have exactly the

same thickness, corresponding to 40 nm; while the other two sizes range between 300 nm and 1 μm.

The characteristic Si/Al molar ratio and the type of extra-framework cations contained in the sample have been established by Energy-Dispersive Spectroscopy (EDS). The EDSspectrum of the sample is shown in Figure 2b; the specimen contained the following types of cations: K+, Ca2+, Mg2+, and Fe3+ (in a very little amount). Owing to the low charge density, potassium cations (K+) were the only charge-carriers active in the transport for this ionic conductor. The Si/Al atomic ratio of the clinoptilolite sample can be approximately evaluated by using the silicon and aluminum intensities of signals in the EDS spectrum, and it corresponded to ca. 5.4. Such a value is typical for this kind of mineral. According to this moderately high Si/Al value, the clinoptilolite sample can be considered as a quite hydrophobic zeolite type. The scarce hydrophilicity of the mineral is a property of fundamental importance for a moisture-sensing material, since it allows the sensor to behave reversibly in service.

Since voltage is an electrical property that is easy to measure, the possibility to use a root-mean-square (RMS) digital voltmeter to detect the wearable sensor response is very convenient. Usually, the response to stimuli of an impedimetric sensor is detected as impedance (Z) variation that is measured by an LCR-meter, or as variation of the intensity of current flowing in the device. However, if the impedimetric sensor is biased by a very low power generator, the sensor response can be detected also as voltage drop at sensor electrodes. In particular, the lower the generator power is, the higher the sensitivity of this method. In addition, the I–V characteristics of the generator can be determined and the voltage response can be converted to current intensity or impedance variation by using the generator I–V characteristics.

Owing to the prompt response and reversible behavior, the moisture sensor based on clinoptilolite can be used as apnea syndrome monitor. The human exhalation pattern could be obtained by measuring the current intensity flowing at sample surface after biasing this sensor, for example, with a 5 kHz sinusoidal voltage signal (20 Vpp). However, a similar pattern with much higher resolution can be generated by using the here proposed "voltage-drop" method. To the best of our knowledge, such an instrumental approach, based on a combination of an extremely low power ac generator and true-RMS digital voltmeter, has never been proposed to detect the response of an impedimetric ceramic sensor, such as, for example, a geomorphic zeolite moisture sensor.

When a zeolitic sensor is biased by a sinusoidal or square-wave voltage signal, produced by a standard ac source (function generator with power >1 W), the very high impedance (few MΩ) of the zeolitic sensor can cause just a negligible voltage drop during the exhalation detection by the sensor. However, many digital multimeters (DMMs) incorporate a square-wave source (signal tracer) of very low power (of the μW order), and if these generators are used to bias the sensor, a significant voltage drop can be experienced

even for a slight variation of the high load impedance. Such a type of signal generator is adequate to detect stimuli by an impedimetric zeolitic sensor by measuring the voltage drop at the sensor electrodes. An alternated voltage source is required to avoid ion accumulation at electrodes surface. In particular, the square-wave output incorporated in the DT830D multimeter originates from the liquid crystal display (LCD) driver in the 7106 multimeter chip. The integrated 7106 is an analog-to-digital converter (ADC), which also provides the LCD back panel driving signal with flipping polarity.

The zeolite sensor was connected to the square-wave output, and a true-RMS digital voltmeter (UT71D digital multimeter, UNI-Trend Technology, Dongguan, China) was placed in parallel to this generator, according to the electrical scheme shown in Figure 3. The square-wave signal characteristics are show in Figure 4a,b. As visible in the oscilloscope trace, when a coupling capacitor of 10 nF was used to filter the small offset voltage contained in the signal, the power source gave a perfectly symmetrical square-wave with an amplitude of 5.0 Vpp (for a square-wave signal, the effective voltage, Veff, as measured by a true-RMS digital voltmeter, exactly corresponds to the peak voltage value, Vp = 2.5). As is visible in Figure 4b, the square-wave signal analysis in the frequency domain showed that the signal was composed by the fundamental at the frequency of 50 Hz and three main odd harmonics.

**Figure 3.** Electrical circuit used to record the voltage-drop measurements.

**Figure 4.** Square-wave signal (**a**) with its harmonic analysis by FFT oscilloscope (**b**) and sensor prototype (**c**).

Time-resolved true-RMS voltage measurements were recorder during exposition to human breathing by using the digital voltmeter UT71D data-logging system. An example of a breathing pattern, including three exhalation steps obtained by the voltage drop technique, is shown in Figure 5. According to this breathing pattern, the surface resistivity was quickly modified in the presence of the water vapor, and the voltage decreased by ca. 100 mV in average. As is visible, water adsorption was a very fast process, and it was described by a linear temporal behavior, while water desorption was slightly slower and followed a parabolic law.

**Figure 5.** Human-breathing pattern (exhalations) obtained by the voltage drop technique.

The I–V characteristics of the square-wave voltage source are shown in Figure 6a. This curve was obtained by applying to the generator a gradually increasing pure resistive load of a precisely known value (tolerance of resistors: ±1%), and measuring the output voltage and the absorbed current by true-RMS digital voltmeter and digital ammeter, respectively. The generator I–V characteristics clearly show a non-ideal behavior, with an electromotive force (EMF) of ca. 2.57 V and an internal resistance value (ratio between the EMF and Imax) of ca. 120 kΩ. In particular, the short-circuit voltage (V = 0 and I = Imax) and the open-circuit voltage, that is, the voltage without load (V = Vmax and I = 0), were obtained by extrapolation, since direct measurement of the short-circuit voltage could damage the generator because of the high current intensity flowing in the circuit (overload protection is not present in such simple type of generators), while the open-circuit voltage is limited by the digital multimeter input impedance (usually 10 MΩ at a frequency of 50–60 Hz). In order to avoid the influence of the digital voltmeter input impedance, the voltage and current values were measured for resistive loads inferior to 10 MΩ. As is visible in Figure 6b, the power curve of the square-wave generator shows a dependence of the output power on the applied resistive load, with a maximum value of ca. 14 μW. Such maximum power is achieved with a pure resistive load of 120 kΩ that corresponds exactly to the internal resistance of the generator.

As indicated above, the exhalation pattern displayed as a voltage-drop sequence (see Figure 5) can be converted to a true-RMS current intensity variation or to a normalized impedance variation by using the determined I–V characteristics of the square-wave generator (see Figure 7a,b). In particular, the following relationship was used to convert the time-resolved voltage-drop measurements to time-resolved current-intensity data:

$$\mathbf{I} = \mathbf{I}\_{\text{sc}} \cdot \left(\mathbf{1} - \frac{\mathbf{V}}{\mathbf{E}}\right) \tag{1}$$

where Isc = Imax = 21.3 μA is the short-circuit current intensity, and E is the electromotive force. Similarly, time-resolved zeolite-sensor impedance (Z) values can be obtained from the experimental voltage data by the following relationship:

$$Z = \frac{\mathbf{V}}{\mathbf{I}} = \frac{\mathbf{V}}{\mathbf{I}\_{\text{SC}} \cdot \left(\mathbf{1} - \frac{\mathbf{V}}{\mathbf{E}}\right)} = \frac{\mathbf{V}}{\left(\mathbf{I}\_{\text{SC}} - \frac{1}{\mathbf{r}} \cdot \mathbf{V}\right)},\tag{2}$$

where r = 120 kΩ is the generator internal resistance. As visible, when the breath pattern is expressed as temporal variation of the impedance, the exhalation signals result in being quite deformed (see Figure 7b).

**Figure 6.** I-V characteristics of the square-wave voltage source (**a**) and its corresponding power curve (**b**).

**Figure 7.** Exhalation signals displayed as time-resolved current intensity (**a**) and impedance (**b**).

The observed very fast response of the clinoptilolite-based sensor to the exhaled humidity is related to the special ionic transport mechanism that involves exclusively extra-framework cations located on the external surface of the zeolite slab. In fact, since the electrical contacts are located on the surface, only surface cations are subjected to the applied alternated electric field. During exhalation, these cations are directly exposed to the external humid environment, and, therefore, they readily interact with water molecular dipole by ion-cation electrostatic forces, thus leading to their hydration (solvation). The hydration process (involving one or more water molecules) significantly reduces the cation charge density, weakening, as a consequence, the Coulomb interaction between cations and the negative charge spread in the closed zeolite framework region (nucleophilic area). The consequent strong increase of cation mobility determines an increase in the current intensity moving on the slab surface, with a drop of the voltage between the silver electrodes. During inhalation, the low humidity content characterizing the environment close to the sensor surface determines a desorption of the water molecules from the cations (ion-dipole is a quite weak physical interaction) with restoration of the original very low surface conductivity of the zeolite slab. Such equilibrium between surface and environmental

water molecules allows the electrical monitoring of the moisture fluctuations in the space immediately close to the electrode surface.

#### **4. Conclusions**

A variety of ac square-wave voltage sources embedded into commercial entry-level digital multimeters (DMM) that are connected in parallel with a true-RMS voltmeter (voltagedrop method) have been tested to detect the signal coming from a simple clinoptilolitebased impedimetric water-vapor sensor. The most sensible response was observed with a square-wave voltage source of 14 μW (maximum) embedded in the DT830D multimeter model. In particular, the water-vapor sensor was fabricated by cutting a piece of geomorphic natural clinoptilolite stone in form of small slab and painting two 1 mm–spaced rectangular electrodes on its surface. When the sensor was exposed to water vapor contained in the exhaled human breath, the slab surface impedance rapidly decreased, and this event was recorded as a voltage drop by the voltmeter data-logging system. Owing to the clinoptilolite surface hydrophobic nature, during the inhalation stage of the human breathing, water molecules rapidly desorbed from cations located on the external surface of the impedimetric sensor, thus determining a fast increase of its impedance value. The sensor showed a very prompt and completely reversible behavior at room temperature, thus allowing its technological exploitation as breath sensor. For example, this instrumental approach can be conveniently used for human-exhalation detection in apnea-syndrome monitoring. The principal physicochemical characteristics of the geomorphic clinoptilolite used for sensors fabrication were also determined by XRD and SEM/EDS.

**Author Contributions:** Conceptualization, G.C. and L.N.; Formal analysis, G.C.; Funding acquisition, G.C. Both authors have contributed equally to this research. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Article* **Functional Polymeric Coatings for CsI(Tl) Scintillators**

#### **Gianfranco Carotenuto 1, Angela Longo 1, Giuseppe Nenna 2, Ubaldo Coscia <sup>3</sup> and Mariano Palomba 1,\***

	- I-80126 Napoli, NA, Italy; ubaldo.coscia@unina.it

**Abstract:** The handling of inorganic scintillators (e.g., alkali metal halides) can benefit from the availability of polymeric materials able to adhere to their surface. Polymeric materials, such as epoxy resins, can act as protective coatings, as adhesives for photodiodes to be connected with the scintillator surface, and as a matrix for functional fillers to improve the optical properties of scintillators. Here, the optical properties of two epoxy resins (E-30 by Prochima, and Technovit Epox by Heraeus Kulzer) deposited on the surface of a scintillator crystal made of CsI(Tl) were investigated, in order to improve the detection of high-energy radiation. It is found that these resins are capable of adhering to the surface of alkali metal halides. Adhesion, active at the epoxy–CsI(Tl) interface, can be explained on the basis of Coulomb forces acting between the ionic solid surface and an ionic intermediate of synthesis generated during the epoxy setting reaction. Technovit Epox showed higher transparency, and it was also functionalized by embedding white powdered pigments (PTFE or BaSO4) to achieve an optically reflective coating on the scintillator surface.

**Keywords:** optical-grade epoxy; inorganic scintillator; alkali metal halides; adhesion; interface; Coulomb forces; optical properties

#### **1. Introduction**

Alkali metal halide crystals, such as NaCl, NaI(Tl), CsI, and CsI(Tl), are excellent optical windows and scintillator materials to detect high-energy radiation (e.g., γ-rays, and X-rays) [1–13]. These inorganic scintillators offer (i) a high output and energy resolution, (ii) a fast and high linear response, and (iii) a very stable light output over a wide range of temperatures; however, they are moisture sensitive and, therefore, quite difficult to handle [4]. In addition, the difficult processing of these materials strongly limits their technological exploitation. Usually, to solve these problems, the crystal is covered by PTFE tape; however, this technological approach has several limitations (tape breaking, non-uniform shape, etc.). Therefore, in this paper, we propose replacing the tape with structural adhesive polymeric materials such as epoxy resins containing reflective powders.

It is known that ionic scintillators are solids made of close-packed alkali metal cations and halogen anions, interacting by Coulomb's electrostatic forces, and adhesion is not a critical issue for ionic solids such as ceramic materials (e.g., potteries and glasses) since their surfaces have a layer of hydroxyl groups that allows a strong grafting with the adhesive phase [14]. However, such a layer is not present on the surface of an alkali metal halide crystal, and therefore the polymer–crystal interfacial adhesion is a very critical issue for these materials.

It is well known that polymers have a more or less effective capability of adhering to different solid surfaces by physical or even chemical interactions, depending on the type of side group [15]. For such a reason, the adhesion of thermoplastic polymers to an ionic surface should improve upon increasing the side group polarity, but a mechanically

**Citation:** Carotenuto, G.; Longo, A.; Nenna, G.; Coscia, U.; Palomba, M. Functional Polymeric Coatings for CsI(Tl) Scintillators. *Coatings* **2021**, *11*, 1279. https://doi.org/10.3390/ coatings11111279

Academic Editor: Je Moon Yun

Received: 16 September 2021 Accepted: 18 October 2021 Published: 21 October 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

stable polymer–ionic solid interface can never be achieved by ion–dipole interactions. On the other hand, epoxy resin is an almost universal structural adhesive class, able to guarantee a strong interface with solids such as ceramics, inorganic glasses, metals, and even most plastics [16,17]. Since ionic groups (alkoxide and ammonium) are generated in the epoxy structure during the setting reaction [18], these materials can determine an effective interaction with the cations/anions present on the alkali metal halide surface, which provides the adhesion with the crystal.

Here, the adhesion of two optical-grade epoxy resins to a CsI(Tl) crystal was tested after their characterization by UV–Vis spectroscopy, fluorescence spectroscopy, scanning electron microscopy (SEM), and energy-dispersive X-ray spectroscopy (EDS). To improve the reflectance of the coating layer, these resins were filled by white organic/inorganic pigments of PTFE or BaSO4.

#### **2. Materials and Methods**

The experimental activity was focused on the surface modification of inorganic scintillators, based on thallium-doped cesium iodide, CsI(Tl), monocrystals, by epoxy resins. The aim of the surface treatment was the protection/functionalization of the scintillator surface, and it was carried out on single crystals, having a cubic shape with a side of 35 mm. Since CsI is highly hygroscopic and sensitive to oxidation, the scintillators were stored in a desiccator with well-activated silica gel. Two types of optical-grade epoxies: E-30 by Prochima and Technovit Epox by Heraeus Kulzer, were tested as scintillator adhesives. The base resin-to-hardener weight ratio was 5:3 for the E-30 epoxy resin, and 3:1 for Technovit Epox. In the case of Technovit Epox, a fast hardener was used. An as-prepared epoxy resin mixture was applied by spraying it on the crystal surface. After that, to eliminate air bubbles and other defects formed in the deposited layer, the covered crystals were kept in oven, under vacuum, at a temperature of 40 ◦C (for E-30) and 80 ◦C (for Technovit Epox), for ca. 2 h, and then the samples were left to cure in air for 2 days. Teflon nanopowder (PTFE, Aldrich, St. Louis, MO, USA) and barium sulphate (BaSO4, 99%, Alfa-Aesar, Haverhill, MA, USA) were selected as a white reflective filler for Technovit Epox. In particular, the filler suspension in the base resin component was treated by an ultrasonic bath for 30 min to improve dispersion. After the hardener addition to the base epoxy, accurate mixture stirring, and its deposition on the CsI(Tl) crystal, the coating was allowed to cure at a temperature of 80 ◦C for 2 h. To obtain an optimal compromise between visible reflectivity and coating uniformity, it was required for the epoxy-based composites were required to be filled with 5.4% by weight of PTFE or 21.3% by weight of BaSO4.

Microscopic characterization and elemental analysis were performed by a scanning electron microscope (FEI Quanta 200 FEG microscope, Theromo Fischer Scientific, Hillsboro, OR, USA )and energy-dispersive X-ray spectroscopy (Inca Oxford 250, Oxford, UK), respectively. Absorption and emission optical spectroscopy measurements were conducted by a UV–Vis–NIR spectrophotometer (PerkinElmer, Lambda-900, Hong Kong, China), equipped with an integrating sphere (diameter of 15 cm, PerkinElmer, Hong Kong, China), covered by Spectralon, and a spectrofluorometer (PerkinElmer, LS-55, Hong Kong, China), respectively.

#### **3. Results and Discussion**

The surface morphology of the "as-received" CsI(Tl) crystals was investigated by scanning electron microscopy (SEM, FEI Quanta 200 FEG microscope, Theromo Fischer Scientific, Hillsboro, OR, USA), and energy-dispersive X-ray spectroscopy (EDS) was used in order to quantify the surface oxidation extent. Figure 1a shows the SEM micrograph of the fractured crystal surface. In the inset, the brittle fracture clearly evidences the many dislocations present in the crystal lattice. The EDS spectrum of the fractured surface is shown in Figure 1b where intensive signals of cesium and iodine elements can be appreciated, while the thallium signal is not visible because it is present in a very small amount. A low-intensity oxygen signal is also present, indicating a slight surface oxidation, and the very small carbon peak is due to the SEM sample preparation.

**Figure 1.** SEM micrograph of the CsI(Tl) crystal surface (**a**) and related EDS spectrum (**b**). The inset of (**a**) shows a magnification of the surface dislocations.

The absorption properties of the CsI(Tl) crystal were measured by absorption spectroscopy. As shown in Figure 2a, even a very thin slice of CsI(Tl) showed a rapidly saturating absorption band starting at 315 nm, while the sample was uniformly transparent in the 320–800 nm spectral range.

**Figure 2.** UV–Vis absorption spectrum of CsI(Tl) (slice of 0.883 mm) (**a**), and UV–Vis excitation– emission spectra of CsI(Tl) (**b**). In the inset, a magnification of the emission peak is shown.

The fluorescence characteristics of the scintillator crystals were explored by fluorimetry. As shown in Figure 2b, one visible light emission of green color was found under ultraviolet excitation. In particular, the excitation spectrum (blue curve) shows a cut-off in the UV region, while the emission spectrum (red curve) presents a symmetric peak, centered at 550 nm (see the inset of Figure 2b).

For optimal scintillator operation, it is required to avoid the full attenuation of the fluorescence signal emitted under exposure to high-energy radiation (γ-rays, X-rays, etc.). Therefore, the optical transparency of the epoxy resins (E-30 by Prochima and Technovit Epox by Heraeus Kulzer) in the green spectral range (520–560 nm) was measured. As it can be seen in Figure 3, after curing, these two resins, which were 0.5 mm thick, they had a cut-off value of 272 nm and 290 nm, respectively. Both cut-off values are quite close to the CsI(Tl) crystal self-absorption wavelength, which is 315 nm; however, the optimal optical features were found for the Technovit Epox resin because of the higher transmittance in the UV–Vis–NIR spectral range.

**Figure 3.** Transmittance spectra of the cured E-30 and Technovit Epox resins, and CsI(Tl) crystal.

Spray technology was used to deposit the Technovit Epox resin on the CsI(Tl) crystal surface. As shown in Figure 4a, a continuous and defect-free coating layer of 0.5 mm thickness was deposited, leading to a coated crystal with a size of 35 mm. Such a type of coating can only slightly attenuate the fluorescent green light emitted by the scintillator crystal under UV radiation (see Figure 4b). According to the difficult peel-out of the Technovit Epox coating, a mechanically robust epoxy–crystal interface resulted. This very good adhesion property could be attributed to some special chemical interaction active at the epoxy–CsI(Tl) interface.

**Figure 4.** CsI(Tl) scintillator crystal coated by Technovit Epox resin under visible (**a**) and UV light (**b**).

In this work, it was observed that epoxy resin showed good adhesion to the CsI(Tl) crystal surface even without preliminary treatments. Such bonding can be explained on the basis of the formation, during the curing reaction, of an ionic intermediate of synthesis (i.e., -CH(O−)-CH2-NH2 +-) that could persist at the organic/inorganic interface because of electrostatic interactions with the Cs<sup>+</sup> and I<sup>−</sup> ions present on the crystal surface.

To improve the collection of the emitted fluorescence signal by multiple light reflections, the possibility of filling the epoxy layer with a white pigment was also verified. The deposition of such a functionalized coating layer increases the reflectance of the crystal walls.

In the case of the white pigment based on PTFE nanopowder, a not very uniform dispersion was obtained for a filling factor higher than 5.4% by weight, Figure 5a, and large aggregates of PTFE grains appeared in the deposited layer, Figure 5b.

**Figure 5.** Surface of a CsI(Tl) scintillator coated by a reflective layer of Technovit Epox resin filled with PTFE nanopowder (**a**). An optical micrograph of the achieved surface microstructure, where the arrows indicate the aggregates present in the coating layer (**b**).

To improve the coating uniformity, BaSO4 powder was tested as a filler, since it is a white reflective pigment widely used in optics [19–21]. According to the SEM micrograph and the EDS spectrum shown in Figure 6a,b, the BaSO4 powder had an average size of 431 nm and contained copper impurity (2.3% by weight).

**Figure 6.** SEM micrograph of the BaSO4 powder (**a**), related EDS spectrum (**b**), and BaSO4/epoxy sample image (**c**).

When sonication was applied during the preparation of the epoxy/BaSO4 mixture, a much higher filling factor was achieved (21.3% by weight), without observing significant grain aggregation in the coating (see Figure 6c).

A reasonably good result was achieved by filling the epoxy resin with BaSO4 powder. Figure 7 shows the total reflectance spectrum in the visible range of the pure CsI(Tl) scintillator crystal (blue curve) and the CsI(Tl) scintillator crystal coated by a BaSO4/epoxy layer (green curve). These optical measurements clearly show a higher reflectance value for the coated crystal.

**Figure 7.** Total reflectance spectra of: CsI(Tl) scintillator (blue curve) and CsI(Tl) scintillator coated by a BaSO4/epoxy layer (green curve).

#### **4. Conclusions**

The optical characteristics of two commercial optical-grade epoxy resins used as coatings for CsI(Tl) crystals were compared. The best optical properties were found for Technovit Epox that also showed good adhesion characteristics for the CsI(Tl) crystal surface. Such adhesive properties of epoxies toward alkali metal halides could be ascribed to the possibility for these macromolecules to generate, during the setting reaction, an ionic intermediate that may electrostatically interact with cations and anions at the salt–resin interface. The functionalization the epoxy coating with a white pigment (BaSO4 powder), in order to improve the reflectance on the scintillator surface in the visible spectral region, was also investigated.

**Author Contributions:** Conceptualization, G.C.; methodology, G.C.; validation, G.C. and U.C.; formal analysis, A.L., M.P. and G.N.; investigation, G.C., A.L., M.P. and G.N.; data curation, G.C., A.L. and M.P.; writing—original draft preparation, G.C., A.L. and M.P.; writing—review and editing, G.C., A.L., M.P., U.C. and G.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors kindly acknowledge the projects titled: "Nanocompositi polimerici per applicazione ottiche" (CNR) and "Calocube" (INFN). The authors are grateful to Maria Cristina Del Barone of the LAMEST laboratory (IPCB-CNR) for the SEM/EDS measurements.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

