*Article* **Enhanced Crystallinity and Luminescence Characteristics of Hexagonal Boron Nitride Doped with Cerium Ions According to Tempering Temperatures**

#### **Jae Yong Jung 1, Juna Kim 2, Yang Do Kim 2, Young-Kuk Kim 3, Hee-Ryoung Cha 3, Jung-Goo Lee 3, Chang Sik Son 1,\* and Donghyun Hwang 1,\***


Received: 3 December 2020; Accepted: 29 December 2020; Published: 3 January 2021

**Abstract:** Hexagonal boron nitride was synthesized by pyrolysis using boric acid and melamine. At this time, to impart luminescence, rare earth cerium ions were added to synthesize hexagonal boron nitride nanophosphor particles exhibiting deep blue emission. To investigate the changes in crystallinity and luminescence according to the re-heating temperature, samples which had been subjected to pyrolysis at 900 ◦C were subjected to re-heating from 1100 ◦C to 1400 ◦C. Crystallinity and luminescence were enhanced according to changes in the reheating temperature. The synthesized cerium ion-doped hexagonal boron nitride nanoparticle phosphor was applied to the anti-counterfeiting field to prepare an ink that can only be identified under UV light.

**Keywords:** hexagonal boron nitride; photoluminescence; cerium; anti-counterfeiting; crystals

#### **1. Introduction**

Hexagonal boron nitride (h-BN) is an important material with excellent properties, such as a wide band gap (4.4~6.0 eV), high electrical insulation, low dielectric constant, high temperature stability and large-scale oxidation cross-sectional area, and thermal neutrons [1–3]. All these features have made h-BN a promising material in the aviation industry. Moreover, it has many uses in microelectronic mechanical systems (MEMs), biomedicine, fireproofing, laser devices, solid-state neutron detectors, lubricants and electrical insulators [4–6]. Among the features of h-BN that make it suitable for such varied uses, the wide band gap allows it to be used in applications that require a material with unique luminescence characteristics.

Since the first observation of intense far-ultraviolet (UV) excitation emission, the unique optical and fluorescent properties of h-BN have attracted special attention in the past ten years, making this material a candidate for new light-emitting devices, and it is expected to be used as a future material for photovoltaic applications [7–9]. In addition, the band gap energy of h-BN can be adjusted by designing a super lattice, doping, recombination of organic functional groups and surface functionalization. Doping with rare earth (RE) ions is the most widely known method of tuning the band gap energy of h-BN. Continued interest in rare earth-doped nitride-based materials appears to be increasing, especially for h-BN, partly because of the discovery that the thermal quenching of luminescence decreases

as the bandgap of the host materials increases [10,11]. In particular, metal nitrides or oxynitrides containing activator ions are often used as color-converting phosphors for white light-emitting diodes (wLEDs), because they have excellent visible light emission and less thermal quenching. However, most nitride-based host materials used in phosphors are ternary or quaternary nitride semiconductors. Binary nitride semiconductors, such as aluminum nitride (AlN), gallium nitride (GaN) and BN are rarely used as host materials for luminescent activators [12–14]. Steckl et al. proposed a diode that emits green light in the visible region by doping erbium ions using GaN as a host material [15]. In a study by Jadwisienczak et al., AlN thin film doped with terbium ions by sputtering was grown on a silicon wafer and a light-emitting device in the visible region was produced by a re-heating process [16]. Kim et al. investigated the changes in the luminescence of deep blue with various doping concentrations of cerium ions using h-BN as a host [17]. Mauro et al. described the optical properties of dilute semiconductor materials by presenting a set of highly efficient analytical equations that focus on the evolution of the peak luminescence gain with temperature and the relationship to sample quality [18].

In this study, the crystallinity and luminescence properties of cerium ion-doped hexagonal boron nitride (h-BN) according to the change of reheating temperature were investigated by X-ray diffraction, transmission electron microscopy, Raman spectroscopy, photoluminescence and photoluminescence excitation analysis.

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

#### *2.1. Synthesis of Hexagonal Boron Nitride Nanophosphor Doped with Ce3*<sup>+</sup>

The experimental procedure is schematically shown in Figure 1. Hexagonal boron nitride nanophosphors were synthesized from boric acid (H3BO3, Sigma-Aldrich, St. Louis, MO, USA, ≥98.5%) and melamine (C3H6N6, Sigma-Aldrich, 99%). First, 7 mmol of H3BO3 and 1 mmol of C3H6N6 were dissolved in 150 mL deionized (D.I) water. Cerium 0.05 mmol was also incorporated by dissolution of cerium nitrate (Ce(NO3)3·6H2O, Sigma-Aldrich, St. Louis, MO, USA, 99.999%). These materials were completely dissolved in D.I water, evaporated at 120 ◦C while stirring at 500 rpm to obtain the precursor, and then dried at 80 ◦C for 24 h. The dried precursor underwent pyrolysis in an alumina tubular furnace (AJEON FURNACE, Namyangju, Korea) at 900 ◦C under nitrogen atmosphere. After heat treatment, the sample was rinsed with D.I water to remove any remaining unreacted material. When the powder was collected by agglomeration in the direction of gravity using a centrifuge, the solution was discarded, and the recovered powder specimen was dried at 80 ◦C for 12 h. The samples of h-BN synthesized at 900 ◦C were subjected to re-heating at 1100, 1200, 1300 and 1400 ◦C for 2 h under nitrogen atmosphere to investigate their crystallinity and luminescence in relation to the tempering temperature.

**Figure 1.** Schematic of preparation of the Ce3<sup>+</sup>-doped h-BN nanophosphors.

#### *2.2. Characterization of h-BN Nanophosphors*

X-ray powder diffraction (XRD, Rigaku, Ultima IV, Tokyo, Japan) analysis was performed to investigate the crystallinity of the samples in relation to various temperatures with Cu-Kα radiation (0.15406 nm) generated at 40 kV and 20 mA. The measurement range was performed from 20 to 80 ◦C and the step scan was performed for 4 s at 0.02 degree intervals. Raman spectra were recorded using a dispersive Raman spectrometer (LabRam-HR 800, Horiba Jobin-Yvon, Longjumeau, France) equipped with a microscope and a 522-nm laser as the excitation source. In addition, the specimen was made into a 1 mm × 1 mm cylindrical pellet and measured in the range of 700 to 2300 after focusing at 1800 gv magnification and focus using the E2g mode of the optical lens. The photoluminescence (PL) of the samples was measured and photoluminescence excitation spectroscopy (PLE) was conducted with a fluorescence spectrometer (FP-6500, JASCO, Tokyo, Japan) equipped with a xenon (Xe) flash lamp. At this time, the supplied energy of the spectral lamp was measured with 200 photomultiplier tube modules (PMT). The PL measurement range was from 250 to 800 nm, and the PLE was from 200 to 550 nm. The morphology and crystallinity of the nanophosphors were observed by transmission electron microscopy (TEM, JEM 2100F, JEOL, Tokyo, Japan). The element distribution was analyzed by energy dispersive X-ray spectroscopy (EDX, X-Max 150, Oxford Instruments, Abingdon, UK) to confirm the presence of rare earths in the samples. The resolution was about 129 eV and the analysis time was X-ray exposure for 2 min for each specimen. At this time, the components were compared with B, N, Ce and C for the component detection analysis of the synthesized parent and the doped material. Spectrometry (V-570, JASCO, Tokyo, Japan) was conducted to investigate the transmittance of the bare glass and thin film-coated glass substrate. The analysis range was from 250 to 800 nm, and analysis was performed using an integrating sphere.

#### *2.3. Applied Anti-Counterfeiting and Fingerprinting*

Anti-counterfeiting inks prepared using Ce3+-doped h-BN nanophosphors 1 wt% were dispersed in an aqueous solution containing 10 wt% of polyvinylpyrrolidone (PVP, M.W. = 14,000). The solution was spin coated on glass at 2000 rpm and banknotes were painted with a brush and then dried at 80 ◦C for an hour. The thin film coated on glass was photographed under UV light. Fingerprints were marked on glass substrates. Then Ce3+-doped h-BN nanophosphor powders were applied to the glass substrate surface and latent fingerprints on the surfaces were carefully wiped off. The latent fingerprints coated with the nanophosphors were developed using a UV lamp, and the appearance of the fingerprints was confirmed by photography.

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

#### *3.1. Crystallinity and Morphology of Ce3*+*-Doped h-BN Nanophosphors*

Ce3+-doped h-BN nanophosphors were synthesized by pyrolysis of a precursor in the form of a chemical compound prepared from boric acid, melamine and cerium nitrate at 900 ◦C under a nitrogen atmosphere. The samples were re-heated at 1100, 1200, 1300 and 1400 ◦C and their crystallinity was investigated by analysis of their XRD patterns as shown in Figure 2a. In all of the synthesized samples, the peak of the (002) phase matched the Joint Committee on Powder Diffraction Standards (JCPDS 34-0421) and the main diffraction peak of hexagonal boron nitride crystalline was observed in the XRD patterns. In the case of the sample synthesized at 900 ◦C, the shape of the (002) phase, which was the main diffraction peak, was slightly broad and showed a relatively weak signal. However, as the re-heating temperature increased, the XRD signal of the main diffraction peak (002) phase changed strongly and clearly. These changes were not the perfect form of h-BN for a sample synthesized at a relatively low temperature. Rather, the formed phase was turbo-stratic boron nitride (t-BN), which has the same crystal structure as h-BN but has hexagonal layers stacked and randomly rotated along the c-axis identified [19]. As the re-heating temperature was increased, the profiles of the diffraction peaks became clearer and their full width at half maximum (FWHM) narrowed (Figure 2b, black symbol); i.e., the crystallinity of t-BN formed by the pyrolysis of the sample synthesized at 900 ◦C improved as the re-heating temperature increased. The interplanar spacing of the (002) peak in the diffraction patterns of the Ce3<sup>+</sup>-doped h-BN nanophosphors gradually increased as the re-heating temperature increased. The FWHM of the samples decreased (Figure 2b, black symbol) and the interplanar spacing of the (002) phase increased (Figure 2b, red symbol).

**Figure 2.** (**a**) X-ray powder diffraction (XRD) patterns, (**b**) full width at half maximum (FWHM) and interplanar spacing of (002) phase according to 900, 1100, 1200, 1300 and 1400 ◦C temperatures.

Increased interplanar spacing and decreased FWHM of (002) peaks are commonly found during crystallization of h-BN [20]. We can see that the crystallinity of h-BN is improved with increased re-heating temperature. For these assumptions to be valid, Raman analysis expressed as the lattice frequency of an intrinsic constant was performed, and the results are shown in Figure 3. The Ce3<sup>+</sup>-doped h-BN nanophosphors' position converts the E2g mode near 1363 cm−<sup>1</sup> of Raman spectra to low frequencies. Because the frequency of the Raman spectrum mode is inversely proportional to the square root of the constituent atomic mass [21], the transition to the lower frequency of the Raman spectrum means that heavy Ce atoms are incorporated into the BN lattice of light elements.

−1 **Figure 3.** (**a**) Raman shift, (**b**) change of integrated and FWHM (inset) of Raman shift according to 900, 1100, 1200, 1300 and 1400 ◦C temperatures.

In addition, according to the re-heating temperature, the increased integrated area and decreased FWHM of the Raman shift indicate enhanced crystallinity, which is consistent with the XRD data. In the case of the specimen pyrolyzed at 900 ◦C and reheated at 1100 ◦C, Raman signals were hardly observed. However, when the reheating temperature was increased to 1200 ◦C or higher, the broad Raman signal strength became stronger and a distinct and strong Raman signal was observed at 1400 ◦C. This tendency is thought to be a result of the increase in crystallinity with increased reheating temperature.

Transmission electron microscopy was performed to observe the shape and morphology of the synthesized Ce3+-doped h-BN nanophosphors as shown in Figure 4. Figure 4a shows a sample pyrolyzed at 900 °C and the shape of the nanoparticles is unclear. When the lattice constant (0.332 nm) was profiled with high-resolution, it was found to be close to the XRD result. In addition, carbon was

observed (Figure 4c) throughout the sample; this was unreacted and unburned material residue due to the relatively low heat-treatment temperature. However, the sample synthesized at 900 ◦C and re-heated at 1400 ◦C showed enhanced crystallinity according to the XRD results and showed a distinct elliptical plate shape of about 20-nm in the TEM image. The lattice constant d(002) observed with high-resolution also decreased (0.334 nm) from the XRD data, and components of B, N and Ce were detected by EDX analysis (Figure 4d), confirming that Ce ions were doped in the h-BN lattice. The XRD and TEM results showed that, as the re-heating temperature increased, the enhanced crystallinity and the size of the particles increased [22]. It is thought to be an important process for synthesizing h-BN.

**Figure 4.** (**a**) Transmission electron microscopy (TEM) image of pyrolysis at 900 ◦C of Ce3<sup>+</sup>-dope h-BN nanophosphors (inset up; high-resolution, inset down; line profile), (**b**) TEM image of re-heating at 1400 ◦C of Ce3<sup>+</sup>-dope h-BN nanophosphors (inset up; high-resolution, inset down; line profile), (**c**) TEM image of observed remain carbon in pyrolysis at 900 ◦C and (**d**) X-ray spectroscopy (EDX) data of pyrolysis at different tempering temperatures (element identification: B, C, O, N and Ce).

#### *3.2. Luminescence of Ce3*+*-Doped h-BN Nanophosphors*

Figure 5 shows the changes in luminescence according to the re-heating temperature of the Ce3+-doped h-BN synthesized at 900 ◦C. When the excitation wavelength was controlled at 304 nm, the photoluminescence (PL) wavelength showed deep-blue emission at 396 nm. The PL intensity was enhanced remarkably as the re-heating temperature increased as shown in Figure 5b. The increase in the PL intensity of the sample with increased re-heating temperature may be attributed to the substitutional incorporation of luminescent Ce3<sup>+</sup> ions into the enhanced crystallinity lattice, as revealed in the previous section. The photoluminescence excitation (PLE) spectra of Ce3<sup>+</sup>-doped h-BN nanophosphors irradiated for emission at 396 nm showed an asymmetric curve centered at 304 nm. This is actually a mirror image of the PL spectra. This has been reported for organic dyes in the emission spectra [23]. Mirror symmetry is usually found for defect centers with weak Jahn-Teller interactions [24], which can be attributed to the substitution of Ce3<sup>+</sup> ions for B atoms in the hexagonal boron nitride framework, as shown in Raman spectroscopy. The broad PL spectrum of cerium-doped nitride materials is usually attributed to the 4f-5d excitation of Ce3<sup>+</sup> ions [25].

**Figure 5.** (**a**) Photoluminescence excitation (PLE) and photoluminescence (PL) spectra and (**b**) integrated PL intensity of Ce3<sup>+</sup>-doped h-BN nanophosphors in relation to re-heating temperatures of 900, 1100, 1200, 1300 and 1400 ◦C.

Because 5d electrons contribute to chemical bonds, the PL intensity and emission band are highly dependent on the chemical environment surrounding the Ce3<sup>+</sup> ions [26]. As a result, the PL spectrum was highly dependent on the crystallinity of the host material. Interestingly, further increase in the re-heating temperature enhanced the crystallinity of Ce3<sup>+</sup>-doped h-BN nanophosphors.

#### *3.3. Anti-Counterfeiting Application of Ce3*+*-Doped h-BN Nanophosphors*

The large intrinsic bandgap (≥5 eV) of BN and the plate-like particle shape allow rare earth-doped boron nitride to be deposited as transparent phosphors on a flat surface. Figure 6a shows a Ce3+-doped h-BN nanophosphors thin film deposited on a glass substrate.

**Figure 6.** (**a**) Transmittance of Ce3+-doped h-BN nanophosphors coated on glass and (**b**) images of Ce3+-doped h-BN nanophosphors fingerprinting development on glass substrate (up) and painted on Korean bank notes (down); used samples were re-heated at 1400 ◦C.

The transmittance of the glass substrate decreased slightly after coating with the phosphor thin film with a colloidal solution of Ce3+-doped h-BN nanophosphors. In daylight, it was difficult to distinguish between the bare glass and the phosphor-coated glass with the naked eye due to the transparency of the phosphor thin film. However, when irradiated with UV light, strong blue light emission can be seen from the phosphor film. The images of fingerprinting development are shown in the upper part of Figure 6b. The bare, donor and powdered images used Ce3<sup>+</sup>-doped h-BN nanophosphors re-heated at 1400 ◦C. The emission image was taken under UV light. The fingerprint was obtained from a person whose fingerprint has a whorl loop. The bare image was blurred in daylight. The powdered images produced by using prepared nanophosphors were eye-catching, because these particles were attached to the moisture component of the fingerprint. Under UV light, the blue emission confirmed that the contrast and resolution of the fingerprint had been improved. A Ce3<sup>+</sup>-doped h-BN nanophosphor colloidal solution was painted on a Korean bank note surface as shown in the lower images in Figure 6b. In daylight, it is difficult to distinguish between bare bank notes and those coated with the phosphor thin film. However, the phosphor-coated bank notes showed an intense blue emission text 'Silla' and label under UV light. Due to its light transmission and visible light emission under UV irradiation, this is hidden under normal conditions and can be recognized by the naked eye under UV radiation. Labels and text composed of h-BN based nanophosphors can be effectively hidden in daylight but can be easily identified under UV light, which is essential for a variety of anti-counterfeiting applications.

#### **4. Conclusions**

Hexagonal boron nitride doped with cerium ions was successfully synthesized by pyrolysis and re-heating of the precursor prepared from a chemical mixture of boric acid, melamine and cerium nitrate. Deep blue emission was detected at 396 nm under excitation with 304 nm from Ce3<sup>+</sup>-doped h-BN nanophosphors. The deep blue emission intensity was notably enhanced by increasing re-heating temperature. Increased re-heating temperature resulted in improved crystallinity and luminescence. It also affected particle growth. The deep blue emission is attributed to the transition of Ce3<sup>+</sup> ions from the 5d level crystal field component to the 4f ground state. In the case of the specimen synthesized at 900 ◦C, the results of XRD patterns and Raman spectrum inferior crystallinity due to relatively low temperature were shown. However, as the reheating treatment temperature increased, the crystallinity improved, which is thought to be the effect of recrystallization caused by heat energy supplied from the outside. In addition, in the TEM image, particles with clear crystals were observed at 1400 ◦C, where the reheating treatment temperature was the highest and, as a result of component analysis, B, N and Ce were identified. This improvement in crystallinity has resulted in an increase in luminescence properties. The solution-based coating of Ce3<sup>+</sup>-doped h-BN nanophosphors was applied to various substrates; the films showed excellent transparency and luminescence with a strong visible blue color. The hiding and easy identification of nanophosphors under normal conditions demonstrates the feasibility of using rare earth-doped hexagonal boron nitride in anti-counterfeiting inks in a variety of applications.

**Author Contributions:** Conceptualization, J.Y.J., D.H. and Y.-K.K.; methodology, J.Y.J., Y.-K.K., C.S.S., Y.D.K. and J.-G.L.; formal analysis, J.Y.J., H.-R.C., J.K. and D.H.; investigation, J.Y.J., J.-G.L., H.-R.C. and J.K.; writing—original draft preparation, J.Y.J.; writing—review and editing, J.Y.J., Y.D.K., D.H. and C.S.S.; visualization, J.Y.J., and Y.D.K.; supervision, D.H.; project administration, C.S.S. 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:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available in the database of the authors at the Faculty of Materials Science and Engineering.

**Acknowledgments:** This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (No. NRF-2019R1A6A3A01095400); and by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF-2018R1A5A1025594); and by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (No. 20193010014850).

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

#### **References**


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

© 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 (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Thermal Properties of Bayfol® HX200 Photopolymer**

#### **Pierre-Alexandre Blanche 1,***∗***, Adoum H. Mahamat <sup>2</sup> and Emmanuel Buoye <sup>2</sup>**


Received: 22 October 2020; Accepted: 26 November 2020; Published: 2 December 2020

**Abstract:** Bayfol® HX200 photopolymer is a holographic recording material used in a variety of applications such as a holographic combiner for a heads-up display and augmented reality, dispersive grating for spectrometers, and notch filters for Raman spectroscopy. For these systems, the thermal properties of the holographic material are extremely important to consider since temperature can affect the diffraction efficiency of the hologram as well as its spectral bandwidth and diffraction angle. These thermal variations are a consequence of the distance and geometry change of the diffraction Bragg planes recorded inside the material. Because temperatures can vary by a large margin in industrial applications (e.g., automotive industry standards require withstanding temperature up to 125 ◦C), it is also essential to know at which temperature the material starts to be affected by permanent damage if the temperature is raised too high. Using thermogravimetric analysis, as well as spectral measurement on samples with and without hologram, we measured that the Bayfol® HX200 material does not suffer from any permanent thermal degradation below 160 ◦C. From that point, a further increase in temperature induces a decrease in transmission throughout the entire visible region of the spectrum, leading to a reduced transmission for an original 82% down to 27% (including Fresnel reflection). We measured the refractive index change over the temperature range from 24 ◦C to 100 ◦C. Linear interpolation give a slope 4.5 × <sup>10</sup>−<sup>4</sup> *<sup>K</sup>*−<sup>1</sup> for unexposed film, with the extrapolated refractive index at 0 ◦C equal to *<sup>n</sup>*<sup>0</sup> = 1.51. This refractive index change decreases to 3 × <sup>10</sup>−<sup>4</sup> *<sup>K</sup>*−<sup>1</sup> when the material is fully cured with UV light, with a 0 ◦C refractive index equal to *n*<sup>0</sup> = 1.495. Spectral properties of a reflection hologram recorded at 532 nm was measured from 23 ◦C to 171 ◦C. A consistent 10 nm spectral shift increase was observed for the diffraction peak wavelength when the temperature reaches 171 ◦C. From these spectral measurements, we calculated a coefficient of thermal expansion (CTE) of 384 × <sup>10</sup>−<sup>6</sup> *<sup>K</sup>*−<sup>1</sup> by using the coupled wave theory in order to determine the increase of the Bragg plane spacing with temperature.

**Keywords:** photopolymer; temperature; hologram; CTE; thermal degradation; refractive index

#### **1. Introduction**

Over the past few decades, most imaging and nonimaging systems have been designed and built using conventional bulky glass- and metal-based optical elements. Those optical elements have been proved to perform well under many conditions, but they are heavy, expensive, and require long lead time to manufacture. In recent years, optical designers started to shift their focus on the design of thin, lightweight, and easy-to-manufacture optical elements using the method of holographic recording [1–3]. Typically, holographic optical elements can be recorded on several different photosensitive materials such as dichromated gelatin, silver halide, photoresist, and photopolymer [4].

As the demand for small, lightweight, and compact optical systems has grown since augmented and virtual reality technologies have entered the optics industry, interest in holographic optical elements has also rapidly increased. Holographic optical elements are cheap, easy to manufacture, and sensitive to wavelength and incidence angle; they also provide high diffraction efficiency. These optical elements diffract light through the refractive index modulation obtained through their recording and development processes [5,6].

For many applications, the thermal response of the holographic material is particularly important. The linear coefficient of thermal expansion (CTE) can have an effect on both the diffraction spectrum and diffraction angle of the hologram, changing the color and modifying the direction of the diffracted beam [7].

In the case of a holographic optical element such as a lens, the temperature can change the focal length [8]. In the case of dispersion gratings or holographic notch filters, there can be a shift in the wavelength distribution that affects their use in spectrometry or in a laser cavity [9–11]. For a grating coupler such as those used in holographic combiners for augmented reality and for a heads-up display, this could modify both the color and the field of view of the system [12–14].

It is also important to consider the thermal stability of the holographic material for integration into industrial processes and applications [15]. The holographic material and the diffractive structure contained within should be able to withstand the high temperatures encountered during thermoplastic molding and extrusion, multilayers lamination (such as in windshield and security windows), and their use in extreme environments such as defined in automotive, military, and aerospace specifications (up to +125 ◦C) [16,17].

Volume holograms recorded in acrylamide-based photopolymers were investigated for their operational range and reversibility over temperature range of 15–50 ◦C, and relative humidity of 10–80% [18–20]. This material experiences a red wavelength shift in its diffraction response with increased temperature and humidity. Liu et al. [21] investigated the spectral properties of the DCG-based holograms under different temperatures and humidity conditions. They found that the peak diffracted wavelength decreased with temperature and thermal processing time. Lin et al. [22] studied the temperature effect in PQ:PMMA photopolymer and reported enhancement of the diffraction efficiency with temperature postprocessing. SU-8, a commercial photoresin used for photolithography and surface relief holographic gratings, has been shown to shrink, soften, and even collapse at temperatures just above 100 ◦C. [23]. However, to our knowledge, no one has studied the thermal response of the Bayfol® HX200 photopolymer material.

Bayfol® HX200 photopolymer is a holographic recording material distributed by Covestro that has found uses in many applications such as the heads-up display, augmented reality glasses, solar concentrator, and disperser for a spectrometer [14,24,25]. Bayfol® HX200 is an acrylate-based 2-chemistry photopolymer material; its composition and chemistry are described in detail in [26]. The chemical and physical properties of this material, such as the photo-initiation process, spectral photosensitivity, refractive index modulation, and bleaching, have been characterized [26–28]. The Bayfol® HX200 film has an average refractive index of *no* = 1.49 and a thickness of *d* = 16 μm, and it can provide a maximum refractive index modulation of Δ*n* = 0.03, per the manufacturer's specifications [29].

Unfortunately, little is known about its thermal response. The only available information, to our knowledge, is the fact that the Bayfol® HX200 is able to withstand the injection molding procedure with a mold temperature of 70 ◦C [30].

In this paper, we present a study of several thermal characteristics of the Bayfol® HX200 that are critical for its use as an holographic material in industrial processes. We measured the thermal degradation using both thermogravimetry and the transmission spectrum. Both techniques are complementary, giving the maximum temperature under which the material should be kept to ensure its proper optical operation. We quantified the variation of the refractive index as a function of temperature between 24 ◦C and 100 ◦C, for unexposed as well as for exposed materials. Knowing the refractive index allows determination of the precise optical path of the light rays inside the sample and permits the accurate optical design of systems that include this material. This is particularly important for holographic waveguides that rely on total internal reflection, which is

determined by the refractive index. Finally, we are also reporting the measurement of the CTE of the material calculated from the wavelength drift experienced by a reflection hologram with temperature. The CTE and the refractive index are absolutely necessary parameters to calculate the thermal behavior of an optical system that use the Bayfol® HX200 material. Together, these two parameters determine the change in spectral and angular dispersion of holographic optical elements according to temperature, and they were not known before this study.

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

#### *2.1. Thermal Degradation*

To measure the thermal stability of the Bayfol® HX200 material (manufactured by Covestro AG, Leverkusen, Germany), we used both thermogravimetric analysis and spectroscopy.

The thermogravimetric analysis was performed on a TA Instruments (New Castle, DE, USA) TGA 550 machine at a rate of 10 ◦C/min. The photopolymer sample was separated from the backing film and both compounds were measured separately.

To make sure the material does not suffer any optical damage at temperatures below the weight loss recorded by the thermogravimetric experiment, we recorded the transmission spectrum depending on temperature. For this experiment, the sample was prepared as follows: a 4 × <sup>4</sup> cm<sup>2</sup> piece of Bayfol® HX200 material was laminated on a float glass and exposed to sunlight for 5 min until fully bleached. The material was then encapsulated with another float glass using UV curing optical glue NOA61, which can withstand a temperature of 260 ◦C for three hours according to the manufacturer.

The sample was mounted parallel in front of a first surface aluminum mirror and placed into a temperature-regulated oven. The front side of the oven had a transparent window so that the sample could be illuminated. We used an Oceanoptics (now Ocean Insight, Orlando, FL, USA) USB-4000 fiber fed UV-VIS spectrometer to measure the transmission spectrum of the sample in a double-pass experiment where the back mirror reflected the light back at the spetrometer input fiber. The light source was a halogen lamp that was projected on a diffuser, passing through an aperture, and collimated. The amplitude of the spectrum was calibrated using a dark measurement where the light source was turned off, and a 100% measurement where the sample was removed from the optical path.

A thermocouple was used to measure the temperature at the sample location. The temperature was gradually increased from ambient 23 ◦C to 260 ◦C at a rate of 60 ◦C/h in incremental steps. The transmission spectrum was acquired at each step once the temperature reached equilibrium. Then, the temperature was increased to the next step. Since the temperatures recorded at the oven controller and at the thermocouple were different, the temperature steps appeared uneven. The reported temperatures were those measured at the thermocouple, which has a greater precision (0.1 ◦C) than the oven controller. A final spectrum was acquired after the sample was left for 12 h in the oven at 260 ◦C.

#### *2.2. Measurement of the Refractive Index Variation with Temperature and Exposure*

Holographic optical elements are recorded by interference of two mutually coherent beams within the Bayfol® HX200 photopolymer material. As a result of the interference, a refractive index modulation is created within the active region of the material [5,6]. The refractive index spatial modulation *ng*(*x*), within the active region of the grating is approximated to have a sinusoidal structure and it is defined by Equation (1),

$$n\_{\%}(\mathbf{x}) = n\_o + \Delta n\_{\%} \cos \left(\frac{2\pi \mathbf{x}}{\Lambda}\right) \tag{1}$$

where *no* is the average refractive index of the cured Bayfol® HX200 photopolymer, Δ*ng* is the refractive index modulation created by the interference between the signal and reference beams, and Λ is the period of the refractive index modulation.

Although both the average refractive index and refractive index modulation are assumed to be fixed after recording and processing, it is expected that they might change as a result of heat or refrigeration.

The refractive index measurement was performed using a Metricon 2010 Model (Pennington, NJ, USA). This system consists of a red laser at 637 nm, high refractive index prism, pneumatically operated coupling head, and photodetector. The prism and coupling head are heated to temperature values comprised between 24 ◦C and 100 ◦C with increments of 5 ◦C or less.

In Figure 1a, images of the fresh unexposed Bayfol® HX200 material are shown; in Figure 1b, pictures of the material after being fully cured by UV exposure are presented. The sample size was about 5 × <sup>10</sup> cm2. The purple coloration of the fresh unexposed samples presented in Figure 1a was due to the photoinitiator that was responsible for the polymerization of the monomer under light exposure. The Bayfol® HX200 material uses a two-component system composed of a dye and an organoborate salt [26]. Once the material has been cured under UV light, the photoinitiator dye is left in a bleached state that is transparent to the visible light. Consequently, the cured samples presented in Figure 1b are colorless. Measurements on the unexposed material were performed with the room light off to avoid any photochemical reaction.

**Figure 1.** Pictures of Bayfol® HX200 film samples for (**a**) unrecorded/unexposed and (**b**) cured samples.

The Bayfol® HX200 photopolymer film was placed on the base of the high refractive index prism and the coupling head was brought pneumatically to ensure that the film was in near contact with the prism. Due to the soft nature of the Bayfol® HX200 photopolymer material, and the fact that any excessive pressure could alter the bonding of its chemical composition, the pneumatic air pressure was set to 20 psi. A piece of electrical tape was placed over the coupling head to minimize stress in the sample under test. For each measurement, the thermocouple controller was set to the desired temperature, and allowed enough time to heat up the sample prior to taking measurement.

For each refractive index versus temperature measurement, ten measurements were collected from different locations within a piece of the film. The mean of the data was then recorded as the average refractive index, *no*. The temperature of the prism and the coupling were changed to various values ranging between 24◦ and 100 ◦C with increments of 5 ◦C or less, and a linear function was fitted for all the collected refractive indices.

Extrapolation of the fitted function to an input temperature of 0 ◦C yields the average refractive index at 0 ◦C. The slope of the line indicates the variation of refractive index with temperature, *<sup>∂</sup><sup>n</sup> ∂T* . The linear fit function can be expressed as

$$m\_0 = \frac{\partial n}{\partial T}T + n[0].\tag{2}$$

Since the holographic recording is usually done at room temperature, the refractive index modulation is calculated as the difference between the values of *n*<sup>0</sup> for unexposed/unrecorded film and cured film. The cured sample was laminated over 3 mm thick BK7 substrate glass (Thorlabs Inc., Newton, NJ, USA) and placed under 275W Xenon lamp (Osram HLX 64656-FNT Xenophot, Osram Sylvania, Wilmington, MA, USA) for fourteen hours, followed by one hour exposure under the sun to make sure the sample was fully cured.

#### *2.3. Linear Coefficient of Thermal Expansion*

For volume phase holographic gratings such as the one recorded in Bayfol® HX200, the energy distribution depending on the angle and wavelength can be calculated using electromagnetic propagation theory or coupled wave analysis [5]. However, the direction (*θB*) and wavelength (*λB*) of the maximum efficiency can be predicted simply by using Bragg's equation:

$$\sin(\theta\_B) = \frac{\lambda\_B}{2n\Lambda} \tag{3}$$

where *n* is the refractive index of the material and Λ is the distance between the modulation planes (Bragg planes).

When the material in which the hologram is recorded shrinks or swells, the spacing and angle of the modulation planes are affected, changing the *K* vector of the grating (with |*K*| = 2*π*/Λ). This impacts the diffraction angle and the diffracted spectrum of the hologram.

In the simplified configuration of a transmission hologram with no slant angle, only the lateral spacing of the modulation planes changes (|*K* | = 2*π*/(Λ + ΔΛ)), which influences the diffraction angle for a specific wavelength:

$$
\sin(\theta\_B') = \sin(\theta\_B) \frac{1}{1 + \frac{\Delta\Delta}{\Delta\_B}}\tag{4}
$$

Usually, this effect is not particularly visible since the material is laterally constrained by the substrate on which it is laminated. In the case where the two materials (hologram and substrate) have different coefficients of thermal expansion, the system is bending, introducing even more aberrations that could not easily be separated.

On the other hand, for a reflection grating with its modulation planes parallel to the substrate, a change of the material in the thickness direction affects the diffracted wavelength, keeping the diffraction angle the same:

$$
\lambda\_B' = \lambda\_B \left( 1 + \frac{\Delta\Lambda}{\Lambda\_B} \right) \tag{5}
$$

This wavelength change can easily be detected by a spectrometer and is not coupled to any other effect, making it a good candidate for the measurement of the material CTE.

In the more general case where the modulation planes have a slant angle different from 0 or *π*/2, a more rigorous calculation, such as coupled wave analysis, is required to determine the perturbation on both diffraction angle and wavelength.

It should also be noted that the wavelength shift in reflection hologram due to the material swelling has been extensively studied in other materials such as silver halide or dichromated gelatin [31]. This effect has been used for tuning the hologram color and for the production of three colors (red, green, and blue) 3D images with a single laser [32]. However, in the case of these collagen-based materials, the swelling is due to the absorption of water rather than because of CTE, and the effect is orders of magnitude larger than what is expected with temperature change.

For testing the temperature dependence on the spectrum of a reflection hologram, the Bayfol® HX200 sample was prepared as in the previous section: a 3 × <sup>6</sup> cm2 sample was laminated on float glass, then encapsulated after the hologram was recorded. The reflection hologram was recorded with a 532 nm laser with one beam orthogonal to the sample and the other beam incident at 5◦ angle. This angle was introduced so the diffracted beam could be easily separated from the front face reflection.

The transmission spectrum (zero-order) was measured in a double-pass experiment where the illumination light was reflected back by a mirror located behind the sample. The orientation of the incident illumination and the mirror angle were optimized to achieve maximum diffraction efficiency at 532 nm. In Figure 2, the interpolation of the spectrum measured at room temperature with coupled wave analysis is presented. The best fit parameters were Λ = 5637 lp/mm (532 nm wavelength with 5◦ and 180◦ (=0◦ opposite direction) incidence angles), 7.2 μm effective thickness, and a refractive index modulation of Δ*n* = 0.026.

**Figure 2.** Double-pass transmission spectrum (zero-order) of the reflection hologram interpolated by coupled wave analysis.

The temperature was gradually raised from ambient to 170 ◦C over the course of 3 h, during which spectra were acquired at regular intervals. The sample was then allowed to cool down back to ambient temperature over 12 h. Finally, another temperature cycle was run with more spectra acquisition.

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

#### *3.1. Thermal Degradation*

The thermogravimetric analysis of both the Bayfol® HX200 photopolymer material and its cellulose triacetate backing (without photopolymer) is presented in Figure 3.

The photopolymer followed a multistage decomposition that started at 160 ◦C (first onset) and reached its midpoint at 180 ◦C. The backing film showed an initial degradation at the same temperature that is believed to be due to some photopolymer residue on our sample. A much larger weight loss appeared at 325 ◦C, which is consistent with the literature on cellulose triacetate [33].

**Figure 3.** Thermogravimetric analysis of the Bayfol® HX200 photopolymer and the cellulose triacetate backing.

#### *3.2. Spectral Measurement*

In Figure 4, the results of the spectral measurement depending on temperature are presented. The initial spectrum taken at 47 ◦C shows the residual absorption of the sample composed of glass, Bayfol® HX200, and NOA60 glue, as well as the Fresnel reflection from the different interfaces. It can be noted from the picture taken of the sample before thermal treatment (Figure 5a) that the Bayfol® HX200 material is slightly yellow. No spectral change was observed up to 138 ◦C. The next spectrum recorded at 190 ◦C shows a decrease in transmission around 530 nm consistent with the thermogravimetric measurement. This indicates that the decomposition temperature was reached. As the temperature kept increasing, the transmission kept decreasing over the entire 400 nm to 840 nm band.

**Figure 4.** Transmission spectrum of the Bayfol® HX200 sample without hologram, depending on temperature.

Pictures of the sample before and after thermal treatment are presented in Figure 5. It can be noticed from Figure 5b that the brown coloration of the sample was only due to the Bayfol® HX200 material, since there was no color change for the NOA61 optical glue around the material. The optical transmittance of the sample before thermal treatment was 82% integrated over the entire visible spectrum (450–750 nm) and included the Fresnel reflection. After the thermal treatment (262 ◦C for 12 h), the optical transmittance was reduced to 27%.

**Figure 5.** Pictures of the sample before thermal treatment (**a**), and after 12 h at 262 ◦C (**b**).

The precise chemical reaction responsible for the darkening of the photopolymer with temperature is beyond the scope of this research, but it can be deduced from similar behaviors observed in other polymers that the material undergoes thermal oxidation and further breakdown of the polymer chains [34]. To avoid these irreversible reactions, the Bayfol® HX200 material should be kept below 160 ◦C at all times.

#### *3.3. Refractive Index*

Figure 6 shows the change of the refractive index of the Bayfol® HX200 samples as they were heated to temperatures between 24 ◦C and 100 ◦C. The plots show that increased temperatures caused the refractive index to drop for both samples. The highest change is seen with the unrecorded samples as *<sup>∂</sup><sup>n</sup> <sup>∂</sup><sup>T</sup>* is 4.5 <sup>×</sup> <sup>10</sup>−<sup>4</sup> *<sup>K</sup>*<sup>−</sup>1, when the cured samples have the value 3 <sup>×</sup> <sup>10</sup>−<sup>4</sup> *<sup>K</sup>*<sup>−</sup>1.

**Figure 6.** Refractive index variation depending on temperature for the (**a**) unexposed/unrecorded samples and (**b**) cured sample.

From the values of the refractive index presented in Figure 6, it can be noted that the refractive index change due to material curing by light exposure was around 0.015 for the cured sample. This Δ*n* did not reach the maximum of 0.03 as indicated by the Bayfol® HX200 manufacturer data sheet [35]. Knowing the fact that the photopolymer materials rely on the photoinitiated cross-linking, monomer diffusion, and further polymerization to achieve high refractive index modulation [4], it can be understood that uniform illumination as used in this study does not allow for maximum refractive index change. In the case of large area illumination, the monomers could not diffuse far enough from the unexposed regions of the material to increase the material density and refractive index.

#### *3.4. Linear Coefficient of Thermal Expansion*

Figure 7 shows how the transmission spectrum of the reflection hologram drifted with temperature. It can be seen that from 23 ◦C to 171 ◦C, the transmission minimum drifted from 565 nm to 602 nm.

In Figure 8, the wavelength of the transmission minimum was plotted depending on temperature for two temperature cycles (back and forth) between 23 ◦C and 170 ◦C. During the initial temperature increase, an irreversible change occurred at 170 ◦C with a sudden increase in the minimum wavelength. After observation of the sample, we noticed that the NOA glue has softened and allowed the sample to shift from the cover glass. This shift induced a color change in the reflection spectrum of the hologram that can be seen in Figure 9b.

Figure 9 shows pictures of the reflection hologram sample before thermal treatment (left) and after the initial temperature increase to 170 ◦C (right). One can see some permanent change in the form of nonuniform color (pink or yellow) of the hologram after the thermal treatment. Note that the coloration of the sample was not due to material absorption, but to the diffraction from the hologram. It has to be noted that after this initial irreversible change, further temperature variation from 20 ◦C to 170 ◦C induces a linear and fully reversible change in diffracted wavelength.

**Figure 7.** Transmission spectrum of the Bayfol® HX200 reflection hologram depending on temperature.

**Figure 8.** Wavelength minimum of the transmission spectra of the Bayfol® HX200 reflection hologram depending on temperature for two temperature cycles.

**Figure 9.** Pictures of the reflection hologram sample before thermal treatment (**a**) and after initial temperature increase from ambient temperature to 170 ◦C (**b**). Note that the pink coloration of the sample was not due to the material absorption but to the diffraction of the green wavelength that was no more present in the transmission spectra (white light − green = pink).

To calculate the linear coefficient of thermal expansion (CTE) for the Bayfol® HX200 material, we used Equation (5) to retrieve the spacing of the Bragg planes spacing from the measurement of

the diffraction peak minimum wavelength as a function of temperature. This spacing parameter is presented in Figure 10 where it is interpolated by a linear regression. The origin of the straight line interpolation gives the original distance between the planes (*L*), and the slope gives the distance increase depending on temperature (Δ*L*/Δ*T*). The CTE is defined as:

$$a = \frac{\Delta L}{L \Delta T} = 384 \times 10^{-6} \text{ K}^{-1} \tag{6}$$

**Figure 10.** Bragg planes spacing for the Bayfol® HX200 reflection hologram depending on temperature for 1.5 temperature cycles.

The CTE calculated value <sup>384</sup> × <sup>10</sup>−<sup>6</sup> *<sup>K</sup>*−<sup>1</sup> for the Bayfol® HX200 is of the same order of magnitude as the literature values for other polymeric materials such as polypropylene <sup>150</sup> × <sup>10</sup>−<sup>6</sup> *<sup>K</sup>*<sup>−</sup>1, polystyrene 150 × <sup>10</sup>−<sup>6</sup> *<sup>K</sup>*<sup>−</sup>1, or teflon 175 × <sup>10</sup>−<sup>6</sup> *<sup>K</sup>*−<sup>1</sup> [36–38].

#### **4. Conclusions**

Thermogravimetric measurements show that the Bayfol® HX200 material does not suffer from any thermal degradation before 160 ◦C. Up to that temperature, holograms recorded in this material experience only reversible change, expressed as thermal dilatation between the Bragg planes. Irreversible change in the hologram diffraction spectrum started at 170 ◦C, and irreversible change in the transmission spectrum of the material was recorded starting at 190 ◦C.

Refractive index variation depending on the temperature was measured as 4.5 × <sup>10</sup>−<sup>4</sup> *<sup>K</sup>*−<sup>1</sup> for the unexposed material and 3.3 × <sup>10</sup>−<sup>4</sup> *<sup>K</sup>*−<sup>1</sup> once the material was fully cured. These changes are linear depending on the temperature over a range from 24 ◦C to 100 ◦C.

A CTE of 384 × <sup>10</sup>−<sup>6</sup> *<sup>K</sup>*−<sup>1</sup> was calculated by measuring the spectral shift of a reflection hologram depending on temperature. This color shift was due to the dilatation between the Bragg planes experienced when the photopolymer material swelled under temperature increase.

These results will be valuable to predict the characteristics of holographic optical elements and other holograms in applications and processes where thermal variation is expected.

**Author Contributions:** Conceptualization, P.-A.B., A.H.M., and E.B.; methodology, P.-A.B., A.H.M. and E.B.; manuscript writing, review, and editing, P.-A.B., A.H.M. and E.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** Authors acknowledge the support of the NavAir under grant #N162-103-0420 NISE space section 219 program.

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

#### **References**


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

© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

*Article*
