Cellulose derivative/barium titanate composites with high refractive index, conductivity and energy density

This work deals with hydroxypropyl cellulose (HPC)/barium titanate (BT) composites, having small levels of perovskite-like BT loading, i.e. 0.5–2% and being designed for electric energy storage applications. The films were obtained by solution “casting method and their structural properties were confirmed by FTIR. Scanning Electron Microscopy scans reveal that ceramic filler is well and uniformly dispersed within the cellulosic matrix. UV–VIS spectroscopy data indicate a slight increase of the absorption and the decrease of optical band gap from 5.71 eV for the HPC matrix (0% BT) to 5.20 for the sample containing 2% BT. According to refractometry analyses, the presence of BT particles determines an increase in sample polarizability, reflected in a higher refractive index, namely 1.533 and 1.605 at 486 nm for 0% BT and 2% BT samples, respectively. The values of the dispersion energy and single-oscillator energy are decreasing after filling HPC with BT, whereas first/third order optical susceptibilities and nonlinear refractive index are increasing. Mechanical tests show that incorporation of BT particles increase the value of Young modulus from 239 MPa for pure HPC to 342 MPa for 2% BT in the matrix. Dielectric studies prove that the samples exhibit a significant change in the real part of the permittivity with filler loading, ranging from 4.2 for neat HPC up to 8.5 for HPC/BT 2%.


Introduction
In the global context of increasing energy demands and associated issues of planet pollution, the polymer materials with low toxicity seem to be the key for replacing parts of high-tech devices, like high power electronics and energy storage. Dielectric capacitors with high energy density opened fresh perspectives due to lightweight, flexible, and elevate power densities (i.e. rapid discharge of stored electric energy). For such purposes, the performance of the material can be controlled via dielectric constant and breakdown strength by combining plastics with inorganic fillers. Regarding the latter, barium titanate (BT) is an inorganic lead-free ferroelectric compound, largely known for its photorefractive properties (Takenaka 2013), and high permittivity (100-2000) (Petrovsky et al. 2008). Recent reports emphasize the importance of chemical modification of cellulose with the purpose of adapting its properties to widen the applicative horizon (Rana et al. 2021;Zielinska et al. 2021;Voicu and Thakur 2021).
Hydroxypropyl cellulose (HPC) can be regarded as a suitable polymer matrix for composites. This is a thermoplastic cellulose derivative, where some of the hydroxyl groups from the main chain are hydroxypropylated. This polymer displays prevalent plastic properties, combined with relative hydrophobic character. HPC is easily solving in water and polar organic solvents. This cellulose ether is able to render liquid crystalline phases as a function of either temperature (Wojciechowski 2000) or concentration in solution (Navard and Haudin 1986). HPC presents unique properties, such as high transparency and optical activity (Barzic et al. 2014) suitable for fabrication of chiroptical filters (Charlet and Gray 1989), thermoresponsive features adequate for pharmaceutical purposes (Weißenborn and Braunschweig 2019), ordered surface texture in peculiar conditions, which is transferable to other polymers (Cosutchi et al. 2011), excellent biodegradability and biocompatibility useful for medicine (Seddiqi et al. 2021) or green electronics (Irimia-Vladu 2014). One of the main disadvantages of HPC arises from its relatively low termostability, which can affect its performance (Rials and Glasser 1988). To the best of our knowledge, there are no studies that deal with HPC as dielectrics used in energy storage. However, the quite high values of the dielectric constant of HPC found in only few investigations devoted to the dielectric analysis of this cellulosic material in solid state (Rachocki et al. 2005;Shinouda and Abdel Moteleb 2005;Manaila Maximean et al., 2018) showed that HPC is qualified for use in the energy storage applications. Additionally, this artificial polymer is introducing the advantage of eco-compatibility of the device.
Many scientific papers describe BT composites using cellulosic matrix. Ali et al. (2017) showed that when inserting BT nanoparticles in ethyl cellulose up to 13 vol%, they obtain a dielectric constant of * 7.5 at 1 Hz, high dielectric loss and raised dc conductivity. At low amounts of BT and below 60°C, the composites are good for antistatic applications, whereas above 7 vol% of BT and at high temperatures the films are suitable for electrostatic dissipation purposes (Ali et al. 2017). Rotaru et al. (2016) proposed an ultrasound assisted method for the preparation of regenerated cellulose (viscose fiber)/BT composites. The procedure involved heating of the precursors and homogenizing the system via ultrasonication. The resulted composites displayed high dielectric constants, as a function of the content of BT and preparation procedure. The materials were found to be useful as electromagnetic shields. Hassan et al. (2018) studied cellulose nanofiber/BT composites and emphasized the influence of spontaneous polarization of BT filler on the magnitude of the dielectric constant. Jia et al. (2016) used BT particles for enhancing the dielectric performance of cyanoethyl cellulose/antimony tin oxide (CEC/ATO) composites. They found a permittivity value of * 149 at 100 Hz for BT and ATO contents of 30% and 28%, which makes these films attractive for energy storage capacitors. Yin et al. (2020) prepared regenerated cellulose/BT nanofibers composites by using dopamine as compatibilizer; the composites are characterized by a discharged energy density of 17.1 J/cm 3 at 520 MV/m, which is ideal for use in capacitors applications.
Most reported studies are focused on structural, morphological, thermal and dielectric properties of BT-polymer composites and very few address aspects related to their optical properties (Mimura et al. 2010;Nagao et al. 2011;Morsi et al. 2019). This is probably because, at high BT loadings, the transparency of the material is compromised and limits the possibility of material investigation. Optical absorption features, particularly the absorption edges are essential for describing optical transitions, the band structure, the band tail and the band gap energies of composite materials. To our knowledge, only the study of Morsi et al. (2019) describes optical constants of BT composites with polyethylene oxide/carboxymethyl cellulose (PEO/CMC) as matrix. They show that such materials can be used for sensors or as ceramic capacitors. On the other hand, Nagao et al. (2011) prove that the incorporation of BT in poly(methyl methacrylate) increases both the refractive index and the permittivity of the composites, which become ideal for use in electronics.
Previously, we have prepared viscose fiber/BT composites of low BT content and proved the appropriateness of their properties for applications as absorbent materials in electromagnetic fields (Rotaru et al. 2016). In this work, new BT composites with biodegradable hydroxypropyl cellulose (HPC) matrix were prepared. Structural, optical, mechanical and electrical properties were analyzed to demonstrate their usefulness for energy storage. The performed investigations will prove that the HPC/BT materials have higher refractive index than PMMA/ethylene glycol monovinyl ether/BT (Mimura et al. 2010) as well as larger permittivity and conductivity in comparison with PEO/CMC/BT (Morsi et al. 2019) and polypropylene/BT composites (Yao et al. 2018). Energy density evaluations reveal similar values to those reported for polyimide/BT composites (Yue et al. 2019).

Materials
Klucel TM Hydroxypropyl cellulose (HPC) of 100 000 g/mol molar mass and substitution degree of 2.5 was purchased from Aqualon Company.
Barium carbonate and titanium dioxide were purchased from Sigma Aldrich and used as received.

Preparation of BT
Barium titanate (BT) was prepared by using a wetstate reaction under ultrasonic irradiation. The detailed procedure is described in a previous work (Rotaru et al. 2017). Shortly, a 1/2 w/w titanium dioxide/barium carbonate mixture was ultrasonically irradiated at 20 kHz frequency, in Milli-Q ultrapure water for 60 min (119 kJ energy were dissipated and the temperature reached 92°C) by using a generator equipped with probe and sensor for temperature. The obtained powder was further decanted and dried in a microwave furnace for 10 min. The last procedure involved a thermal heating during 3 h, undertaken in a furnace at 500°C to obtain perovskite-like BT particles of submicronic dimensions (400-600 nm).

Preparation of HPC/BT composite films
Film sample preparation was done as follows: a stock polymer solution was obtained by dissolving 5 g HPC in 250 mL of deionized water at room temperature. 50 mL of HPC solution and known amounts of BT particles of different filler loadings (0.5, 1 and 2% BT) were homogenized by ultrasonication for 10 min; during this procedure, 7 kJ energy was dissipated and the temperature increased from 20 to 40°C. The prepared dispersions were poured onto Petri dishes of 10 cm diameter. The films were obtained by drying overnight at room temperature and then they were kept in a vacuum oven at 90°C for 6 h. The pure HPC film was obtained by using a similar procedure.
The thickness of the obtained HPC and composite films ranged between 70 and 140 lm as measured with a digital micrometer.

Characterization
The structures of barium titanate (BT) powder, pristine HPC film and the HPC/BT composites films were investigated by FTIR spectroscopy in potassium bromide pellets using a Bruker Vertex 70 Spectrometer (data are provided in Supplementary Information file).
X-ray diffraction analysis was performed on a Rigaku Miniflex 600 diffractometer using CuKaemission in the angular range 2°-90°(2h) with a scanning step of 0.0025°and a recording rate of 2°/ min.
The mechanical tests performed with a crosshead speed of 10 mm/min were done on an Instron 3365 machine, on dumbbell shaped specimens with a width of 4 mm and a length of 50 mm.
In order to highlight the morphology of the samples, they were examined on the surface and in cross-section. Surface morphology for the control HPC sample and HPC/BT composites were analyzed on a Verios G4 UC scanning electron microscope (SEM) from Thermo Scientific, Czech Republic. The samples were coated with 10 nm platinum using a Leica EM ACE200 Sputter coater to render electrical conductivity and to obstruct charge buildup during exposure to the electron beam. SEM investigations were performed in high vacuum mode using a secondary electron detector (Everhart-Thornley detector, ETD) at accelerating voltage of 5 kV. Cross-sectional analysis of films was performed on a Quanta 200 device (FEI, Czech Republic). To achieve the cross section, the films were immersed in liquid nitrogen to impede deformation during fracture. The presence of the BT was confirmed by energy dispersive X-ray (EDX) analysis using the Quanta 200 microscope coupled with EDX detector.
UV-VIS features of the free-standing films were examined on a SPECORD 210 PLUS device.
Refraction properties of film samples were measured on a DR-M4 Abbe refractometer at several wavelengths from visible range.
Dielectric measurements were carried out on a broadband dielectric spectrometerequipped with a high resolution Alpha-A analyzer. The dielectric spectra were recorded at room temperature, in a frequency range between 1 and 10 6 Hz. The dielectric data were determined for three different samples from each film composition and presented as mediated values.

Results and discussion
HPC was used as matrix for the preparation of organic-inorganic composites of different BT contents. To obtain optically transparent composites a maximum of 2% of BT filler was used.

X-ray diffraction analysis
All X-ray diffractograms of the films (Fig. 1) contain the characteristic 2 signals of HPC centered at 2h = 8.37°accounting for the crystalline portion of HPC and at 2h = 19.3°accounting for the slightly ordered amorphous portion of HPC (Echeverria et al. 2015;Basta et al. 2021). In addition, the diffraction peaks of the BT are visible, their intensity being correlated with the concentration of the salt.
The presence of BT in the films does not alter the diffraction profile of HPC. The degree of crystallinity increases from 39.8% as calculated for HPC up to 45% for the samples with 0.5 or 1% BT. Further increasing the BT content up to 2% in the last sample did not further improve the crystallinity of the sample, the found value of the crystallinity being 36.8, lower than the crystallinity of the neat HPC sample. Similar results were reported for other composites (Mngomezulu et al. 2021;Ram et al. 2020) where no linear dependence of the crystallinity as a function of the filler content as well as the filler dimensions were noticed.

Mechanical properties
The mechanical properties of HPC/BT composites films were investigated by tensile tests. The stressstrain curves for pristine sample and samples with 1% and 2% BT are given in Supplementary Information ( Figure S1). The elongation at break is slightly increasing from about 50% for HPC to about 56% for HPC/BT 2%. The tensile strength of the prepared films increased from 9.9 MPa for HPC matrix to 12.8 MPa for the composite containing 1% BT and to 16.2 MPa for the composite with 2% of BT. Similarly, the increase of the amount of filler resulted in higher Young's modulus. For example, the sample without filler has a Young's modulus of 239 MPa while the composites have a Young's modulus of 298 MPa and 342 MPa, respectively. Overall, the mechanical results showed that the addition of the filler led to higher deformation at break, tensile strength and Young's modulus, these suggesting a good compatibility of BT nanoparticles with the cellulose matrix.

Morphology
The morphological investigations were made by SEM coupled with EDX for the neat HPC and BT filled films. Figure 2 shows the surface micrographs and EDX compositional results. The HPC matrix displays mainly a relatively textured surface sprinkled with small defects probably formed due to partially crystalline nature of cellulose material (Vanderghem et al. 2012;Mohd Ishak et al. 2020). In a previous work (Rotaru et al. 2017), the morphology of the used BT powder was found to consist in agglomerated particles having almost spherical shape, while their diameters varied between 400 and 600 nm. Here, these ceramic particles can be easily observed as white approximately spherical dots which are present for all filler concentrations and with increasing particles' density on the film surface with increasing BT content. The ceramic filler is well and uniformly dispersed within the cellulose derivative matrix on the surfaces of the samples. Relatively similar morphology was noticed for BT/polyaniline composites (Pant et al. 2006).
The qualitative results on film compositions provided by the EDX analysis are presented in the tables inserted in Fig. 2. EDX recordings show the lack of Ba and Ti elements in images corresponding to pristine HPC, while, upon reinforcement, the percentage of these chemical elements is closed to the content of BT filler on both surface and cross-section. Thus, EDX analysis proves the uniform dispersion of the ceramic BT particles not only on the surface but also inside the reinforced polymer films. The FTIR analysis (see Figure S2 and comments in Supplementary Information) confirms the presence of the HPC matrix and of incorporated BT filler in composite samples. Cross-section SEM images of HPC and HPC/BT samples are depicted in Fig. 3. HPC, known for its thermotropic liquid crystalline behavior generated a film which displayed a granular structure, with small pores, typical for a three-dimensional network (Nishio and Takahashi 1984;Silva et al. 2008). The incorporation of small amounts (0.5, 1%) of BT in the HPC matrix does not change this morphology. However, when 2% BT were added into HPC matrix, a pleated morphology of the cross-section of the resulted film was observed. Moreover, HPC/BT films show more compact morphology with a lower porosity as compared to the pristine HPC film, especially for HPC/BT 2% sample, also evidenced by decreasing values of film thickness as a function of BT content (Table 1). Increased compactness of the composite films was earlier evidenced by Banerjee et al. (2019) for poly(vinyl alcohol)/manganese chloride composites. One may also observe a change in the composites roughness as compared to that of the polymer matrix. Similar effect was observed in other polymers/inorganic composites-structures like ''depressions'' which were attributed to the filler distribution in the polymer matrix (Banerjee et al. 2012). The transparency properties of HPC and HPC/BT composite samples were investigated by UV-VIS-NIR spectroscopy on free-standing films. Figure 4a, b show the spectra samples. A sharp absorption edge is noticed in the spectrum of the cellulosic matrix, which is slightly enhanced by the addition of BT particles in the HPC matrix. The observed absorbance behavior of the prepared composite samples presents changes in the magnitude of the optical band gap energy with further loading of BT filler in the system, reflecting the possible interactions between the two phases. Similar spectral features were reported by Morsi et al. (2019) for composites of PEO/CMC filled with BT particles. For the investigated samples, the optical clarity of the prepared materials is gradually diminished and the absorbance is enhanced as BT loading increases from 0 to 2% in the HPC matrix. Thus, the reinforcement of the composite films favors light scattering and enhances the absorption coefficient. This aspect can be observed from a simple visual examination of the free-standing films, since their aspect gradually changes from transparent to almost white color (Fig. 4b). Absorption and scattering phenomena caused by 1% and 2% BT reinforcement of cellulose derivative are altering the transparency of pristine HPC film.
The optical absorption coefficient characterizes the material ability to absorb electromagnetic radiations, particularly from visible range. Optical absorption coefficient was estimated for each of the investigated samples by introducing UV-VIS data in Eq. (1): where a is the optical absorption coefficient, t is the thickness of the film sample and T is the transmittance.  For all spectra illustrated in Fig. 4a, the absorption edge, representing the low energy wing of the first absorption band, is analogous to the edge suggested by Tauc approach (Tauc 1967) for ideal amorphous semiconductors. Given this similarity, the absorption edge parameters were determined from the dependence of absorption coefficient on photon energy, as depicted in Eq. (2): where E is the photon energy (E = ht), E 0 can be either Tauc energy (Et) or Urbach energy (Eu). Graphical representations of Eq.
(2) as semi-logarithmic plots are following the Urbach relation as shown in Fig. 5. Absorption edge energies were extracted from the reverse of slopes of the straight lines and the results are listed in Table 1. Et is the optical parameter extracted from the low-energy exponential zone of the dependence of absorption coefficient on photon energy, while Eu is related to the high-energy exponential region. For pristine HPC matrix, as well as for HPC/BT composites the values of Et are higher than those of Eu. Therefore, the absorption related to possible structural defects (i.e. break or torsion of polymer chains) is more pronounced with regard to absorption determined by structural disorder. No report on Eu and Et was found in literature for polymer/BT composites. Moreover, BT reinforcement generates intensification of these absorption processes as seen in the increase of the Et and Eu parameters.
The photon energy that is demanded for transporting the electrons from the valence to the conduction band can be estimated by examining the region of the absorption boundary. Based on literature data (Choudhary and Sengwa 2018; Morsi et al. 2019;Sengwa et al. 2019), it is well known that the majority of photon absorbing materials are characterized by a direct and an indirect band gap close to fundamental absorption edge. The theory proposed by Tauc (Tauc 1967) allows the calculation of the band gap energy (Eg), as expressed by Eq. (3): where C is a constant and m is a parameter that describes the nature of the electronic transitions, taking the value of 0.5 for direct allowed transitions and 2 for indirect allowed transitions. Figure 6 displays the plots of (aE) 1/m versus photon energy for m = 0.5 and m = 2. Linear parts are noticeable in all plots and their intersection with x-axis is essential for establishing the Eg values for the prepared samples. Electrons are travelling from the valence to the conduction band, while conserving total energy and the wave-vector k-space, denoting a direct transition. On the other side, the movement of the electrons in the conduction band at distinct k-spaces produces an indirect transition. The estimated values of indirect band gap (Egi) and direct band gap (Egd) for the investigated samples are illustrated in Table 1. Regardless the filling levels of HPC, the results indicate lower values for Egi as compared to Egd. Moreover, the reduction of optical band gap energy with enhancement of inorganic filler loading was registered, a common feature for polymer composites (Choudhary and Sengwa 2018;Morsi et al. 2019;Sengwa et al. 2019). The obtained band gap values for HPC/BT are slightly higher than those reported by Morsi et al. (2019) for PEO/CMC filled with BT. The presence of ceramic BT particles within HPC matrix facilitates electron crossing in the valence band with production of supplementary localized energy states in the forbidden energy zone and therefore lowering the band gap of the BT/HPC materials.
The Eg reduction for composites with the variation in content of the constituting phases is believed to enhance with the degree of disorder (higher Eu) in the material.

Refractive index and optical dispersion parameters
The refractive index (n) is an optical parameter that is connected to the light speed through the analyzed material. The experimentally determined n values of pristine HPC and HPC/BT composite films as a function of wavelength are displayed in Fig. 7a. The presented refractive index data are acquired by refractometry method. In contrast to ellipsometry, this technique does not allow evaluation of the imaginary part of the refractive index, which is related to absorption of light. However, information on the absorption properties of the samples that are not negligible can be observed in the UV-VIS spectra from Fig. 4. For all samples the refractive index is decreasing as the light wavelength is gradually increasing. Furthermore, incorporation of BT filler determines the increase of the refractive index comparatively to the pure HPC, namely, at 486 nm, n ranges from 1.533 at 0% BT to 1.605 at 2% BT in the matrix. This kind of enhancement in the refractive index magnitude might be the result of an increase of the composite polarizability and density, as supported by literature data for polymer composites (Banerjee et al. 2019). Comparable results were reported for other polymer/BT composites, namely the refractive index of HPC/BT samples is smaller than that of poly(arylene ether ketone)/BT (1.65 at 12 wt% BT) (Imai et al. 2010), similar to that of PMMA/BT (1.56 at 10 vol% BT) (Nagao et al. 2011) and higher than Fig. 6 Plots of (aE) 0.5 versus E (a) and (aE) 2 versus E (b) for pristine HPC and HPC/BT composite films Fig. 7 Refractive index dependence of wavelength (a) and plots of 1/n 2 -1 versus E 2 (b) for neat HPC and HPC/BT composite films that of PMMA/ethylene glycol monovinyl ether/BT (1.53 for BT/8EGMVE/10H 2 O/5PMMA) (Mimura et al. 2010).
Optical dispersion parameters can be obtained by making a close analysis of the refractive index dependence on photon energy. This is described by the Wemple and DiDomenico (WDD) single oscillator model, as shown by Eq. (4): where E 0 is the average excitation energy for electronic transitions and E d is the dispersion energy. Figure 7b displays the plots of 1/(n 2 -1) against E 2 for HPC matrix and its BT-containing composites. The values of E 0 and E d were estimated from the linear fit of the obtained graphs (Fig. 7b) and the results are provided in Table 2.
The obtained data indicate that the strength of the interband optical transitions, expressed by E d , is decreasing as the filling level is higher. The average excitation energy E 0 follows the same trend. This behavior was also noticed for polyvinyl alcohol (PVA) filled with BT (Barzic et al. 2020). The oscillator energy varies analogously to the optical band gap energy as the reinforcement level in HPC is higher. However, the values of E d , E 0 and n 0 for HPC/BT are slightly smaller than those reported for PVA/BT (Barzic et al. 2020).
The data concerning E 0 and E d can be utilized to calculate the static refractive index (n 0 ) at zero photon energy, as revealed by Eq. (5): As seen in Table 2, the static refractive index is higher as the inorganic particle amount in HPC matrix is increasing.
Optical dispersion parameters lie at the basis for determining nonlinear optical properties (Soliman et al. 2020), namely optical susceptibility and nonlinear refractive index, defined by Eqs. (6-8): where v (1) is the first-order optical susceptibility, v (3) is the third-order optical susceptibility and n NL is the nonlinear refractive index. The achieved data for nonlinear optical properties are listed in Table 2. It was found that the insertion of ceramic particles in HPC generates the enhancement of the first and third order optical susceptibility and also of the nonlinear refractive index. The values of v (1) and v (3) for HPC/BT are smaller comparatively to those of PVA/BT (Barzic et al. 2020), while the nonlinear refractive index of studied samples is higher. These aspects sustain the suitability of the HPC/BT materials for nonlinear optical and photonic devices.

Optical and electrical conductivities
The optical conductivity is connected to the refractive index and the optical absorption coefficient at specific wavelengths. For the analyzed composite materials this parameter was evaluated using Eq. (9): where a was estimated above, as seen in Fig. 5,r opt is the optical conductivity and c is the speed of light in empty space. Table 2 The values of dispersion energy (E d ), single-oscillator energy (E 0 ), static refractive index (n 0 ), linear optical susceptibility (v (1) ), third order optical susceptibility (v (3) ), and nonlinear refractive index (n NL ) for HPC and HPC/BT composite films The conductivity at optical frequencies is depicting the material's optical response caused by the travelling of the charge carriers under the action of external electromagnetic waves. Figure 8 illustrates the dependence of r opt with photon energy for the prepared samples. The optical conductivity of the pristine HPC film is not changing very much within energy interval of 1.85-2.55 eV. A similar behavior is noticed for the HPC composites containing 0.5-2%BT filler. At 2.1 eV, the optical conductivity increases with an order of magnitude from 10 8 for HPC to 10 9 for HPC/ BT 2%, which is smaller than that of PVA/BT (Barzic et al. 2020). The enhancement of optical conductivity could be attributed to the formation of new states within the band gap that enable electron crossing between the valence to the nearest state, as observed in the lowering of optical band gap (Table 1).
The electrical conduction properties of HPC and HPC/BT composites were also analyzed at low frequencies. As seen in Fig. 9, the measured conductivity (r) of all examined films is affected by the applied electric field frequency, so that it ranges from 10 -11 to 10 -7 S/cm. The conductivity ascended with the increase of frequency. A closer analysis shows that in the lower frequency interval of 1-10 3 Hz, the variation of r was gradual, whereas after 10 3 Hz it ranged faster with further augmentation of the frequency. The recorded conductivity and the phase angle (h) versus frequency are represented in Fig. 9a, b. The plateau region of conductivity retrieved at low frequencies connected with the values close to 0°of phase angles are assigned with the DC conductivity (e.g. transport of charge carriers through the polymer matrix). On the other hand, the linear increase of conductivity at high frequencies and phase angle values close to -90°are generally specific to AC conductivity. Such behavior can be explained on the basis of Dyre's random free energy barrier approach (Dyre 1988). The latter implies that conduction in solids occurs by bouncing of the charge carriers in the localized and highest states in the conduction band. This phenomenon is hastened by applying higher frequencies (Hassan et al. 2018;Tao et al. 2020).
On the other hand, the conductivity of HPC material is affected by the addition of the ceramic filler. The increase in conduction features is more obvious up to 10 3 Hz in the plateau region of conductivity, whereas after this point the slope of the conductivity curve is slightly increasing with addition of BT particles in the polymer. In other words, the increase of conductivity in high frequency zone takes place owing to higher amounts of free charges as a result of more polymer-particle interactions, combined with the existence of BT inside the polymer, which is known to promote the charge conduction (Morsi et al. 2019). Relatively constant conductivity in low frequency range might be ascribed to polar functional groups from HPC matrix, while for composites, this plateau results from electronic or ionic conductivity. As the filler amount in the system increases, the recorded conductivity increase could be attributed to carrier hopping among the macromolecular chains (Mendes et al. 2012). The level of conductivity at the highest frequency slightly exceeds that reported by Morsi et al. (2019) for BT incorporated in PEO/CMC at low frequencies.
Dielectric constant and electric energy density Dielectric spectroscopy was also used to test the dielectric performance of the samples. Figure 10a, b illustrate the double-logarithmic dependence of the real (e') and imaginary (e'') dielectric constant on frequency (f) for neat HPC and HPC/BT composite films. The prepared film of HPC matrix has a permittivity of 3.4-5.2 in the recorded frequency domain. This result is in agreement with that reported by Abdelaziz (2015). However, the obtained e' magnitude for HPC is slightly smaller than that measured by Shinouda and Abdel Moteleb (2005) (e' * 7.6) or by Manaila Maximean et al. (2018) (e' * 19.9). The differences could reside from the degree of sample drying, synthesis conditions (influencing the degree of substitution) or polymer processing (film, powder, fiber). It is evident that e' lowers with increasing frequency over the interval of 1-10 6 Hz. The intensive decline of e', observed at low frequencies for HPC/BT composite samples, is specific to the electrode polarization (EP) process (Samet et al. 2015). This effect may appear due to the agglomeration of mobile charges mainly originated from BT source of charge carriers at the contact between sample and the electrodes used for dielectric measurements. As frequency increases, the EP effect is gradually diminished and a slighter reduction of e' is noticed. The latter behavior reveals the 'pure' dipolar activity of the samples. If the frequency is very high, the dipoles can no longer cope-up with oscillations of the electric field, so the macromolecules are not orienting along the direction of the applied field (Abdelaziz and Ghannam 2010). For this reason, as the frequency increases, the dielectric constant gradually becomes smaller and is almost constant at high frequencies.
Generally, for reinforced polymers, the dielectric properties are linked to the filler content and particle distribution (Morsi et al. 2019). As one may see in   Fig. 10a, the real part of the dielectric constant is enhanced upon BT addition within HPC matrix. The permittivity values of HPC/BT samples below 10 Hz were noticed to be high. At 1 kHz, the measured dielectric constant ranges from 4.2 for neat HPC up to 8.5 for HPC/BT 2% (see the inset graph from Fig. 10a).
As observed in Fig. 10b, the dielectric loss is also higher with further addition of BT particles. This increase reveals that the charge carrying capacity of the prepared samples is notably larger with respect to the unfilled HPC film. In addition, the high value of the dielectric loss, noted at low frequencies (under 1 kHz), could be attributable to the mobile charges from the cellulosic matrix. The linear decrease of e'' retrieved at low frequencies is connected with the r DC plateau region from Fig. 9a and may be attributed to mobile charge carriers from the material.
On the other hand, the low signal from high frequencies is connected with the linear increase of conductivity and attributed to the loss of energy due to polarization of chemical dipoles. Similar dielectric behavior was reported for other polymers filled with BT ceramic material (Yao et al. 2018;Ahmed et al. 2018;Beena and Jayanna 2019). However, at 1 kHz and 2% BT in HPC, the prepared composites have slightly smaller values for e' and e'' than those found in literature for PVA/BT (Beena and Jayanna 2019), polypyrrole/BT (Ahmed et al. 2018), while the permittivity of our samples is higher than that reported for polypropylene/BT (Yao et al. 2018) and PEO/ CMC/BT (Morsi et al. 2019).
Dielectric properties affect the level of energy density within the prepared materials. From practical point of view, this aspect is essential for energy storage applications. The electric energy density (U) for a dielectric is described by Eq. (10): where e 0 is permittivity of free space, e r is the relative permittivity and E is the intensity of the electric field. It is obvious that the discharged energy density of a dielectric can be amended by enhancing either the permittivity of the material or the field intensity. A close analysis of Eq. (10) indicates that the effect of field's intensity has a greater impact on U magnitude because it appears in relation in the second order. The highest energy density is obtained at the point of material's breakdown strength. Reinforced polymers have received special attention owing to the possibility to control the electrical characteristics in regard to the properties of single phase materials (Huang and Jiang 2015), whilst changing the mechanical properties. Ferroelectric particles, such as BT, are ideal for improving the dielectric performance of the polymers owing to their high permittivity.
According to Stark and Garton model (Stark and Garton 1955), the dielectric breakdown strength ranges proportionally with yields stress and is the reverse of the materials permittivity. The yield stress for HPC was reported to be 3.5 MPa (Borges et al. 2004). Based on these aspects, for HPC matrix the breakdown strength is 4.367 9 10 8 V/m. Literature shows that addition of BT particles in a polymer determines lowering of the breakdown strength (Dou et al. 2009;Yu et al. 2013). Knowing that the dielectric breakdown of BT is 38 9 10 5 V/m (Branwood et al. 1962), the corresponding values for the studied composites were estimated and depicted in Fig. 11. In order to determine the maximum values of the electric energy density, this parameter was assessed at electric field intensities corresponding to the breakdown strength. Figure 11 shows the variation of electric energy density with BT amount introduced in HPC. Similar values to our results were reported for polyimide/ reduced BT and poly(vinyl alcohol)/BT composites (Yue et al. 2019;Liu et al. 2020). The increase of U parameter with further addition of ceramic inclusions in the cellulose derivative matrix is a promising indicative for energy storage applications.

Conclusions
New polymer composites were fabricated by embedding ceramic particles in a cellulose derivative matrix. Addition of BT filler favor production of supplementary localized energy states in the forbidden energy zone, thus reducing transparency of films and direct band gap energy from 6.12 eV for neat HPC to 5.74 eV for HPC/BT 2%. Optical micrographs indicate a good dispersion of ceramic particles within HPC material owing to ultrasonication of the samples. Also, the presence of BT inclusions increases polarizability and implicitly the refractive index from 1.533 to 1.605 at 486 nm. Moreover, optical and AC conductivity of HPC is improved after reinforcement, together with dielectric constant that increases from 4.2 for pristine HPC up to 8.5 for HPC/BT 2%. This is reflected in a higher electric energy density and recommends the studied materials for energy storage applications.

Declarations
Conflict of interest The authors declare no conflict of interests.
Consent to participate All authors consent to participate at the research included in this work and approve the manuscript submission.
Consent for publication All authors agree with article publication.