2019 Volume 60 Issue 12 Pages 2499-2505
In present work, a comparative study on electrophysical properties of two ferroelectric nanocomposites based on cellulose nanoparticles (CNP) mixed with Rochelle salt (RS) and sodium nitrite (NaNO2) at different composition mass ratios was conducted to clarify the influence of cellulose on dielectric properties of primitive ferroelectrics. The composites were carefully characterized XRD and FTIR techniques. The experiments were carried out in a wide temperature range under a weak electric field with an amplitude of 2 V/cm at 1 kHz. The obtained results revealed that at relatively high cellulose content, the phase transition temperature of Rochelle salt in the composite increased, while for sodium nitrite – decreased as compared to those of volume ferroelectrics. Besides, the higher the cellulose content was, the stronger the shift of phase transition point was observed. At small cellulose content, the Curie point characteristic for volume ferroelectrics occurred. The intermolecular hydrogen bonds and size effects were supposed to be responsible for the observed anomalies.
Fig. 4 Temperature dependences of dielectric constant and dielectric loss tangent for (a), (b) CNP/RS and (c), (d) CNP/NaNO2 composites at different composition mass ratios. The results of cellulose were also added for comparison. Tdec – decomposition temperature of RS. The insets – for polycrystalline RS (a), (b) and NaNO2 (c), (d).
The development of technology cannot be separated from the exploration of new materials. In the context, reducing of material size from normal view 3D to 2D and even nanosize is a breakthrough step in technological progress.1–3) Nanocomposites are such advance materials due to the combination of advantages from their components.4–7)
Ferroelectric nanocomposites are multifunctional materials with several advantages over primitive ferroelectrics due to anomalous properties created by the heterogeneity of composite system and size effects of components at nanoscale size.4,7) The study on properties of ferroelectric nanocomposites is important for both fundamental scientific interest and technological applications.
Cellulose is not a new material in electronics. Recently, cellulose as well as other eco-friendly materials have been encouraged to use in electronics industry in the context of ever-growing amount of toxic e-waste discharged into eco-systems. Because of advantages as low cost, light weight, high electrical stability, flexibility and biodegradability, cellulose was utilized as the main material for processing stable and highly flexible optoelectronic devices,8) or an ideal support material for preparing high performance transistors,9,10) photovoltaic cells11) and OLEDs.12) In this regard, development of new based-cellulose electronics materials has attracted great attention from researchers.
As reported in previous studies,13–21) nanocomposites based on cellulose and various ferroelectric inclusions were synthesized. The presence of cellulose in the composites led to the appearance of several promising anomalous properties. For instance, the introduction of ferroelectrics into nanosized pores of cellulose resulted in the shift of Curie point toward higher temperatures,13,19) the reduction of relaxation frequency and strong dielectric dispersion at low frequencies21) as well as the increase in dielectric characteristic under the influence of humidity.16) In addition, the mechanical strength of composites can be adjusted by changing the cellulose content in ferroelectric nanocomposites.22)
Up to now, there has been a few studies devoted to preparation of cellulose-based ferroelectric materials and therefore the whole picture is not explored completely with several drawbacks and omission that should be addressed: (i) in previous studies,13–22) cellulose in the form of porous matrices with nanochannels into which ferroelectrics were embedded was used to prepare composites. These composites had a huge drawback related to difficulties in determining the ferroelectric content. Thus, the influence of composition mass ratios on properties of primitive ferroelectrics was not reported; (ii) authors considered mainly effects of cellulose on properties of ferroelectrics which contains hydrogen bonds as cellulose. The influence of cellulose on properties of non-hydrogen-containing ferroelectrics has not been investigated yet. In the present work, a comparative study was conducted by synthesizing two composites based on cellulose nanoparticles (CNP) mixed with hydrogen-containing Rochelle salt (RS) and non-hydrogen-containing sodium nitrite (NaNO2) at different composition mass ratios for investigating the influence of cellulose on dielectric properties of ferroelectrics. Rochelle salt (RS) is a classical ferroelectric having two Curie points at −18°C and 24°C with the occurrence of ferroelectric phase between them,23) while sodium nitrite (NaNO2) undergoes transition from ferroelectric to paraelectric phase at 164°C.24) The results are thoroughly discussed.
The cellulose was prepared from cotton waste according to the procedure described in the work.27) After freeze drying, the near spherical shape of particles was formed with the size of 40–80 nm as shown in Fig. 1(a) and Fig. 1(b). The size of nanoparticles and their morphology were examined by a Zetasizer analyzer with a detection range of 0.3 nm–10 µm and a FE-SEM S4800 HITACHI scanning electron microscope with an accelerating voltage of 10.0 kV, respectively. The Rochelle salt and sodium nitrite were reagent-grade, purchased from Merck supplier and utilized in preparation process without further purification. The reliability of starting materials (CNP, RS and NaNO2) was confirmed by XRD (Fig. 1(c)) and FTIR (Fig. 1(d)) results in the present work. For cellulose nanoparticles, three XRD typical peaks for Iβ crystalline phase at 2θ = 14.7°, 16.3° and 22.5° corresponding to (101), $(10\bar{1})$ and (002) crystallographic planes25,26) were detected (Fig. 1(c)). Besides, FTIR pattern of cellulose (Fig. 1(d)) contains all characteristic peaks and bands as 3273; 2900 (2840), 1433 (1378) and 1635 (707) cm−1 attributed to O–H stretching,28) CH2 asymmetric vibrations, OCH in-plane vibrations27) and OH out of plane bending.29) XRD characteristic peaks obtained for RS and NaNO2 are in good agreement with the ICPDS cards no 04-0836 and 03-0628, respectively (Fig. 1(c)). For RS, the band of 3500–3000 cm−1 for assigned to OH stretching30) along with several adsorption peaks at 2930, 1595, 1350 and 1120 cm−1 respectively characterized by CH stretching,31) C=O stretching, δ(OH) vibrations32) and C–C stretching vibrations33) were clearly revealed (Fig. 1(d)). In the case of NaNO2, FTIR pattern shows characteristic peaks at 1241 cm−1 (NO2 bending) and 822 cm−1 (asymmetric stretching mode).34) The above obtained results indicated adequacy of the used reagents for preparation of the composite.
(a) SEM image and (b) size distribution of cellulose nanoparticles; (c) XRD patterns and (d) FTIR spectra for starting materials (cellulose, polycrystalline Rochelle salt, polycrystalline NaNO2).
For preparation of the composites, saturated solutions obtained by dissolving RS or NaNO2 in distilled water at 25°C were prepared. Then, a determined amount of RS solution was taken out to mix with cellulose at different CNP:RS mass ratios (0.2:1, 1:1, 3:1, 5:1, 7:1). Similar procedure for NaNO2 solution was carried out to get mixtures with CNP:NaNO2 mass ratios of 0.2:1, 1:1, 3:1, 5:1, 7:1 and 9:1. Temperature was stabilized to be slightly higher than 25°C to maintain saturation level of the solutions. Each of mixtures was stirred in a closed bottle for 8 h. Next, the bottle was partially opened to slowly evaporate residual water at room temperature (20°C). After about 3 days, a solid part was obtained, freeze dried before crushing in motar and then compressed into tablets of 6 mm in diameter and 1 mm in thickness. The silver electrodes were applied on the large surfaces of composite samples.
The information of crystalline structure and functional groups for samples were tested by using a Rigaku Ultima IV X-ray diffractometer and a Bruker Tensor 37 spectrophotometer (USA), respectively. The phase transition in the composite was measured on a model GW Instek LCR-821 meter at 1 kHz. The temperature for all experiments was stabilized with an accuracy of 0.1 K. The setup for measuring the third harmonic coefficient for the composite included a harmonic oscillator with an operating frequency of 1 kHz. The signal was taken from a resistor connected in series with the sample, then fed to a digital spectrum analyzer consisting of a computer with a 24-bit ZET 230 analog-digital converter and ZETView software. During the measurement amplitudes of the third harmonic and the main signal were recorded. The relative measurement error did not exceed 0.1%.
The results for crystalline structure (Fig. 2) and functional groups (Fig. 3) in CNP/RS and CNP/NaNO2 composites at different composition mass ratios revealed that most of bands/peaks characteristic for composite components were observed, and there was no the shift of XRD and FTIR peaks in comparison with those for their components. However, there were several special points to mention here. Firstly, a strong overlapping was detected. For examples, for CNP/RS composite, the XRD peaks at 2θ = 14.7° (120), 16.2° (210) of RS were overlapped with those at 2θ = 14.7° (101), 16.3° $(10\bar{1})$ of CNP, respectively (Fig. 2(a)). In addition, the overlapping of FTIR adsorption peak couple at 2930 cm−1 (RS) – 2900 cm−1 (CNP), 1120 cm−1 (RS) – 1164 cm−1 (CNP) and 692 cm−1 (RS) – 707 cm−1 (CNP) was also detected. Secondly, the expansion of the region of 3500–3000 cm−1 with increasing the CNP content in the case of CNP/RS composite was observed (Fig. 3(a)). Finally, for both composites, the peak intensity of a component increased with increasing its content in the composite.
XRD patterns for (a) CNP/RS and (b) CNP/NaNO2 composites at different composition mass ratios. The results of cellulose, polycrystalline Rochelle salt and NaNO2 are added for comparison.
FTIR spectra for (a) CNP/RS and (b) CNP/NaNO2 composites at different composition mass ratios. The results of cellulose, polycrystalline Rochelle salt and NaNO2 are added for comparison.
The influence of cellulose nanoparticles on phase transition of RS and NaNO2 inclusions is presented in Fig. 4(a), (b) and Fig. 4(c), (d) respectively. For these experiments, temperature dependences of dielectric constant ε(T) and dielectric loss tangent tgδ(T) for composite samples at different composition mass ratios were measured. In the case of CNP/RS composite, the upper phase transition characteristic for pure RS (Tc = 24°C) (the inset in Fig. 4(a)) was detected except for samples with relatively small RS content (CNP:RS = 7:1) (Fig. 4(a)). Similarly, the peak characteristic for phase transition in bulk NaNO2 (Tc = 164°C) (the inset in Fig. 4(c)) was also observed at CNP:NaNO2 ratios of 0.2:1, 1:1 and 3:1 (Fig. 4(c)). These phase transitions became more blurred with decreasing CNP content. At the same time, with increasing CNP content, a higher-temperature peak in ε(T) for CNP/RS composite appeared, shifted toward higher temperatures and became more pronounced (Fig. 4(a)). It is worth to notice that at relatively low RS i.e. high CNP content (the curve for samples with CNP:RS = 7:1 in Fig. 4(a)), dielectric constant constantly increases with temperature rising without any peaks detected up to decomposition temperature of RS (Tdec = 56.5°C35)). The increase in Curie point and disappearance of phase transition peak were reported for RS embedded into nanosized pores of alumina matrices and mixed composites.35,36) For CNP/NaNO2 composite, only one peak of ε(T) was observed, shifted to lower temperatures and then disappeared at high CNP content (Fig. 4(c)). The absence of transition peak upon heating was also reported for some ferroelectric nanocomposites with NaNO2 inclusions.37,38) The similar shape of tgδ(T) as of ε(T) for both composites was obtained with values comparable to literature data.37,38) The nature of all obtained anomalies will be thoroughly discussed in this study.
Temperature dependences of dielectric constant and dielectric loss tangent for (a), (b) CNP/RS and (c), (d) CNP/NaNO2 composites at different composition mass ratios. The results of cellulose were also added for comparison. Tdec – decomposition temperature of RS. The insets – for polycrystalline RS (a), (b) and NaNO2 (c), (d).
In order to get more information about phase transition in the studied composites, nonlinear dielectric measurements were conducted. The results for temperature dependences of the third harmonic coefficient γ3ω for CNP/RS and CNP/NaNO2 composites at different composition mass ratios are shown in Fig. 5. The results for polycrystalline samples of RS and NaNO2 are added for comparison. It can be seen in Fig. 5 that the values of γ3ω reach minimum at phase transition point. For CNP/RS composite with CNP:RS ratio of 7:1, the monotonic increase of γ3ω(T) was observed in the entire temperature range up to decomposition point of RS as detected above in ε(T) (Fig. 5(a)). This allows us to conclude that the phase transition did not occur in CNP/RS (7:1) samples. Meanwhile, although the transition peaks were absent for CNP/NaNO2 composite at CNP:NaNO2 ratios of 7:1 and 9:1 in above ε(T), minima in γ3ω(T) appeared at 152 and 141°C i.e. the phase transition point continue shifting toward lower temperatures with rising cellulose concentration as mentioned above for other composition mass ratios. All phase transition temperatures determined in γ3ω(T) spectra for both composites are listed in Table 1 and Table 2.
Temperature dependences of the third harmonic coefficient for CNP/RS and CNP/NaNO2 composites at different composition mass ratios. The results for polycrystalline RS and NaNO2 are added for comparison.
The presence of Curie point which is characteristic for the volume NaNO2 (Tc = 164°C) as well as the appearance of the upper phase transition of bulk Rochelle salt (Tc = 24°C) are obviously related to the formation of ferroelectric clusters in composite samples at low concentration of cellulose content.
The question about why the second peak at higher temperature was observed in the case of CNP/RS composite, but not detected for CNP/NaNO2 samples can be answered by the existence of interaction between CNP and RS through hydrogen bonds. The interaction led to the fixation of polarization of RS, resulting in the shift of phase transition toward higher temperatures. This assumption is reasonable because both CNP and RS are hydrogen-containing materials. It was proved that the bands of 3500–3000 bands which corresponds to the OH stretching as indicated in FTIR patterns for CNP/RS composite are expanded with increasing CNP content. Known28,39,40) that the change in the number and strength of hydrogen bonds brings the change in intensity and the width of the related bands. Since NaNO2 is non-hydrogen-containing ferroelectric and therefore the CNP/NaNO2 interaction did not occur.
The shift of the higher-temperature phase transition point in CNP/RS composite to higher temperatures with increasing CNP content might be associated with the stronger interaction between CNP and RS components. It can be explained by the fact that the higher the concentration of cellulose was, the higher the isolation level of RS could be and therefore the closer the CNP/RS contact occurred.
The rising values of dielectric constant with the absence of peak in ε(T) for CNP/RS composite at CNP:RS of 7:1 or higher CNP content can be referred to the maintenance of ferroelectric phase in RS inclusion up to decomposition temperature of RS. In order to prove this assumption, additional XRD spectra (Fig. 6(a)) and FTIR patterns (Fig. 6(b)) for CNP/RS composite with CNP:RS mass ratio of 7:1 at different temperatures higher and lower decomposition temperature of Rochelle salt (Tdec = 56.5°C) were taken. At 15°C, the XRD peaks and FTIR bands characteristic for ferroelectric phase of RS component in the composite were observed as mentioned above. At 56°C i.e. close to but lower than Tdec, some XRD and FTIR peaks characteristic for RS crystals disappeared, indicating that the decomposition process started. At 58°C higher than Tdec, almost all XRD and FTIR peaks of RS were absent. The remaining peaks are of cellulose. Importantly, in the FTIR pattern at 56°C (Fig. 6(b)), the two functional groups (C=O at 1575 cm−1 and O–H at 1350 cm−1) which are responsible for ferroelectricity in RS36) started to disappear, proving the maintenance of ferroelectricity in samples with CNP:RS of 7:1 up to decomposition point. It is worth to notice that the decomposition process accompanies breaking hydrogen bonds,36) intensifying the evaporation speed of water from the samples and therefore the band of 3500–3000 cm−1 became narrower (Fig. 6(b)).
XRD spectra (a) and FTIR patterns (b) for CNP/RS composite with CNP:RS mass ratio of 7:1 at different temperatures higher and lower decomposition temperature of Rochelle salt (Tdec = 56.5°C).
The reduction of phase transition temperature in CNP/NaNO2 composite with increasing cellulose content can be related to the contribution of size effects. The influence of size effects was probably significant at high CNP concentration because the NaNO2 could be naturally distributed in the composite as isolated particles with smaller size. Indeed, the Landau phenomenological theory developed for isolated particles also predicted that the temperature of structural phase transition should decrease with decreasing particle size if the order parameter or the value of the exchange integral at the particle boundaries is less than those of the volume samples.41,42) The question here is why this did not happen in the case of CNP/RS composite and reduced its phase transition temperature. In our opinion, although the isolation of RS particles took place, but the contribution of hydrogen bonds to phase transition was greater because the phase transition for CNP/RS composite occurred at low temperatures (<60°C) (Fig. 4(a)). It should be noticed that the sharp increase in dielectric constant and dielectric loss tangent in NaNO2 is associated with the increase in conductivity due to the presence of high-mobility free Na+ ions.43–45) Just these ions create a screening effect for the action of the depolarizing field and therefore determine the real position of the Curie point in the composite. In addition, as indicated in studies,43–45) Na+ ions become more mobile with decreasing NaNO2 particle size and the phase transition occurs faster. This might be responsible for the absence of transition peak in temperature dependence of dielectric constant for CNP/NaNO2 composite at high cellulose concentrations.
In summary, the presence of cellulose in ferroelectric nanocomposites with Rochelle salt and sodium nitrite affects their phase transition by two factors. The first one is related to the interaction through hydrogen bonds that fix polarization and lead to the increase in phase transition temperature of hydrogen-containing Rochelle salt. This influence was not observed in the case of non-hydrogen-containing sodium nitrite. The second factor is size effects which was present for both composites and became significant at high cellulose content. Besides, at relatively high cellulose content, the ferroelectric phase of Rochelle salt maintained up to its decomposition temperature, and the characteristic peak for phase transition in sodium nitrite was absent. The obtained results are important for practical applications by means of bringing an effective method for adjusting properties of ferroelectric materials.
