Bacterial cellulose (BC) exhibits several outstanding characteristics compared to conventional cellulose, including high purity, high crystallinity, and an exceptional holding capacity to retain hydrogen-bonding substances [1, 2, 3, 4, 5, 6, 7, 8]. In comparison with plant-derived cellulose, BC typically possesses higher purity and can be synthesized more readily. For plant-derived cellulose, the processing procedure is more complex, as it requires the removal of numerous impurities and the elimination of the amorphous phase. Commonly, BC is known as a stabilizer and a fat replacer in making foods [5], or as an ingredient in biomedical applications [6, 7]. Another less-addressed aspect is that BC also plays a crucial role as a green polymer for the fabrication of flexible devices [8, 9], in which the constituent materials predominantly exist in the form of BC-based composites. This poses a challenge for researchers, requiring continuous improvement of material properties to meet desired future functionalities.

Structurally, BC contains interconnected nanochannels formed by entangled nanofibers, creating a porous 3D network that facilitates molecular diffusion. As a result, substances can be embedded into cellulose nanochannels and confined at the nanoscale. Due to size effects and interactions with cellulose channel walls, useful anomalies may appear. For instance, in studies [10, 11], the authors used ferroelectrics of NaNO₂ and KIO₃ to confine within cellulose nanochannels. The results revealed that the phase transition in NaNO₂ was shifted toward lower temperatures due to the size effects created by the enhanced mobility of ions at the nanoscale [10]. Conversely, once these interactions became dominant, the ferroelectric structure was reinforced, leading to an increase in the phase transition temperature [11]. The detected anomalies can be highly valuable for practical applications.

Potassium dihydrogen phosphate (KDP, KH₂PO₄) is an inorganic ferroelectric widely employed in various advanced technologies, where it is typically utilized in the conventional crystalline form [12, 13, 14]. The phase transition point of KDP from polar to non-polar phases is relatively low (T_c = 123 K) [15], and therefore its applications are mostly focused on the non-polar phase at room temperature. To date, the practical applications of bulk KDP have largely reached a plateau. In this context, downsizing it to the nanoscale is an effective approach that has been increasingly adopted by researchers to enable new functionalities. However, there has been very limited fundamental research addressing the properties of KDP at the nanoscale [16, 17]. Consequently, the nanoscale properties of KDP remain insufficiently understood. This represents a significant gap in both fundamental knowledge and the potential applications of nanosized KDP.

As far as we are concerned, the nanoconfinement of ferroelectrics in BC was reported in literature, where triglycine sulfate was used [18]. However, the ferroelectric material KDP has not been employed to fill in BC yet. The only study on the combination of KDP with cellulose can be found in the work [19] but the used cellulose was not BC. In addition, the KDP component in this composite grew to a size far larger than nanoscale. The present work aims to fill this gap by downsizing the size of KDP to the nanoscale within the nanochannels of BC. Beyond the novelty, the idea of incorporating KDP into BC is particularly intriguing because KDP contains hydrogen bonds and therefore, its interaction with a naturally hydrogen-bonded matrix, such as BC may create interesting anomalous behaviors. Specifically, a comparative study is conducted by simultaneously analyzing the data of structural characteristics, phase transition behavior, relaxation times and activation energies for both pure KDP at normal size and its nanosized form in BC nanochannels. The structural and property-related anomalies are discussed in detail.

The nanosized KDP was obtained by infiltrating it into cellulose nanochannels. The cellulose-containing nanochannels with diameters ranging from 40 to 80 nm were biosynthesized by the bacteria Acetobacter xylinum following a common procedure [20]. The pure KDP was purchased from Sigma-Aldrich and used as received without further purification. However, for this study, both KDP and Acetobacter xylinum cellulose (AxBC) were carefully characterized before preparation of the AxBC/KDP composite using scanning electron microscopy (SEM), X-ray diffraction (XRD) and Fourier-transform infrared spectroscopy (FTIR) techniques.

Because the as-prepared AxBC was stored in distilled water, it was gradually solvent-exchanged into acetone prior to the filling process to achieve more stable rafting porosity and to minimize the shrinkage of the cellulose nanochannels. Subsequently, a room-temperature saturated KDP solution was dropwise added to the AxBC/acetone mixture. The resulting AxBC/acetone/KDP/water suspension was then placed in a vacuum chamber at a low pressure of 100 kPa until acetone bubbles ceased to form. Finally, natural evaporation yielded the AxBC/KDP composite. After partial drying, the AxBC/KDP samples were immersed and gently agitated in ethanol for approximately 10 seconds and subsequently blown with nitrogen gas to remove residual KDP adhering to the surface. The obtained samples were dried at 105 °C for 1 h to remove excess moisture. To determine the filling level, the samples were immersed in boiling water for 4 h to gradually remove KDP from the BC matrix. The recovered BC was then completely dried and weighed again, and the mass was compared with that of the composite sample. The results indicated that the filling loading ranged from approximately 36–55 wt%. The sample selected for the present study corresponds to a loading of 55 wt%. After preparation, the samples were stored in a humidity-controlled chamber. For electrophysical parameter measurements, the samples were cut into square pieces with lateral dimensions of 3 $\times$ 3 mm². The thickness depends on the specific sample and ranges from 0.5 to 0.7 mm. Electrodes were applied using a conductive silver paste. To confirm the existence of anomalies, each experiment was repeated three times for verification.

The morphology was captured on a field emission scanning electron microscope (FE-SEM S-4800, Hitachi High-Tech Corporation, Tokyo, Japan). The crystalline structure and functional groups of materials were explored by an X-ray diffractometer (Rigaku Ultima IV, Rigaku Corporation, Tokyo, Japan) and an infrared spectrometer (Bruker Tensor 37, Bruker Corp., MA, USA). For the testing phase transition, a combined system consisting of an inductance-capacitance-resistance (LCR) meter (GW Instek LCR-821, Accusource Electronics Inc., Newark, DE, USA) with a cooling module using liquid nitrogen was employed. This module was also used to connect an impedance gain/phase analyzer (Solartron-1260A, Ametek Co., Ltd., Berwyn, PA, USA) to obtain frequency dependences of the permittivity at different temperatures. The experimental results shown in this study were measured at least three times with an error not exceeding 0.1%.

The morphology of AxBC and the AxBC/KDP samples is shown in Fig. 1. As described above, the residual KDP was removed from the surface of BC after filling and drying. Therefore, the SEM images do not reveal the confined KDP within the BC nanotubes. Due to the softness of cellulose, it is nearly impossible to capture images clearly showing the nanotube openings. For this reason, as reported in the preparation process, the KDP filling percentage was determined by comparing the mass of BC after KDP removal with that of the initial composite sample. What we can see in the surface pattern is the nanochannel shape, the diameter of which is estimated to be in the range of approximately 40–80 nm. Importantly, based on the mass difference and the characteristic phase transition for KDP as shown below in this study, we believe that KDP has been successfully incorporated into the nanotubes.

Fig. 1.

SEM images for the AxBC (a) and AxBC/KDP (b) samples. SEM, scanning electron microscopy; AxBC, Acetobacter xylinum cellulose; KDP, potassium dihydrogen phosphate. Scale bar = 150 nm.

The results for crystalline structures and functional groups of KDP, AxBC and the composite of KDP confined in AxBC at 250 K are shown in Fig. 2. As shown in Fig. 2a, the pure KDP contains characteristic XRD peaks, including 2$\theta$ = 17.6° (101), 24.2° (200), 29.9° (211), 30.9° (112), 34.3° (220), 35.4° (202), 38.7° (301), 40.9° (103), 46.2° (321) and 46.9° (312), which are consistent with JCPDS ICDD card No. 84-0520. Besides, FTIR characteristic functional groups were also detected (Fig. 2b), such as C-H stretching (2950 and 2863 cm^-1), O=P-OH (2433 cm^-1), P=O stretching (1644 cm^-1), [H₂PO₄]^– (1308 cm^-1) and O-P-O (1082 and 889 cm^-1) [21]. Meanwhile, the characterization results for crystalline structure of AxBC showed the presence of typical cellulose I structure with 2$\theta$ = 14.5° (101), 16.6° (101^–), and 22.6° (002); along with functional groups of C-H (2931 cm^-1), -COOH (1625 cm^-1), -OH in-plane bending (1335 cm^-1) and C-OH (1027 cm^-1) [20].

Fig. 2.

XRD (a) and FTIR (b) data for samples of KDP, AxBC and nanosized KDP in AxBC at the temperature of 250 K. FTIR, Fourier-transform infrared spectroscopy; XRD, X-ray diffraction.

Let us consider the structural changes in samples of KDP-filled in AxBC as compared to those of the pure AxBC and KDP. Based on the results obtained, several significant shifts of the AxBC and KDP components in the composite medium were detected in both XRD pattern (Fig. 2a) and FTIR spectrum (Fig. 2b). Firstly, under nanoconfinement conditions in AxBC nanochannels, two XRD peaks of KDP including 24.2° (200) and 30.9° (112) were found to move to 24.0° and 30.8°, respectively. This indicates that the crystalline structure of KDP was affected at the nanoscale. Meanwhile, the FTIR spectra of KDP-filled samples also demonstrated overlapping and movement. For instance, the peak of 1308 cm^-1 ([H₂PO₄]^–, KDP) overlapped with 1335 cm^-1 (-OH in-plane bending, AxBC), moved to 1346 cm^-1, and became deeper at the new position. Besides, the peak of 1632 cm^-1 (P=O, KDP) overlapped with 1625 cm^-1 (-COOH, AxBC) and moved to a new position of 1635 cm^-1. Importantly, the valley of 3500–2800 cm^-1 created by hydrogen bonds and possible residual water became deeper and wider in the case of AxBC/KDP composite. This anomaly was observed for all synthesized composite samples, proving the formation of new hydrogen bonds between cellulose and nanosized KDP. Other XRD and FTIR peaks of AxBC and KDP components retained their positions in the composite state.

At the nanoscale, an anomalous transition from polar to non-polar phases of KDP was also detected when analyzing temperature-dependent permittivity $\epsilon$(T). Indeed, as shown in Fig. 3, both the pure KDP and AxBC/KDP samples exhibit the presence of a peak at 123 K and 136.8 K, respectively. It is known [15] that the point of 123 K stands for the transition from polar to non-polar phases in KDP at normal size. Thus, the anomalous peak at 136.8 K, i.e by 13.8 K higher than those of the bulk KDP, seems to be created by the nanosized KDP in cellulose nanochannels. This is reasonable because the dependence of $\epsilon$(T) for AxBC did not reveal any peaks (Fig. 3). Moreover, the movement of the phase transition point as observed for the nanosized KDP in this case has also been previously reported in various other materials at the nanoscale [22, 23, 24]. Even more convincingly, theoretically, the phase-transition peak of the nanosized KDP, like the KDP at the conventional size, can still be described by the Curie–Weiss law [25, 26] as $\epsilon$ ~ 1/$|$T – T_c$\mathrm{|}$, where T is the temperature, and T_c is the phase transition point. The detected increase in T_c for the nanosized KDP filled in AxBC will be discussed later in this study.

Fig. 3.

Temperature dependences of the real part of permittivity of AxBC, pure KDP and nanosized KDP in the AxBC matrix.

To further clarify the anomalies observed in the nanosized KDP, the relaxation times and activation energies were also determined. To achieve this, the pure KDP and AxBC/KDP samples were tested for the real $\epsilon$′ and imaginary $\epsilon$″ permittivity under an alternating electric field at various frequencies f. The corresponding results are presented in Fig. 4 and Fig. 5. The calculation results based on the obtained data are shown in Fig. 6. The measurement frequency was smoothly varied from 1 kHz to 100 MHz at different temperatures, which were taken by equal intervals relative to the phase-transition temperature T_c, extending to both sides within the polar and non-polar phases. The relaxation $\tau$ was calculated by $\tau$ = 1/2$\pi$f_r, where f_r is the relaxation frequency determined at the peaks of $\epsilon$″(f). Obviously, the dependences of relaxation times on the reciprocal temperature are in good agreement with the Arrhenius law [27]:

(1)$$\tau ={\tau }_0{e}^{\frac{E_a}{kT}}$$

where $\tau$₀ is a constant, E_a is the activation energy and k is the Boltzmann constant. According to the obtained results, the frequency dependences of $\epsilon$′(f) and imaginary $\epsilon$″(f) for both pure KDP (Fig. 4) and KDP-filled in AxBC (Fig. 5) exhibit a similar behavior. Specifically, the values of $\epsilon$′ decrease with increasing frequency, accompanied by the appearance of peaks in $\epsilon$″(f) corresponding to the relaxation frequency f_r. This similarity is further reflected in the variation trends of the activation energy and relaxation times. For pure KDP, when transitioning from the polar to the non-polar phases, the activation energy decreases from 0.025 eV to 0.019 eV, accompanied by a reduction in relaxation times as the temperature increases (Fig. 6). At the same time, the activation energy of the nanosized KDP in AxBC decreases from 0.035 to 0.028 eV. However, an important difference to note here is that at the nanoscale, both the activation energy and the relaxation times of the nanosized KDP are higher than those of the pure KDP at the normal size in both the polar and non-polar phases (Fig. 6).

Fig. 4.

Frequency spectra of the real and imaginary permittivity for the pure KDP at different temperatures in polar (a,b) and non-polar (c,d) phases.

Fig. 5.

Frequency spectra of the real and imaginary permittivity for the nanosized KDP in AxBC at different temperatures in polar (a,b) and non-polar (c,d) phases.

Fig. 6.

Dependences of relaxation times on the reciprocal temperature for the pure and the nanosized KDP in polar and non-polar phases.

To clarify the relationship between the structural changes and the above-mentioned anomalies, we analyzed the FTIR spectra of the pure KDP, AxBC, and AxBC/KDP samples at different temperatures decreasing from 250 K to 110 K (Fig. 7). The description of the adsorption peaks and corresponding functional groups is provided in Table 1. Obviously, when cooling down samples, the changes of FTIR spectra for the pure KDP and AxBC were insignificant. In this process, the -OH intra-/inter-chain hydrogen bonds and -COOH in AxBC, along with P=O and [H₂SO₄]^– in the pure KDP were found to be slightly shifted (Table 1). Besides, their valleys located in 3500–2800 cm^-1 originated from hydrogen bonds and possible residual water molecules became slightly deeper and wider (Fig. 7). Meanwhile, in the case of the nanosized KDP in AxBC, the observed valleys became strongly deeper and wider upon cooling (Fig. 7). Importantly, the described shifts were observed to move significantly toward higher wavenumbers (Table 1). Interestingly, a strong interaction between the KDP and AxBC components in the composite was detected, including P=O (KDP) and -COOH (AxBC); [H₂SO₄]^– (KDP) and -OH in-plane bending (AxBC); -OH intra- and inter-chain hydrogen bonds (AxBC) and hydrogen bonds (KDP). In other words, the hydrogen bonds formed between KDP and AxBC played a leading role in their interactions.

Fig. 7.

Changes of FTIR patterns with temperature for the pure KDP (a), AxBC (b) and nanosized KDP in AxBC (c).

From the above analyses, the relationship between the structural modifications and the anomalous properties observed for the nanosized KDP under nanoconfinement conditions became evident. The first question to address here is why introducing KDP into the cellulose nanochannels led to an increase in the phase-transition temperature (Fig. 3). According to Landau-Ginzburg-Devonshire theory [28], the transition point T_c increased when the material structure became more rigidly reinforced, which in turn enhanced the viscosity $\eta$ of the material. Indeed, the relaxation frequency f_m is related to the viscosity $\eta$ as f_m ~ 1/$\eta$. Comparing the results shown in Figs. 4,5, the relaxation frequency f_m of the nanosized KDP was lower, i.e., the viscosity $\eta$increased. The question is why was the viscosity enhanced? Since both KDP and BC contain hydrogen bonds, the most plausible assumption is the interaction at the nanochannel walls. Moreover, at relatively low temperatures in our case, hydrogen bonding typically plays a crucial role in KDP [29]. Consequently, the domain walls in ferroelectrics such as KDP became more difficult to reorient. As a result, the polar phase could only transform into the non-polar phase at a higher temperature, leading to an increase in the phase-transition temperature T_c to 136.8 K (Fig. 3). The strong AxBC/KDP interactions through hydrogen bonds might also be responsible for the rise of activation energy and relaxation times as observed above. Logically, once the viscosity increased, the domain walls took more time to switch in a cycle and therefore, this required more thermal energy to activate.

The comparative analyses of experimental data for the pure KDP and AxBC/KDP composite revealed an increase in phase transition temperature from 123 K to 136.8 K for nanosized KDP-filled in nanochannels of Acetobacter xylinum cellulose, as determined in the temperature dependences of permittivity. At the same time, the rise of activation energies from 0.025 eV to 0.035 eV in the polar phase, and 0.019 to 0.028 eV in the non-polar phase was detected, accompanied by longer relaxation times inferred from the shift of relaxation frequencies. Those anomalies were related to the interactions between KDP and cellulose components through hydrogen bonds at low temperatures, resulting in the enhancement of viscosity and, therefore consolidated KDP structures under thermal effects. This assumption was also evidenced by the shift of the functional groups in the FTIR spectra, which directly contributed to hydrogen bonds. Despite the limitations associated with the natural softness of cellulose, causing difficulties in filling and the random distribution of cellulose nanochannels after drying, the obtained results may be useful for researchers in the field of materials manufacturing at low temperatures, where hydrogen bonding plays a more dominant role in stabilizing hydrogen-bonded networks.

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

BDM planned, conducted experiments, analyzed the experimental data and prepared the manuscript. The author read and approved the final manuscript. The author has participated sufficiently in the work and agreed to be accountable for all aspects of the work.

Not applicable.

I would like to express my gratitude to all those who helped us during the writing of this manuscript. Thanks to all the peer reviewers for their opinions and suggestions.

This research received no external funding.

The author declares no conflict of interest.