The ability to encapsulate the liquid crystal (LC) molecules within a polymer matrix, as seen in the polymer-dispersed liquid crystal (PDLC) as well as in the polymer-stabilized liquid crystal (PSLC) systems and their multi-stimuli responsiveness, provides insightful techniques for innovative electro-optical applications^[1–4]. These systems leverage the inherent anisotropic properties of LCs, including the refractive index, permittivity, and other physical parameters. Depending on the initial molecular orientation within the polymer domains and the conditions for refractive index matching between the LC and the polymer, optical properties, such as transmittance and haze, can be manipulated upon application of an electric field. In typical LC-based polymeric systems, the anchoring strength of the LC molecules within the micropores of the polymer network is comparatively greater, which requires a higher electric field to reorient the LC domains, endowing a high-power consumption, and the mechanical strength as well as durability also suffer^[5–7]. Striking a balance between these properties is crucial for optimizing the performance of electro-optical devices. The optical performance is mainly influenced by the anchoring strength of LC molecules, while the order of interfacial interaction with the polymer network determines the anchoring characteristics, which can be controlled by the use of pre-alignment layers in substrates and/or different types of polymerizable monomers^[8,9]. The coexisting polymer-dispersed and stabilized liquid crystal (PD&SLC) system^[10,11] prepared by integrating both mesogenic and non-mesogenic monomers reveals a lower driving voltage of about a few tens of volts^[10,11], but further reduction is desirable for energy-saving applications. Recent advancements, including the use of various monomers, doping of dye or conductive nano-particles, and optimizing curing conditions^[12,13], have minimized the driving electrical power to some extent. However, the introduction of strongly polar and ferroelectric LC material with fascinating electro-optic properties is expected to offer low driving voltage with faster switching.

Recent discovery of the polar nematic phase, named the ferroelectric nematic (NF) LCs, has attracted widespread attention in soft matter due to their fascinating properties, such as large spontaneous polarization^[14–16], high dielectric constant^[17], strong nonlinear optical response^[18], low driving electric field^[14,19], significant piezoelectricity^[20], microsecond switching speed^[19,21], and novel electro-optical properties, which have been envisaged as potential candidates for next-generation faster-switching devices^[14,22–24]. Compared to conventional nematic LCs, the highly polar molecular architecture of ferroelectric nematic promotes polar ordering, aligning all dipoles along a single direction within each ferroelectric domain and making them substantially more sensitive to the electric field. To date, there have been some reports on the application of this ferroelectric nematic^[25–28], highlighting the effects of a high dielectric constant, considerable second-order optical nonlinearity, and polarization of the NF phase. A recent study on integrating ferroelectric nematic in a nano-PDLC system shows the induction of larger birefringence at lower voltages compared to typical nematic LCs^[29]. Additionally, a microlens device with a polymer-stabilized chiral ferroelectric nematic liquid crystal was fabricated with a low electric field driving capability^[21]. Therefore, considering the distinctive characteristics of the ferroelectric nematic and associated tunability of optical properties offers considerable scope for the development of switchable devices with a low voltage and faster response time.

In this work, we demonstrate a polymer-confined LC film consisting of a high dielectric constant ferroelectric nematic, DIO, and photo-polymerizable monomers with lower concentrations. Under the application of an alternating current (AC) electric field, the sample was polymerized to avoid the potential drop due to the pre-alignment layer on the substrates. The effective polymer network structure randomizes the director orientation of molecules at microscales in inter-domains of continuous network architecture. Consequently, the prepared polymerized sample exhibited a highly scattered state without an electric field and transformed to a transparent state under a low-amplitude electric field. The electro-optical characteristics demonstrate a satisfactory driving voltage (less than 4 V) with a faster switching speed (∼0.6ms) in the NF phase, significantly better than the paraelectric N phase.

In the present study, the NF LC, 2,3’,4’,5’-tetrafluoro-[1,1’-biphenyl]-4-yl 2,6-difluoro-4-(5-propyl-1,3-dioxan-2-yl)benzoate (DIO, Nanjing Shuxin Technology Co., Ltd.), is utilized as a primary compound for the preparation of the precursor. Figure 1(a) shows the chemical structures of the LC material, monomer, and photo-initiator used in the experiments. Two types of reactive monomers, a bi-functional LC monomer, 1,4-bis-[4-(3-acryloyloxypropoxy)benzoyloxy]-2-methylbenzene (RM257, Nanjing Leyao Technology Co., Ltd.) and a non-mesogenic monomer 2-Ethylhexyl Acrylate (EHA, Shanghai Macklin Biochemical Technology Co., Ltd.) were added according to the weight ratio of the DIO, RM257, and EHA at 94:3:3 for polymerization of the system. Generally, the polymer strands of the RM257 are flexible, but the acrylate monomer EHA provides a stiffer network in conjunction with the bifunctional monomers. It has been observed that the approximate ratio of 1:1 for the mono-functional to bi-functional monomers results in lower driving voltage and a higher contrast ratio^[30]. Similarly, an equal amount of RM257 and EHA is selected in the present study. The photoinitiator IRG651(Nanjing Leyao Technology Co., Ltd.) is further added at 2% (mass fraction) of the total weight of all compounds. All the materials were used without any further purification.

The processing steps for the preparation of the sample and fabrication of test cells by electric field-dependent UV polymerization are illustrated in Fig. 1(b). Primarily, the DIO and monomers are added according to the specified weight ratio and mixed with dichloromethane (DCM) to prepare a homogeneous solution. Then, IRG651 is added to the solution and placed on the hot plate at a temperature of 70°C for 5 min. In order to prepare the homogenous precursor, we repeated the process of heating the sample for 2 min followed by vigorous shaking with the aid of a vortex shaker for 1 min. Finally, the solvent was evaporated by maintaining the sample temperature at 65°C for 4 h by opening the lid of the bottle. After complete evaporation of the solvent, the precursor was kept at 120°C for 5 min and repeatedly mixed using a vortex shaker. The LC cells were fabricated by bonding two ITO-coated glass substrates without any pre-alignment layer. The thickness of the cells (∼20μm) was controlled by applying a UV curable glue mixed with silica spheres with a diameter of 20μm±0.5μm. The polymerizable mixture was filled into the cells at a temperature of 120°C using capillary action. Further, the sample was cooled down to the NF phase and stabilized for 5 min. In order to polymerize the sample, a square wave AC electric field with an amplitude of 10 V (peak-to-peak voltage) and a frequency of 1 kHz was applied perpendicularly to the sample and subsequently exposed to UV light (365 nm) at an intensity of ∼30mW/cm2 for 2 min for complete polymerization. Here, the applied AC electric field is applied only to align the polymer networks along the field direction without dispersion or stabilization of LC molecules/phases. Thus, such confinement of molecules using this technique can be called polymer-confined liquid crystal.

The electro-optical parameters, such as voltage-transmittance (V-T) profiles, driving voltages, and response time were determined using a laser light (Model DG70962, Changchun Laser Technology Co., Ltd., wavelength = 532 nm) projecting on the sample at 90° (±2°) to the plane of the cell without any polarizers. The transmitted light intensity was detected using a photodetector (DET025AL/M, Thorlabs, Inc.) at a distance of ∼25cm. The orientation of the LC directors was controlled by applying a square-wave positive- or negative-bias voltage using a waveform generator (DG4162, Rigol Technologies Co., Ltd.) connected to an amplifier (ATA-2041, Xian Aigtek Electronic Technology Co., Ltd.). To measure the response times of the prepared device, an electrical pulse (positive bias) with a width of 5 ms was applied to the cell, and the response was recorded in the oscilloscope (SDS6204H10 Pro, Siglent Technologies Co., Ltd.) connected with the photodetector. According to industry standards, 10%–90% switching time is the characteristic electro-optic response time. The rise time (τrise) and fall time (τfall) are determined as the transmittance changes from 10% to 90% and 90% to 10%, respectively. The contrast ratio is calculated as the ratio between the maximum and minimum transmittances.

During cooling of the sample from the isotropic phase, POM images are captured in unaligned cells with crossed polarizers. Figure 1(c) displays these images obtained at different temperatures. Below the isotropic state, characteristic droplets of the nematic phase are generated at 140°C, which gradually expand and encompass the entire region at 120°C. With further cooling, the birefringent color of uniform texture shifts and ultimately stabilizes around 90°C. However, when the temperature drops below 60°C a transition occurs [Fig. 1(c)(iv)] to the intermediate Nx phase (also termed M2, Ns, or SmZA)^[31,32], followed by transition to the ferroelectric nematic (NF) phase [Fig. 1(c)(v)] at 48°C where the uniform texture is covered by some small, randomly oriented spikes. In a planar-aligned sample, similar spikes appear but align uniformly along the rubbing direction. The sample temperature was maintained at this NF phase (45°C), and the AC electric field at 1 kHz was applied to uniformly orient all molecules perpendicularly to the substrates. Figure 2(a)(i–iii) present the images for applied voltages of 0, 5, and 10 V, respectively. With nearly vertical molecular alignment, as indicated by the dark texture in Fig. 2(a)(iii) at 10 V, a UV light (365 nm, 30mW/cm2) was used for photo-polymerization. Although polymerization occurs quickly (∼20 s) after UV exposure, it was continued for 2 min to ensure complete polymerization throughout the bulk sample. Figure S1 shows the POM images during UV-induced polymerization in the presence of an electric field.

Among the two types of monomers, the liquid crystal monomer (LCM), RM257, tends to align with the electric field, while the non-liquid crystal monomer (NLCM), EHA, is insensitive to the electric field and is arbitrarily oriented within the sample. Consequently, during polymerization, the LCM forms a vertically aligned network structure, whereas the NLCM produces a randomly oriented network structure that influences the vertical alignment. Previous studies on PD&SLC systems using both an LCM and an NLCM^[10,33] report a similar kind of electric field-induced vertically aligned polymer network (VAPN) structure of the LCM within the micropores of the bulky NLCM polymerized network. However, their initial polymerization was conducted without an electric field to create porous microdomains of the NLCM. Additionally, we found that the electric-field-dependent polymerization of the NF LC-based sample composed solely of RM257 LCM resulted in a VAPN structure that exhibits only a transparent state before and after applying a driving electric field. Due to the lack of head-tail symmetry in the NF phase, polar molecules prefer to align their dipoles unidirectionally. However, in our system, the effective multidomain network structure offers random director orientations in different domains. This leads to a mismatch between the refractive index of the polymer fiber and the LC molecules, causing the final polymerized sample to appear scattered. Even after the UV exposure and the electric field are removed, a faint dark texture remains, indicating the formation of a stable polymer network with a multidomain structure. Further observation without the polarizer yields a dark state in the absence of an electric field, as depicted in Fig. 2(a)(iv). This confirms that light scattering occurs due to the arbitrary orientation of the molecular directors in different domains of the polymer network [Fig. 2(c)(i)]. Upon application of an alternating current (AC) electric field with a positive or negative bias, the directors in each domain orient to the electric field direction [Fig. 2(c)(ii)], resulting in a bright state without a polarizer [Figs. 2(b)(i) and 2(b)(iii)]. Once the electric field is removed, the directors return to their random orientations, reverting to the dark state, as represented by the OFF state in Figs. 2(b)(i) and 2(b)(iii). However, applying an electric field with alternating positive and negative cycles causes only insignificant changes in director orientation and modifies the dark state, as shown in Fig. 2(b)(ii). Increasing the amplitude of the alternating electric field produces a bright state, but it is not significantly brighter than when using only positive or negative cycles. This signifies that the molecular reorientation impact on the refractive index matching is pronounced when only positive or negative voltages are applied. Therefore, all measurements were performed on the prepared sample using positive or negative AC voltages.

To understand the effect of the temperature, we investigated the electric field-dependent transmittance of the prepared sample under POM without a polarizer. In the high-temperature paraelectric N phase, optical images exhibit a bright state, and the relative change in transmittance between the ON and OFF states is very minimal. Notably, the N phase exhibits head-to-tail molecular symmetry, and also, no alignment layer is utilized in the fabricated cells. Therefore, the molecules oriented irregularly within each domain align with the applied electric field, showing a transition from a less bright state to a completely bright state. However, during the transition to the NF phase, all the molecular dipoles in each domain try to be aligned in a specific direction due to polar symmetry breaking. Owing to the interfacial interaction inside the polymer network, molecules within each domain experience a hindrance to orient collectively along the same direction, resulting in a less bright state, as shown in Fig. S2. Once an electric field with a positive or negative bias is applied, it triggers all the molecules to align with the electric field direction, exhibiting a brighter or transparent state. After removing the electric field, an interplay between inter-molecular interaction and polymeric interfacial interaction leads to orienting the director of each domain in a preferred direction. This causes a random director distribution across the sample, which scatters the incoming light, resembling an extremely dark state. Once this dark texture is activated by applying either a positive or negative bias, the NF phase consistently shows extremely dark and bright images at the OFF and ON states of the electric field, respectively. In this case, the relative change in the dark and bright states is notably pronounced for the electric field of a positive or negative bias compared to the effect of alternating positive and negative cycles, as shown in Fig. S2.

The V-T curves for the prepared sample under the 10 Hz square wave positive bias voltage are displayed in Figs. 3(a) and 3(b) for both the N and NF phases at different temperatures. The corresponding driving voltages and contrast ratio (CR) are presented in Figs. 3(c) and 3(d). In the high-temperature paraelectric N phase, the OFF state of the sample is poorly scattered, with a transmittance of approximately 21%. This value increases to a maximum of ∼71% at an applied voltage of 24 V. The threshold voltage, Vth (10% of the maximum transmittance), is obtained around ∼0.6V for all temperatures in the N phase. Meanwhile, the saturation voltage, Vsat (90% of the maximum transmittance), decreases from 18.1 V at 100°C to 10.1 V at 60°C [Fig. 3(c)]. It is worth noting that after transitioning to the NF phase, the OFF-state transmittance reaches 1.5%, indicating that the dark state is reasonably good, and when the amplitude of the electric field increases, it approaches 85% at 20 V. Although the values of Vth in the NF phase are higher than those of the N phase, they do not exceed 2 V. Conversely, the value of Vsat lies within a range of 9 to 13 V. Generally, the value of Vth is inversely proportional to Δε (Δε denotes the dielectric anisotropy), and due to the high Δε value in the NF LC, it is expected to have a lower value of Vth. However, the switching mechanisms differ between the N and NF phases. In the NF phase, molecules exhibit spontaneous polar ordering. Due to the breaking of inversion symmetry, all the molecular dipoles try to orient along the electric field. Therefore, the spontaneous ordering of molecules individually or in a group requires long-range polar reorientation, which is strongly opposed by the elastic frustration of the confined polymer network, which is less critical in the N phase. Additionally, the high-polarization NF phase produces polarization charges owing to the reorientation of the dipoles^[16], which effectively shields the external voltage within the bulk LC material. As shown in Fig. 3(d), the value of the CR (ratio of maximum transmittance Tmax to minimum transmittance Tmin) significantly increases to a maximum of 57.5 in the NF phase, compared to the N phase (maximum value ∼3.1). The molecular orientation within the NF phase generates different kinds of characteristic topological defects within the medium due to the self-assembly of highly polar molecules, which scatter the incident light more intensely, leading to a higher CR value than the N phase.

To understand the effect of opposite polarity of the bias field, we investigated the optical response of the sample by applying simultaneous positive and negative bias voltages. Figure 4 depicts the transmittance curve obtained by increasing and decreasing the positive and negative voltages at different frequencies. The minimum transmittance obtained in the N phase (at T=80°C) is 22%, while it increases to 66% at 1 Hz frequency for both the positive and negative cycles. Conversely, in the NF phase, the minimum transmittance of 1.2% approaches a maximum of 79% at a voltage of 20 V and 1 Hz for both positive and negative cycles. It is noticeable that the difference in transmittance during increasing and decreasing the voltage (hysteresis) is negligibly small for both phases, and there is a symmetrical nature of transmittance for both positive and negative voltage cycles. However, the highest contrast ratio in the NF phase is observed for 1 Hz frequency, signifying that the group of molecules within all the domains reorient perfectly to the direction of the electric field at lower frequencies, but the response is slower at higher frequencies.

Figure 5(a) exhibits the response curves at different temperatures for an applied electric field of 10 V (positive bias), and the corresponding measured values of the rise time and fall time are shown in Fig. 5(d) for both N and NF phases. It is observed that apart from the transition, the obtained rise time is more or less equal (∼0.3ms) for both phases, while the fall time in the NF phase is less than 0.1 ms, which is faster than that of the N phase (minimum ∼2ms). Conversely, investigating the effect of the applied voltage on the response time in the N and NF phases [Figs. 5(b) and 5(c)], we see that both the rise time and fall time initially decrease up to 6 volts before saturating [Figs. 5(e) and 5(f)]. The rise time is found equivalent (∼0.27ms) in both the N and NF phases, while the fall time is much faster in the NF phase (0.6 ms) than that of the N phase (4.2 ms). This observation implies that the molecules embedded within the polymer network structure recover their initial orientation very slowly in the N phase, but due to the intriguing assembly of molecules in the NF phase, they return quickly. However, the response time for both the saturated rise time and fall time is faster and closer to each other in the NF phase than that of the N phase, which is sufficient for practical applications.

Based on the obtained results, the prepared sample exhibits outstanding electro-optic properties with the best balance of driving voltage and CR. Therefore, we have also examined the performance of the device as a smart window. Figures 6(a) and 6(b) show the photographs of the prepared cell of thickness 20 µm at different electric field conditions. For both near and far objects, the reported device represents a satisfactory haze at the OFF state and transforms to transparent upon applying the electric field (∼4V). Such a low-voltage working device will consume minimal electrical power, which is beneficial for smart windows. A similar film prepared using a conventional 5CB LC requires a higher electric field to operate, exhibiting the advantage of NF LC. Typical commercial LC-polymeric devices require a driving voltage of ∼40V, 1–10 ms response time, and within a range of mW/cm2 power. In contrast, our system offers lower voltage and a faster response, highlighting its superior characteristics. Moreover, we have tested the stability of the sample at 40°C by monitoring the transmittance of the scattered and transparent state over 3000 cycles by applying a 10 V square wave electric field with a frequency of 1 Hz. From Fig. 6(c), we can see that the transmittance at the ON state remains stable at 82% and almost 1.5% at the OFF state. This indicates that the strength of the polymer network favors the switching of molecules for a long time. However, we observed that the crystallization of the samples occurs at 35°C after 1 h of operating, which indicates that this very near crystallization temperature affects the transmittance and deforms the network structure. It is also observed that increasing the polymer concentration may produce a standing film, but the electro-optical properties greatly modified, while increasing the concentration of LC affects the mesophase behavior. The overall results for this optimized NF LC-based polymer confined LC sample show great potential for applications like smart windows and flexible and wearable displays. The results also show great potential for electro-optic shutter/modulator with advantages such as, 1) energy saving, the working power of our device is as low as several μW/cm2 compared to traditional smart windows; 2) operated by DC power; 3) ease of fabrication; 4) low cost; and 5) applicable for flexible substrates. Further improvement in electro-optic parameters can be done by utilizing a high dielectric constant LC or advanced engineering in polymer network structure. These features will provide a unique opportunity for efficiently designing smart polymer-enriched devices for advanced applications employing the NF LCs.

In conclusion, we have developed a polymer-confined liquid crystal film prepared by employing the intriguing properties of high dielectric constant ferroelectric nematic, DIO. The phot-polymerization of ferroelectric nematic utilizing a monomeric mixture, in the presence of an AC electric field, exhibits a multidomain network structure. The resultant network structure implies an inhomogeneous combination of a vertically aligned and cross-linking network structure similar to PD&SLC structure, where the director orientations of the molecules are arbitrarily distributed in inter-domains, revealing a scattered state. Consequently, the polymerized sample exhibits a highly scattering state in the NF phase without an electric field and becomes transparent when subjected to an electric field of low amplitude. The V-T curve shows that the minimum transmittance of 1.2% at the OFF state in the NF phase, which is significantly superior to the paraelectric N phase (20%). Furthermore, the ON-state transmittance approaches 80% at a smaller voltage of less than 4 V with a negligible hysteresis. Therefore, the contrast ratio abruptly increases from 3.1 in the N phase to 57.5 in the NF phase. The rise time is consistent (∼0.27ms) in both the N and NF phases, while the fall time is much faster in the NF phase (0.6 ms) than that of the N phase (4.2 ms). These findings suggest a distinct technique for designing energy-saving and fast-switching applications, including smart windows.