Global energy policies urge us to reduce the power consumption of electronic devices, but it is inconsistent with the user’s needs of enhancing device performances, which discourages the global energy consumption requirement. In this work, the converse flexoelectricity is tailored to dramatically reduce power consumption while superior display image qualities are maintained in fringe-field switching (FFS) liquid crystal displays (LCDs). We report the first 1Hz-refresh-rate flicker-free FFS LCD in a 13.4-inch laptop display. It is inevitable having noticeable brightness deviation when the polarity of driving voltage alters at such low frequency driving, caused by flexoelectric effect and failure of voltage holding. Here, we thoroughly investigated nature of the converse flexoelectric contribution of mesogenic systems with optimum flexoelectric coefficients and molecular polarity, and we developed materials in accordance with the findings. The oxide semiconductor thin-film transistors in the backplane and LC alignment layers with optimum resistivity are adopted to promote voltage-holding characteristic and fully fill the duty cycle. The result shows that the quality of static images and videos at 1Hz- and 10Hz-refresh-rate respectively is as well maintained as that at 60Hz, and the measured power consumption is reduced by 83%, compared to that in a 60Hz-refresh-rate. It is believed that this work will open a new paradigm towards energy consumption in electronic devices, which satisfies the superior display performances in LCDs.
Article
The world’s first 1Hz-refresh-rate liquid crystal display for low power consumption
https://doi.org/10.21203/rs.3.rs-2977463/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Global energy agencies, like international energy agency (IEA), have proposed global energy policies that energy consumption needs to plateau in appliances and equipment following the net energy-zero scenarios to prevent the global warming and environmental crisis1. One indirect and long-term approach to reach this goal is to develop renewable or recyclable energy harvesting systems. In case of solar cells2–4 or piezoelectric micro-/nano-generators5–7, the piezoelectricity enables us to harvest pieces of ambient energy and transduce those to electricity. The piezoelectric effect derives a linear coupling between mechanical deformation and electric polarization in dielectric systems, which turns into flexoelectricity in mesogenic liquid crystal (LC) systems as per the strain gradient exists8–13. With proper rectifiers and batteries upon suitable methodology, the idea of self-powered devices can be implemented. From a materials perspective, LC phases can play a key role in addition to the solid or liquid state for this purpose. In the piezoelectric devices, the homogeneous stress draws responding-mechanical strain, and it causes the electric polarization, which can be described as, \({P}_{i}={P}_{o}+{e}_{ijk}{T}_{jk}\), where Pi, Po, eijk, and Tjk denote electric polarization, spontaneous polarization, a third-rank piezoelectric tensor, and a stress tensor9,10,12. On the other hand, the converse piezoelectricity can functionalize piezoelectric devices, such as actuators, soft robots, etc.11, 13–15. As the word converse implies, the reciprocal response that an electric field application induces mechanical deformation, which can be described as, \({M}_{jk}={M}_{o}+{e}_{ijk}{E}_{i}\), where Mjk and Ei denote a strain tensor and the external electric field, respectively. The symmetric stress pursues symmetric response as Curie’s principle suggests, but when considering symmetry in all materials, e.g., nematic LCs with director n, which is local directional or optic-axial representation of LC molecules, the centrosymmetric response is not valid in the LC elastic strains (splay, twist, and bend deformations) despite of the head-to-tail invariance (n = - n). In this context, flexoelectricity arises, and practically we can utilize it with respect to electric fields, formed via the pixelized interdigitated electrodes in liquid crystal displays (LCDs) in low refresh rates. The converse flexoelectricity, arisen at low-refresh-rate driven by the driving IC gives rise to reduction in power consumption. However, the flexoelectric effect causes brightness difference between the polarity of AC fields, which becomes bad image qualities and can badly irritate visual perception. Because the enhancement of device performances has mostly been in the higher priority, the effort to dig insights into this field has been relatively halfhearted despite of the emergent importance. As the reduction of power consumption in LCDs directly governs that in electronic devices over large variety of applications, studying the relation between low-refresh-rate driving technology and flexoelectricity brings huge impact on the global energy spending status. Here, we demonstrate the first 1Hz-refresh-rate flicker-free flexoelectric LCD with practically meaningful reduction in power consumption while the superior quality of static images and low-frame-rate videos are well maintained. The converse flexoelectricity plays a crucial role in this technology to achieve both high performance and energy saving as the refresh rate is linearly proportional to the power consumption of displays.
In the trajectory of the LCD development, the technological development has navigated to mostly adopt fringe-field switching (FFS) mode because of its suitability in high-resolution displays, high transmittance, wide field-of-view characteristic, and low operating voltage16,17. For the last two decades, high refresh rate like 120 to 240Hz has been developed to play smooth motion in videos, especially in gaming devices. In addition to the high refresh rate, current media environment provides enormous streaming-based information like texts, static images, or short videos, which require relatively low refresh rate. In this aspect, displays that can cover wide range of the refresh rate, from 1 to 240Hz, will dominate the markets, which coherent with the energy saving requirement. As one of the intermediate steps in such technological development, the low-refresh-rate technology at 24Hz has been tried in recent commercialized FFS LCDs. However, image qualities significantly degrade when the refresh rate reduces to 1Hz, and it appears as image flicker at the low refresh rate because the brightness differs from each AC frame, arisen mainly by flexoelectric effect in nematic LCs18–21. Several efforts to achieve the low refresh rate while maintaining the same brightness over time have been reported and can be divided into three major approaches: the first with spatially and/or temporally modulating the symmetric electric field formation when AC field alters (for example, adopting bipolar scheme19, adjusting electrodes22–25 or the thickness of the insulation layer in FFS mode23), the second with engineering the physical properties of LCs (for example, controlling dielectric anisotropy and elastic constants to adjust flexoelectric coefficients by doping bent-core LCs25–27), and the third with thorough understanding of the mechanism of the flexoelectric effects (for example, observation of the phenomenon in simple in-plane switching cells and FFS cells17,21,25). All those approaches have done separately so far, but to implement this technology in real products, exhaustive examination and comprehensive understanding in a variety of fields is required to merge into, such as theory, materials engineering, optics and display engineering and technologies.
More than 50 years ago, Meyer firstly introduced the flexoelectric effect in nematic LCs8. The system consists of elastic felas, dielectric fdielec, and flexoelectric fflexo terms based on the free-energetic consideration. The basic condition that the flexoelectric coupling arises is the broken nematic invariance caused by which there exists sufficient vector-reorientation time of the molecular polarity along the applied electric field. The flexoelectric coupling becomes strengthened locally in which the vectorial coupling between the splay and/or bend deformations and flexoelectric polarization in the same direction (constructive). On the other hand, their vector direction is opposite (destructive) so that the weak flexoelectric coupling suppresses the splay and/or bend deformations and the free-energetic cost is only or mostly paid by the twist deformation instead of splay and/or bend strains21. The flexoelectric polarization is governed by, \({{P}}_{\text{f}}={e}_{\text{s}}{n}\left(\nabla \bullet {n}\right)+{e}_{\text{b}}{n}\times \left(\nabla \times {n}\right)\), where es and eb denote splay and bend flexoelectric coefficients, respectively10,17,21. The statistical approach that Helfrich suggested can describe the es and eb based on the consideration of molecular packing as well as their dipole moment and elastic constants \({e}_{\text{s}}=\alpha {\left(\frac{a}{b}\right)}^{\frac{1}{3}}{N}^{\frac{1}{3}}\left(\frac{2{\mu }_{\parallel }{K}_{11}}{{k}_{B}T}\right)\) and \({e}_{\text{b}}=\alpha {\left(\frac{b}{a}\right)}^{\frac{2}{3}}{N}^{\frac{1}{3}}\left(\frac{{\mu }_{\perp }{K}_{33}}{2{k}_{B}T}\right)\), where a and b are the major and minor axes of molecules to consider an approximated molecular ellipsoid; α is the molecular-kink angle; µ, kB, T, and N are the dipole moment (\({\mu }_{\parallel }\) and \({\mu }_{\perp }\): dipole moments parallel and perpendicular to the long axis of molecules), Boltzmann constant, temperature, and the molecular polarity-excess number (N = N+ - N−)8,28,29. Here, when considering LC mixtures, which contain many single constituents, accurate description or modeling of the flexoelectric coefficients would be hard and imperfect. However, we can predict relatively rough but practically reasonable physical and flexoelectro-optic properties, and the fraction of single constituents via engineering the control parameters of each single in molecular level for the satisfaction of entire LC materials properties. The previously reported values of flexoelectric coefficients seem inconsistent, depending on the LC alignment and electric field formation30–34. Instead of following up all the methods to interconnect to our system, we developed a simple method to minimize errors from various factors, like ion effects, measurement errors of surface anchoring strength of vertical and planar alignments, involvement of twist deformation, and to make the electric field formation consistent to the actual FFS cells. Thus, the result can be directly transferred to the FFS system and further prototyping. The optical flickers can be verified by measuring transmittance difference between the initial transmittance of each (+) and (-) frame. From this, we found that each flexoelectric coefficients are not directly dependent on the appearing optical flicker, but the difference (es – eb) is more meaningful in FFS LCDs. Thus, we focused on how to reduce the difference in flexoelectric coefficients by controlling the parameters in the Helfrich formular to estimate flexoelectric coefficients and to take this estimation into account to the development of LC mixtures while engineering LC materials properties for the best display performances. Also, we measured and analyzed dielectric effect of LCs: dielectric anisotropy, elastic constants, and rotational viscosity, for their contribution to the flexoelectro-optic effect. By utilizing the analysis of the time-dependent luminance curves to extract the contribution of flexoelectric and voltage-holding property, the voltage-holding ratio (VHR) is extremized to fully fill the duty cycles in the signal at 1Hz refresh rate via adopting oxide semiconductor thin-film transistors (TFTs) in a backplane for low off-current characteristics35 and high-resistant alignment layer36. Finally, we demonstrate the first flicker-free 1Hz-refresh-rate 13.4-inch WUXGA FFS-LCD on a laptop. The flickers were thoroughly measured and the RGB pixels in the fabricated panel were observed, and electro-optic properties were measured. Display performances, showing text images and short videos, were evaluated and it exhibits that excellent image qualities are maintained.
Several efforts on analysis and resolving issues of the flickers based on the flexoelectro-optic effect in nematic LCs have been reported that the reason is mainly caused by the electrode structure in the FFS mode at low-refresh-rate driving as shown in Fig. 1a. Because the electric fields which couple with LC deformation are formed between one and another conducting layer by sandwiching an insulation layer, temporally asymmetric fields are formed in each positive (+) and negative (-) frame as previously verified23,25. With one of our starting LC mixtures, named as LC0 (Table 1), obvious brightness deviation in time and space is observed in POM images and the brightness profiles over the electrode location as shown in Fig. 1b. As conducting FEM simulation of the normal mode (es = eb = 0, no flexoelectricity) in an FFS mode, we can visualize the director fields of twist deformation as shown in Fig. 1c. The twist deformation is mostly originated near the electrode edge and propagated around this region because of the elastic torque. The transmittance in the FFS mode can be described as \(T={\text{sin}}^{2}2\phi \left(V\right) {\text{sin}}^{2}\left(\frac{\pi d\varDelta n\left(V\right)}{\lambda }\right)\), where j is an angle between the optic axes of LC and one of polarizers, d is the LC thickness, Dn is birefringence, and l is the wavelength. In the absence of an electric field, j = 0° and theoretical value of T = 0. As applied voltage overcomes the threshold voltage, and LC rotates to twist deformation in which j ~ 45° and \(d\varDelta n=\frac{\lambda }{2}\), T becomes maximal. In flexoelectro-optic mode (es = eb ≠ 0), the spatial brightness profile changes, compared to that in the normal mode depending on how the flexoelectric coupling occurs. In the region that the flexoelectric coupling is constructive, splay and/or bend deformation occurs more than in the normal mode, which leads to the locally reduced brightness. In the coupling-destructive region, on the contrary, the splay and/or bend deformation is suppressed as much as the flexoelectric coupling occurs less. Then, the left unpaid dielectric energy is mostly paid by the twist deformation, and the local brightness increases. Consequently, the overall brightness becomes uneven between (+) and (-) frames21. This phenomenon is clearly verified in the time-dependent transmittance curves of another LC mixture, LCref (Table 1) as shown in Fig. 1d. Each curve shows different aspects of brightness change. The transmittance difference Tdiff at the frame-transient moment (mostly within the LC response time t < 50 ms) is flexoelectric contribution. The transmittance drop Tdrop over designated time in each frame is caused by the failure of the voltage-holding. The rate of the Tdrop is different with respect to applied voltage.
Common molecular shapes that can exhibit flexoelectric effect in splay and bend deformations are illustrated in Fig. 1e,f. However, most of LC constituents in our sample mixtures are not in the pear or banana shapes. Also a single molecule can have different conformations over time as the molecular simulation results were previously reported25. We schematically illustrated how the splay and bend deformations can give rise to flexoelectric effect when LC molecules are rod-like shape in Fig. 1g,h. Assuming dipoles (dark arrows) that are exactly perpendicular to the long axis of rod-like molecules to represent our LC molecules that are negative dielectric anisotropy37, bend deformation is more suitable for the contribution to the flexoelectric polarization (purple arrows) than that with splay deformation. In case of splay deformation, on the other hand, as Prost and Marcerou proposed38, the flexoelectricity can be arisen when quadrupoles are embraced as illustrated in Fig. 1h. In the left column, there is no charge or dipole. Although some portion of plus charges enters the adjacent layer, they are canceled out. When splay deformation is imposed, the volume charge symmetry is broken, and spontaneous volume dipole occurs as illustrated in the right column. Our LC samples contain both dipole and quadrupole as shown in Table 1.
To verify the flexoelectro-optic effect based on the materials properties of nematic LCs, we measured flexoelectro-optic properties: voltage-dependent transmittance (Figure S1) and time-dependent transmittance curves (Fig. 2a). And then we extracted Tdiff of samples LCref,a,b,d,e,g,h as shown in Fig. 2b. As determined in Fig. 1d, the Tdiff between (+) and (-) frames is indicated by the dark arrows, and the Tdrop is measured as shown in Fig. 2a and S2. Among the samples, LCa,b show relatively smaller Tdrop but greater Tdiff whereas LCd,g show relatively smaller Tdiff, but greater Tdrop. In case of the samples LCref,e,h, both Tdrop and Tdiff are bad. From these results, unfortunately, we were not able to see that their elastic constants were correlated with the Tdrop and Tdiff, and thus it is insufficient for the verification of the connection, neither Tdiff nor Tdrop to the flexoelectric coefficients. The rotational viscosity g1 and dielectric anisotropy De, however, both seem to have some similar effects on the curves.
For example, the magnitude of De seems matter to the Tdiff because the higher e⊥ can suppress the LC tilt in response to the vertical component of the fringe electric fields as it is previously reported by FEM simulation results25. However, on the other hand, Tdrop becomes more significant as De gets higher because the ion impurities can be generated more from the strong polar groups in LC molecules for the high dipole moment. The g1 can also be higher as following the tendency of De. In either reason, it is obvious that avoiding this direction is encouraged because increasing De can cause reduction in Tdiff but may increase in Tdrop.
We rather focused more on verification of the flexoelectric effect. The Tdiff was extracted from the time-dependent transmittance curves with respect to applied voltages as shown in Fig. 2b-d. The values correspond to 5, 20, 40, 60, 80, 95% of transmittance to the maximum transmittance, and we respectively indicated the values as T05, T20, T40, T60, T80, T95. We set the LC parameters of LCref with some constants for quantitative comparison between LC samples. To examine the effect of De and g1 on the Tdiff, we compared LCb,e,h (Fig. 2b) of which the dielectric anisotropy varies as De = 0.4e, 1.0e, 1.2e, respectively, (e is a constant, see Table 1 for parameters) and the ratio of bend to splay elastic constant is maintained as K33/K11 ~ 0.8k31 (k31 is a constant). Although the Tdiff above T80 reaches similar Tdiff for LCb,e,h, LCe,h show lower Tdiff than LCb at the voltages lower than T80 because of higher De. By comparing LCd,e,ref, (Fig. 2c), we varied K33/K11 as 1.2k31, 0.8k31, k31, respectively, while maintaining De. The result shows that Tdiff of LCe and LCref behaves similarly with respect to the voltage. If the K33 is close to or below K11, the Tdiff does not change much. In case K33 is greater than K11, on the other hand, the Tdiff is reduced overall as shown in Fig. 2c. We verified the consistency of this result by comparing LCa,d,g as shown in Fig. 2d. The LCa,d,g have similar K33/K11 (~ 1.2k31), but De (0.4e, 1.0e, 1.2e) and g1 (0.7g, 1.4g, 1.6g) are different. Here, the Tdiff of LCa,d,g is reduced although the effect is relatively less in LCa. From this result, we learned that optimization of the ratio of bend to splay elastic constant can reduce the Tdiff.
To adjust the flexoelectric coefficients, we facilitated the measurement method of the es and eb as demonstrated in Fig. 3. The measured coefficient difference (es – eb) of the newly developed LC0-4 (Table 1) is shown in Fig. 3a-d. To minimize errors from various factors, such as ion effects, surface anchoring strength difference between vertical and planar alignments, and inconsistency of electric field formation to FFS cells, we maintained the FFS cell structure, but changed the LC alignment vertical instead of the planar alignment as illustrated in Fig. 3e. As voltage is applied, fringe-electric fields are formed, and the vertically aligned LCs are slightly tilted near the electrode edges because of the relatively intensified vertical component of electric field strength. In this field formation, we can avoid the twist deformation unless the applied voltage reaches some critical voltage that leads to deformed LC director field, which has topological defects. Beyond this defect-generating threshold voltage, topological defects will be populated, which can be originated in which the LC is tilted towards in opposite direction to the adjacent electrode and crashed each other by failing of nematic head-to-tail continuity (Figure S3). We thus put our voltage application limited below the threshold voltage that generates such defects. After obtaining Tdiff with respect to the applied voltages, we conducted FEM simulation and extract the Tdiff with various es and eb. The measured values are in a good agreement with the voltage-dependent Tdiff curves for LC0 (at es – eb = -11), LC1 (at -1.5), LC2 (at -1), and LC3 (at -0.5) as shown in Fig. 3a-d, respectively. From this result, we learned that minimization of (es – eb) is important, which is encouraging and meaningful from the engineering perspective. Previously reported measurement methods, which typically gives the sum or the difference of the coefficients, require preceding knowledge of the materials parameters, for example, anchoring strength of surfaces, Dn, and De32,39,40. In our method, the LC tilt is very slight, and the top and bottom surfaces are all vertical alignments. Also, no twist deformation is involved. Thus, the errors caused by different anchoring strength of the top and bottom alignments and by involvement of twist deformation when applying lateral field to the hybrid-aligned cell to get (es – eb) can be prevented.
We then verified the agreement between FEM simulation and experimental measurement as shown in Fig. 3f-o. The POM images of LC0-4 at V = 4.5 V are shown in Fig. 3f-i. With crossed polarizers whose optic axes are 45° to the electric field direction, we were able to observe the light transmitting the crossed polarizers via experiencing the effective birefringence based on the slight tilt of LCs near the electrode edge. We produced POM-like images of LC0 (es = 2, eb = 13), LC1 (2, 3.5), LC2 (2, 3), LC3 (2, 2.5) by using the FEM simulation (Fig. 3j-m). The produced images are in a good agreement with the POM images (Fig. 3f-i). To understand the flexoelectric contribution in this system, we calculated LC director fields of LC0 and LC3 (Fig. 3n and 3o), which correspond to the POM-like images from Fig. 3j,m.
The solid circles indicate the region where the flexoelectric coupling is constructive whereas the dotted circles indicate the destructive region. For the sake of clarity, the contribution of splay and bend flexoelectric polarizations to the constructive and destructive couplings are separately illustrated in Fig. 3p. In the positive frame, assuming the splay deformation as illustrated in the first column (in the purple solid box (i)), the splay flexoelectric coupling is constructive because the vertical components of the splay polarization and electric field are in the same direction. In contrast, in the second column (in the bluish dotted box (ii)), the vertical component of the electric field is inverted, so the splay flexoelectric coupling is destructive in this region. Assuming the bend deformation under the same electric field formation in the third column (iii), the bend flexoelectric coupling is destructive because the bend flexoelectric polarization and the lateral component of the electric field is in opposite direction. In the fourth column (iv), the bend flexoelectric coupling is constructive because of the lateral component of the bend flexoelectric polarization is reverse. Likewise, we can illustrate the splay and bend flexoelectric couplings in the negative frame. This analysis manifests that the splay and bend flexoelectric couplings in the region (i, iii) contribute to the free energy of the system oppositely those in the region (ii, iv). This implies that if (es – eb) converges to zero, those splay and bend flexoelectric couplings would be mostly canceled out. (Figure S4)
From the measurement result of the flexoelectric coefficient difference (es – eb), we can know the minimization of (es – eb) allows us to suppress the Tdiff. First, based on the rationale about es from Fig. 1g,h, we deduced that the splay flexoelectric polarization of the sample mixtures is from the contribution of quadrupole. There might be small contribution of the perpendicular dipole moment in dipoles, but the fraction would be small. If so, in our sample mixture, es would be mostly dependent on the concentration of the constituent C3. Thus, once the concentration of es is determined, we could control the (es – eb) by adjusting the eb. In case of eb, we must consider that LCs have the perpendicular dipole. Note that, to control eb, we must be aware that increasing K33 also increases eb according to the Helfrich model. Despite the Helfrich model cannot perfectly predict each splay and bend flexoelectric coefficient of mixed constituents with different LC molecular shapes, we could utilize it for estimation of the flexoelectric coefficients because considering the concentration of each constituent may provide a quite simple way for the development.
To estimate eb dependent on parameters, we assumed three mixtures (A, B, C), which contain three-different constituents with concentrations (f1, f2, f3), those are, mixture A: f1 = 0.25, f2 = 0.50, f3 = 0.25; mixture B: f1 = 0.15, f2 = 0.60, f3 = 0.25; mixture C: f1 = 0.05, f2 = 0.70, f3 = 0.25. (See Table S1 for the assumption of materials and physical parameters for the estimation.) The mixture A shows that eb reaches about 12 pC/m by increasing K33 up to 20 pN. On the other hand, the mixture C shows eb reaches about 4.5 pC/m by increasing K33 up to the same value as shown in Fig. 4a. Based on this theoretical estimation, we tuned the LC parameters and evaluated the electro-optic properties of LC0-3 in the fabricated FFS cells as shown in Fig. 4. The measured and simulated voltage-dependent transmittances of LC0-3 are in a good agreement as shown in Fig. 4b. We also measured time-dependent transmittances in Fig. 4c and Figure S5. The coefficient difference of LC0 was measured as (es – eb = -11), which shows large Tdiff whereas LC1 (es – eb = -1.5) and LC2 (es – eb = -1) show reduced Tdiff, but the Tdrop of LC1,2 is huge. However, The LC3 shows enhanced Tdiff and Tdrop (es – eb = -0.5).
The FEM simulation of the time-dependent transmittance curves of LC3 shows similar Tdiff to the result we measured where es < 2 as shown in Fig. 4d. In the simulation results, even if the difference of coefficients is maintained, the Tdiff increases as es and eb increase, which is slightly deviated from result obtained by the FFS cell structure with the vertical alignment in Fig. 3. We believe this proves that the es and eb cannot be increased without consideration of K33 and K11 or other Helfrich parameters. The POM images were observed as shown in Fig. 4e-h, which can visualize the Tdiff measured in Fig. 4c. LC3 shows similar brightness whereas LC0 shows significant brightness difference between (+) and (-) frames. We also noticed that the brightness measured at 60Hz is less than that at 1Hz. Such higher transmittance can also be an advantage when adopting 1Hz refresh rate. For LC3, we verified that the splay and bend deformations are almost evenly occur at the region of interest via the calculated LC director fields as shown in Fig. 4i. The spatial transmittance curves proved the similarity in transmittance between the (+) and (-) frames.
Finally, we demonstrate the 1Hz-refresh-rate driving 13.4-inch WUXGA FFS-LCD in a laptop as shown in Fig. 5. The measured voltage-dependent luminance is shown in Fig. 5a. As we optimized the LC alignment material to high resistivity and improve the ion impurity of LCs, the VHR of LC0-4 was measured as 92, 94.7, 95.7, and 96.3%, respectively. The power consumption P was measured 0.36W at 60Hz and 0.06W at 1Hz as shown in Fig. 5c which we confirmed about 83% reduction. We verified the image quality is well-maintained as shown in the left panel of Video S1,2. The flicker of each sample was measured as LC0 (-47dB), LC1 (-55dB), LC2 (-58dB), and LC3 (-68dB) as shown in Fig. 5d. (See Video S1,2 for verification of the human eye perception to the quantified flicker level, and Methods for the determination of the flicker level.) We adopted so-called dot-inversion driving method to further reduce the flicker level as shown in Fig. 5e,f. The POM image shows R1, G2, B1 at (+) frame while R2, G1, B2 at (-) frame as shown in Fig. 5e and Video S3,4. The driving scheme is illustrated in Fig. 5f. As the oxide TFTs were adopted, we measured the voltage drop at a pixel depending on the off-current Ioff with respect to refresh rates 120, 60, 30, 10, 1Hz as shown in Fig. 5g. The oxide TFT shows the excellent voltage-holding property, which is one of the essential conditions for the 1Hz-refresh-rate driving. The time-dependent luminance curves of the developed panel were measured, which shows excellent Tdiff and Tdrop properties as shown in Fig. 5h. The developed panel is driven at 60Hz and 1Hz in Fig. 5i,j, respectively. In Video S1,2, the left screen clearly demonstrates the flicker-free prototype panel (the sample LC3) at 1Hz-refresh-rate and the right screen is a conventional panel (the sample LC0) driven at 1Hz-refresh-rate. Additionally, we played high-quality videos on both LCDs with 60Hz- and 10Hz-refresh-rates in the left and the right side respectively in Video S5, which shows almost indistinguishable video quality.
We have demonstrated the first flicker-free 1Hz-refresh-rate 13.4-inch WUXGA FFS-LCD in a laptop based on analysis on flexoelectro-optic effect, gauging flexoelectric coefficients, materials development, and adopting oxide semiconductor TFT backplane. First, we addressed and examined the essential reasons of the flickers by flexoelectro-optic measurements of FFS unit cells. The spatial and temporal luminance difference were measured for extracting flexoelectric contribution from the voltage-holding and ion impurity effects. In the time-dependent transmittance curves, transmittance drop Tdrop (transmittance drop over the designated frame) and the transmittance difference Tdiff (the initial transmittance difference at each frame) represent failure of voltage-holding and the flexoelectro-optic effect respectively unless many ion impurities exist affecting the transmittance by the charge-screening effect near the electrode surface. The Tdrop can be dealt by enhancement of ion impurity in LCs, VHR, and adopting oxide TFTs in the backplane.
Thorough investigation of the Tdiff guides us to optimize the ratio of bend to splay elastic constant for the minimization of Tdiff. Optimization of K33/K11 can lead to suppress Tdiff, but the splay and bend elastic constants also influence the flexoelectric coefficients, which causes confliction such that increasing K33 can also increase eb. Investigation of eb based on the Helfrich model allows us being navigated to avoid such confliction although it might be imperfect to apply to mixtures of LCs with different type of constituents. In addition, we can control the concentration of LC mixtures, the molecular shape, dipole moment, dielectric anisotropy and elastic constants when mixing dipoles and quadrupoles for eb and es in negative dielectric anisotropic LC. To evaluate the es and eb, we designed an optimum system for the measurement of flexoelectric coefficients, and the result showed minimization of (es – eb) can reduce the Tdiff.
This result is meaningful because we can tune the eb and es separately by controlling the concentration of different molecules. Then, reducing (es – eb) becomes easier. The developed new LC3 showed tiny coefficient difference (es – eb ~ 0.5), which gives successful level of flicker (~ 68dB) as we thoroughly evaluated. Based on our understanding about the phenomenon and consequent materials we developed, we successfully fabricated the first prototype, flicker-free 1Hz-refresh-rate 13.4-inch FFS-LCD, which displays excellent image qualities of text images. It was also successful to play a high-quality video driven at 10Hz-refresh-rate, which is competitive with that at 60Hz. The power consumption with the 1Hz refresh rate was measured by about 83% reduction from the normal 60Hz refresh rate. We believe this result can open a new chapter of the energy-saving display technologies while maintaining display quality and contribute to the global energy efficiency in electronic devices.
Materials
We prepared LC mixtures with various materials properties along with our experimental analysis. We mixed singles with different types of polar groups and quadrupoles, arranged in Table 1. The materials were used as received without further purification.
Table 1. Relative values of physical properties of liquid crystal (LC) mixtures (Merck Performance Materials, Ltd.).
a) Birefringence; b) Dielectric anisotropy; c Rotational viscosity; d) Bend elastic constant; e) Splay elastic constant; f) The ratio of bend to splay elastic constant; g) Concentration of constituent 1, dipole; h) Constituent 2, dipole; i) Constituent 3, quadrupole; j) These are constants; k) Unspecified concentration; l) 0.4 times of the constant e. All mixtures contain small concentration of more constituents, which are not specified.
Cell preparation
two types of FFS cells that can induce planar alignment and vertical alignment were fabricated (Merck Performance Ltd.). The vertically aligned FFS cells, which are utilized for gauging flexoelectric coefficients, contain the stack of patterned electrodes, an insulation layer, and plain electrodes. Indium-tin-oxide (ITO) was used to produce the electrodes by sputtering, and silicon nitride (SiNx) was used to produce the insulation layer by PECVD. LC alignment angle, the angle between the LC alignment direction and the long axis of the patterned electrodes, is 10° and the cell structure was set as the electrode width w = 4.5 µm, distance between electrodes l = 6.0 µm, and cell gap d = 3.2 µm. The planar FFS cells, which were used for the observation of POM images and measurement of electro-optical properties, have w = 3.5, l = 6, d = 3 µm, and the LC alignment angle was 7°. Each LC mixture was injected into the cells by using the capillary reaction.
Image acquisition
POM (Eclipse LV100N POL, Nikon) images were taken by using a CCD camera (DS-Ri2, Nikon) and high-speed camera (Fastcam UX100, Photron Inc.). The image acquisition was done with 250 frames per second (fps) under applying alternating current (AC) voltage at various frequencies, including 1Hz, by an arbitrary function generator (AFG31021, Tektronix Inc.).
Electro-optic measurement
the electro-optic measurement was done by an automated acquisition system. The voltages were applied and swung between (+) and (-) frames using an arbitrary function generator (AFG31021, Tektronix Inc.), and time- or voltage-dependent transmittance curves were measured using a He-Ne laser (λ = 632.3 nm), photodiode power sensors (S130C, Thorlabs Inc.) and digital optical power meter (PM100D, Thorlabs Inc.). The driving frequency was applied by f = 1kHz at the normal mode and f = 1Hz at the flexoelectric mode.
Measurement of materials properties and the prototype panel performances
VHR was measured (1V, 3Hz, 60°C) based on the formula, where V1 is the given voltage and V2 is voltage at the end of the frame. The leakage current was measured at various off-current Ioff of oxide TFTs (comparison to LTPS TFTs) and various refresh rates (Fig. 5g). To analyze local luminance and flicker level, we used a display color analyzer (CA-410, Konica Minolta Inc.). The flicker level can be determined as, where IAC−rms is integrated maximum power spectrum of AC component and IDC is integrated power spectrum of DC component. The power consumption was measured at the logic, which controls the panel control board (PCB), excluding the power consumption from the backlight unit.
FEM simulation
The free energy density of the system is considered. The calculated electric potential is governed by Laplace's equation. The optical transmittance and POM-like images are calculated by using [2×2] extended Jones matrix method. We used a multi-dimensional finite element method (FEM) simulator (TechWiz LCD, Sanayi system). The system consists of elastic free energy density,
Acknowledgements
MinSu Kim and Seung Hee Lee thank to LG Display Co., Ltd. and Merck Performance Materials Ltd. for the project support. This project is supported by Basic Science Research Program through the National Research Foundation (NRF) of Korea, funded by the Ministry of Education [2021R1I1A1A01060001]; and through the NRF of Korea, funded by the Ministry of Science and ICT (MSIT) [2022R1A2C2091671].
Funding
National Research Foundation grant 2021R1I1A1A01060001 (MSK)
National Research Foundation grant 2022R1A2C2091671 (SHL)
Author contributions statement
M.S.K. conceived the idea, conducted the experiments, analyzed data, wrote, and edited the manuscript; D.Y.L., H.B.P., H.Y.J., J.J. performed FEM simulations and experiments; S.-H.B., J.-W.J., H.-J.P., D.-J.L. performed the experiments and fabricate the 1Hz-refresh-rate panel; S.-M.P., K.-S.K., H.-S.J. prepared and evaluated materials; G.T.K., J.K.P. discussed the project; S.H.L. initiated the project and edited the manuscript.
Conceptualization: M.S.K. S.H.L.
Methodology: M.S.K., S.-H.B., J.-W.J. H.-J.P. D.-J.L.
Investigation: D.Y.L. H.B.P. H.Y.J. J.J. S.-H.B. J.-W.J. H.-J.P. D.-J.L. G.T.K.
Visualization: M.S.K. S.-H.B., D.Y.L. H.B.P.
Supervision: M.S.K. D.-J.L. J.K.P. S.H.L.
Writing—original draft: M.S.K.
Writing—review & editing: M.S.K. S.H.L.
Additional information
Supplementary information The online version contains supplementary material available at .
Competing financial interests The authors declare no competing financial interests.
Data availability Data underlying the results presented in this paper are available from the corresponding author upon reasonable request.
- Appliances and Equipment – Analysis - IEA. https://www.iea.org/reports/appliances-and-equipment.
- Li, Z. et al. Scalable fabrication of perovskite solar cells. Nat. Rev. Mater. 2018 34 3, 1–20 (2018).
- Nayak, P. K., Mahesh, S., Snaith, H. J. & Cahen, D. Photovoltaic solar cell technologies: analysing the state of the art. Nat. Rev. Mater. 2019 44 4, 269–285 (2019).
- Ballif, C., Haug, F. J., Boccard, M., Verlinden, P. J. & Hahn, G. Status and perspectives of crystalline silicon photovoltaics in research and industry. Nat. Rev. Mater. 7, 597–616 (2022).
- Lee, B. Y. et al. Virus-based piezoelectric energy generation. Nat. Nanotechnol. 7, 351–356 (2012).
- Briscoe, J. & Dunn, S. Piezoelectric nanogenerators – a review of nanostructured piezoelectric energy harvesters. Nano Energy 14, 15–29 (2015).
- Gu, L. et al. Enhancing the current density of a piezoelectric nanogenerator using a three-dimensional intercalation electrode. Nat. Commun. 11, 1–9 (2020).
- Meyer, R. B. Piezoelectric Effects in Liquid Crystals. Phys. Rev. Lett. 22, 918–921 (1969).
- Tagantsev, A. K. Piezoelectricity and flexoelectricity in crystalline dielectrics. Phys. Rev. B 34, 5883 (1986).
- Buka, A. & Eber, N. Flexoelectricity in Liquid Crystals: Theory, Experiments and Applications. (World Scientific, 2012).
- Jákli, A. & Éber, N. Soft materials for linear electromechanical energy conversion. Curr. Opin. Chem. Eng. 2, 120–124 (2013).
- Wang, B., Gu, Y., Zhang, S. & Chen, L. Q. Flexoelectricity in solids: Progress, challenges, and perspectives. Prog. Mater. Sci. 106, (2019).
- Grasinger, M., Mozaffari, K. & Sharma, P. Flexoelectricity in soft elastomers and the molecular mechanisms underpinning the design and emergence of giant flexoelectricity. Proc. Natl. Acad. Sci. U. S. A. 118, e2102477118 (2021).
- Jin, T. et al. Triboelectric nanogenerator sensors for soft robotics aiming at digital twin applications. Nat. Commun. 11, 1–12 (2020).
- Zhang, S. et al. Piezo robotic hand for motion manipulation from micro to macro. Nat. Commun. 14, 1–12 (2023).
- Lee, S. H., Lee, S. L. & Kim, H. Y. Electro-optic characteristics and switching principle of a nematic liquid crystal cell controlled by fringe-field switching. Appl. Phys. Lett. 73, 2881 (1998).
- MinSu Kim, Seok Lyul Lee & Seung Hee Lee. Fringe-field switching technology. in High Quality Liquid Crystal Displays and Smart Devices - Volume 1: Development, display applications and components 55–77 (Institution of Engineering and Technology, 2019). doi:10.1049/PBCS068F_ch3.
- Tsuruma, T. et al. Novel image sticking model in the fringe field switching mode based on the flexoelectric effect. in Proc. of 31st International Display Research Conference(EuroDisplay) 13–16 (2011).
- Oh, S.-W., Park, J.-H., Lee, J.-H. & Yoon, T.-H. Elimination of image flicker in a fringe-field switching liquid crystal display by applying a bipolar voltage wave. Opt. Express 23, 24013–8 (2015).
- Lee, D.-J. et al. Transient flickering behavior in fringe-field switching liquid crystal mode analyzed by positional asymmetric flexoelectric dynamics. Opt. Express 23, 34055 (2015).
- Kim, M. S. et al. Flexoelectric effect in an in-plane switching (IPS) liquid crystal cell for low-power consumption display devices. Sci. Rep. 6, 35254 (2016).
- Oh, S.-W., Park, J.-H., Baek, J.-M., Choi, T.-H. & Yoon, T.-H. Effect of electrode spacing on image flicker in fringe-field-switching liquid crystal display. Liq. Cryst. 43, 972–979 (2016).
- Kim, M. S. et al. Field-symmetrization to solve luminance deviation between frames in a low-frequency-driven fringe-field switching liquid crystal cell. Opt. Express 24, 29568 (2016).
- Kim, M. S. et al. Flexoelectric in-plane switching (IPS) mode with ultra-high-transmittance, low-voltage, low-frequency, and a flicker-free image. Opt. Express 25, 5962 (2017).
- Kim, M. S. et al. Liquid Crystals for Superior Electro-Optic Performance Display Device with Power-Saving Mode. Adv. Opt. Mater. 6, 1800022 (2018).
- Kim, J.-W., Choi, T.-H., Yoon, T.-H., Choi, E.-J. & Lee, J.-H. Elimination of image flicker in fringe-field switching liquid crystal display driven with low frequency electric field. Opt. Express 22, 30586 (2014).
- Chen, H., Peng, F., Hu, M. & Wu, S.-T. Flexoelectric effect and human eye perception on the image flickering of a liquid crystal display. Liq. Cryst. 42, 1730–1737 (2015).
- Helfreich, W. The Strength of Piezoelectricity in Liquid Crystals. Zeitschrift für Naturforsch. A 26, 833–835 (1971).
- Helfrich, W. A simple method to observe the piezoelectricity of liquid crystals. Phys. Lett. A 35, 393–394 (1971).
- Castles, F., Morris, S. M. & Coles, H. J. The limits of flexoelectricity in liquid crystals. AIP Adv. 1, 032120 (2011).
- Ponti, S., Ziherl, P., Ferrero, C. & Zumer, S. Flexoelectro-optic effect in a hybrid nematic liquid crystal cell. Liq. Cryst. 26, 1171–1177 (1999).
- Takahashi, T. et al. Novel measurement method for flexoelectric coefficients of nematic liquid crystals. Jpn. J. Appl. Phys. 37, 1865–1869 (1998).
- Madhusudana, N. V. & Durand, G. Linear flexo-electro-optic effect in a hybrid aligned nematic liquid crystal cell. J. Phys. Lettres 46, 195–200 (1985).
- Harden, J. et al. Giant flexoelectricity of bent-core nematic liquid crystals. Phys. Rev. Lett. 97, 157802 (2006).
- Nomura, K. et al. Room-temperature fabrication of transparent flexible thin-film transistors using amorphous oxide semiconductors. Nature 432, 488–492 (2004).
- Reducing ion impurity and increasing resistivity of alignment layer enable us to minimize the Tdrop, which is rather clear and already well-verified in industrial fields, so the detailed development procedure is out of the scope of this paper.
- The direction of dipole moment might not be exactly perpendicular to the long axis of the molecules in real world. This means that the contribution of the perpendicular dipole moment to es can be non-zero, but might be weak.
- Prost, J. & Marcerou, J. P. On the microscopic interpretation of flexoelectricity. J. Phys. 38, 315–324 (1977).
- Barbero, G., Valabrega, P. T., Bartolino, R. & Valenti, B. Evidence for the flexo-electric effect in a nematic liquid crystal cell. Liq. Cryst. 1, 483–490 (1986).
- Valenti, B., Bertoni, C., Barbero, G., Taverna-Valabrega, P. & Bartolino, R. Flexoelectricity in the Hybrid Aligned Nematic Cell. Mol. Cryst. Liq. Cryst. 146, 307–320 (1987).
There is NO Competing Interest.
- 0SI1HzDrivingNatCommun.pdf
Supplementary information
- VS11Hz64grayT05Tmax30Nit.mov
LC3 (left) and LC0 (right) in the prototype panels, driven at 1Hz refresh rate. 64 gray, T05, ~ 30Nit.
- VS21Hz127grayT20Tmax120Nit.mov
LC3 (left) and LC0 (right) in the prototype panels, driven at 1Hz refresh rate. 127 gray, T20, ~ 120Nit.
- VS31Hzrecord250fpsplay250fpsRealtime.mov
Pixels in the LC3 prototype panel, driven at 1Hz refresh rate. Recorded at 250fps, played at 250fps (real time).
- VS41Hzrecord250fpsplay15fpsSlow.mov
Pixels in the LC3 prototype panel, driven at 1Hz refresh rate. Recorded at 250fps, played at 15fps (slow).
- VS560hz10hzvideo.mov
A high-quality video played by a normal panel (left) driven at 60Hz, and the prototype panel (right) driven at 10Hz.