Electric-field-assisted signal amplification for label-free liquid-crystal-based detection of biomolecules

: The fundamental principle of liquid-crystal (LC)-based biosensing is the sensitive response of LC orientation to external stimuli. Biomolecules such as proteins or DNAs immobilized on the glass substrate of a LC cell are detected through disrupting the LC alignment and, in turn, altering the birefringence, resulting in changes in the optical texture that can be readily observed under a polarizing optical microscope. With an additional weak electric ﬁeld across a sandwiched LC cell, we demonstrate in this study a novel label-free biodetection technique with ampliﬁed signal and improved detection limit. By applying the binarization analysis as the quantitative approach, the increase in the light leakage area in the optical texture of LCs with increasing amount of biomolecules can be quantitated with a bright-area-ratio (BAR)-versus-concentration curve. The reported biosensing technique exploits both the optical and electrical properties of LCs and is potentially applicable to other LC-based rapid screening and bioassays. and biomolecular interactions, a label-free DNA hybridization assay using MLC-6884 as the sensing mesogen was manifested with electric-ﬁeld assistance. The capture ssDNA was ﬁrst immobilized on the DMOAP-coated glass substrate at various concentrations for the assessment of optimal hybridization conditions. A dark-to-bright transition in the optical texture of MLC-6884 was observed when ssDNA concentration increased from 10 − 9 M to


Introduction
Liquid crystals (LCs) are anisotropic materials exhibiting unique optical, electro-optical and dielectric properties, which originate from the self-organization of the elongated LC molecules in response to various external stimuli, such as heat, light, electric and magnetic fields, and the chemical composition of the materials in contact with LCs. Optical anisotropy, or birefringence, is the most commonly exploited characteristic of LCs in the emerging field of LC-based biosensing, in which the orientational order of LCs, usually determined by the alignment reagent at the LC-glass or LC-water interface, is disturbed by biomolecules on the aligning layer, thus altering the interaction of LCs with light and the resulting optical appearance when observed under a polarizing optical microscope (POM) [1][2][3][4][5][6][7][8][9]. It was not until recent years that quantitative biodetection technologies derived from capacitance and electro-optical measurements as well as dielectric spectroscopy were developed, in which the effect of an external electric field on LC orientation was analyzed [3,5,10,11].
Under the influence of an electric field, the polar LC molecules are easily polarized to change their orientation from the initially vertical alignment when the applied voltage exceeds the threshold for Fréedericksz's transition-a threshold phenomenon arising from the influence of the boundary opposing the response to the electric field, where K 33 is the bend elastic constant, ε 0 is the permittivity of free space and ∆ε is the dielectric anisotropy corresponding to the difference between the parallel and perpendicular components of the permittivity, | | and ε ⊥ , measured when the long axis (i.e., the director) of LC is parallel and perpendicular to the external electric field, respectively (∆ε = | | -ε ⊥ ). The behavior of a specific LC mesogen subjected to an electric field can therefore be predicted depending on whether it is a positive LC (∆ε > 0) or a negative one (∆ε < 0). A positive LC is polarized along the director and tends to be aligned parallel to the direction of the electric field beyond the threshold, while a negative LC has a stronger component of the dipole moment along its short axis so that its director tends to become perpendicularly oriented to the electric field. Signal amplification is a major challenge in the design of a LC-based biosensor, which is necessary to improve the sensitivity and limit of detection (LOD) in order to meet the requirements of clinical and biomedical applications. While signal amplification of several DNA biosensors based on LC was mediated by labeling the analyte or the detection interface with metal nanoparticles [12][13][14], there is limited discussion in label-free LC-based detection on strategies to enhance the optical signal transduced by biomolecules or biomolecular interactions. We found that the birefringence (∆n) of the sensing mesogen is a determining factor to enhance signal transduction. The brightness of the LC optical texture was enhanced when 5CB (∆n = 0.179 at 589.3 nm and 25°C), the conventional nematic LC selected for most LC-based biosensors, was replaced with HDN (∆n = 0.333 at 589.3 nm and 20°C), a nematic LC of larger birefringence and much wider nematic range [15][16][17]. Chemical modification of the glass surface constituting the LC-glass interface for biodetection increased the binding affinity of analytes, leading to a desired lower LOD [15]. In addition, incident light linearly polarized along the rubbing direction of the glass substrate promoted the detection sensitivity of a protein assay based on dye-doped LC [4].
In this study, a novel approach of field-assisted signal amplification is proposed by subjecting a negative LC (MLC-6884) to an electric field near (i.e., slightly weaker or stronger than) the threshold field required to induce Fréedericksz's transition, thereby weakening the penetrating anchoring force of the alignment layer. Each LC cell was constructed with a pair of glass substrates coated with the homeotropic alignment reagent, dimethyloctadecyl [3-trimethoxysilyl)propyl] ammonium chloride (DMOAP). Electric-field assistance was exploited in a wide variety of studies such as protein crystallization, anion−π catalysis and formation of perovskite films for solar cells [18][19][20]. Our early work utilized an AC voltage (of 3.5 V rms at 1 kHz) below Fréedericksz's transition threshold to meliorate the acousto-optical effect in a 200-µm-thick nematic LC film [21]. Here again the externally applied electric field was not strong enough to effectively overcome the anchoring strength of DMOAP. Neither could this weak field alter the structure and function of biomolecules such as protein and DNA, which are usually analyzed in gel electrophoresis at a much higher voltage (100-150 V) depending on the experimental setup with their three-dimensional structures intact. Presumably an assisting voltage weakens the surface anchoring force, contributing to a higher propensity for MLC-6884 to be loosen from the homeotropic state, thereby facilitating its response to the boundary condition and enhancing the light leakage when the vertical alignment disrupted by immobilized bovine serum albumin (BSA) at the LC-glass interface. The signal enhancement was quantitated through image processing and binarization analysis, which improved on our earlier quantitative approach toward optical texture images [17]. Finally, an electric-field-assisted label-free DNA hybridization assay based on MLC-6884 was extendedly performed to demonstrate the feasibility of such signal enhancement in the detection of DNA and DNA interactions.

Materials
with a pair of conductive glass substrates, each LC cell has an overlapped electrode area of 5.0 mm × 5.0 mm. Dimethyloctadecyl [3-trimethoxysilyl)propyl] ammonium chloride (DMOAP), a silane surfactant used as the homeotropic alignment agent, was purchased from Sigma-Aldrich (St. Louis, Missouri). The nematic LC, MLC-6884, used in this study as the sensing mesogen was obtained from Merck (Darmstadt, Germany). MLC-6884 exhibits an extraordinary refractive index n e of 1.5775 and an ordinary refractive index n o of 1.4805, yielding a birefringence (∆n ≡ n e − n o ) of 0.097 at a wavelength of 589.3 nm and a temperature of 20°C. The dielectric constants of MLC-6884 parallel and perpendicular to the nematic director are 4.0 and 9.0 (at 1.0 kHz and 20°C), respectively, indicating that the nematic possesses negative dielectric anisotropy (∆ε < 0). Bovine serum albumin (BSA), a protein standard with a molecular weight of 66.43 kDa, was received from Sigma-Aldrich (St. Louis, Missouri). Two complementary 48-bp single-stranded DNA (ssDNA) oligonucleotides, 5'-CGC CCC CAT GTT CGT CAT GGG TGT GAA CCA TGA GAA GTA TGA CAA CAG-3' and 5'-GCG GGG GTA CAA GCA CTA GTA CCC ACA CTT GGT ACT CTT CAT ACT GTT GTC-3' were synthesized by GeneDireX (Taichung, Taiwan). Deionized distilled water (DI water) was purified by an RDI reverse osmosis/deionizer. All BSA and DNA solutions were prepared or diluted in sterilized DI water, pH 7, which is DI water autoclaved at 121°C and 15 psi for at least 30 min to eliminate bacteria and deoxyribonuclease (DNase) contamination. The pH of sterilized DI water was checked regularly and can be maintained at pH 7 for over six months at room temperature when stored in glass bottles.

LC-based protein and DNA assays
The ITO-coated glass slides were chemically modified with DMOAP by reacting with 1% (v/v) DMOAP in DI water at room temperature for 15 min. After rinsed with DI water, the glass substrates were dried under a stream of nitrogen and baked at 85°C for 15 min. To prepare the LC cell for protein detection, BSA was immobilized on the electrode area of a DMOAP-modified ITO-coated glass slide in a protein array format by manually dispensing BSA solution at the designated concentration on the glass substrate using a P20 micropipette accurate for a liquid volume of 2-20 µl. Each protein array consisted of 2 replicates of BSA protein of 3 µl/dot arranged in a 1 × 2 format. The protein array was allowed to dry in an oven for 1 h at 35°C, followed by rinsing with DI water to remove unbound BSA. The LC cell was then assembled by pairing a DMOAP-coated glass substrate with that immobilized with the BSA protein array, separated by ball spacers of 4.2 µm in diameter to form a cell gap of 4.7 ± 0.5 µm as determined by white-light interference. The resulting LC cell with an overlapped electrode area of 5.0 mm × 5.0 mm for the electric field-assisted detection to take effect was sealed with epoxy glue and completely filled with MLC-6884 by capillary action. For DNA detection, 1 nM of the ssDNA capture probe was immobilized on the DMOAP-coated substrate in a 1 × 2 array format at 3 µl/dot, followed by rinsing and drying. DNA hybridization was performed by dispensing 10 µl of the complementary ssDNA at the designated concentration on the glass substrate with the immobilized DNA array and covering it with a clean cover glass. The hybridization reaction was allowed to occur for 1 h at 50°C, which was close to the annealing temperature of the DNA sequences. After removing the cover glass and rinsing the glass substrate with DI water, the LC cell was assembled as described above.

Electric field-assisted biosensing and quantitative analysis
The optical texture of MLC-6884 was examined with an Olympus BX51 POM (Tokyo, Japan) with crossed polarizers in the transmission or reflection mode. In order to experimental find the threshold voltage of an MLC-6884 cell, voltage-dependent transmission was performed at room temperature by employing a probe beam derived from an unpolarized He-Ne laser of 633 nm in wavelength. The transmitted light through the sample situated between two crossed linear polarizers was measured by a photodiode detector. Field-assisted biodetection was achieved by applying an external voltage of the order of 1 V rms at a frequency of 1 kHz across the LC cell thickness during texture observation. The AC voltage was supplied with a Tektronix AFG3021B function generator (USA, Oregon). To obtain quantitative results for texture analysis, Digimizer, an image analysis software with various image processing and analysis capabilities, was exploited in the measurement of the perimeter and area of the optical texture of LCs as well as the subsequent binarization analysis.

Results and discussion
3.1. The optical texture of MLC-6884 in response to various concentrations of BSA As illustrated in Fig. 1, the homeotropic alignment of MLC-6884 directed by DMOAP coated on both glass substrates was disrupted due to the presence of immobilized BSA at the LC-glass interface. When observed under a POM with crossed polarizers, the MLC-6884 molecules in contact with BSA deviated from the original orientation and gave rise to light leakage. The level of molecular disturbance in MLC-6884 in response to increasing concentrations of BSA can therefore be examined through the dark-to-bright transformation in LC optical texture (Fig. 2, upper panel). In each micrograph, immobilized BSA and thus the change in optical texture were confined within the circled area, outside of which MLC-6884 remained homeotropically aligned, suggesting that the concentration-dependent light leakage resulted primarily from the immobilized BSA. The brightness of the optical texture in the presence of 10 −11 -g/ml BSA was discernible from the dark background, but that of 10 −12 -g/ml BSA was completely dark, indicating that the amount of BSA was insufficient to induce an optically observable disturbance. Consequently, we can estimate the LOD to be ∼10 −11 -g/ml BSA for the protein assay based on the optical texture of MLC-6884. two crossed biodetection frequency of was supplied quantitative r image process and area of th

The opti
As illustrated on both glass glass interfac molecules in leakage. The concentration in LC optical the change i MLC-6884 re light leakage texture in the of 10 −12 -g/ml to induce an õ 10 −11 -g/ml B

Electric field-assisted protein biosensing
The Fréedericksz threshold voltage of MLC-6884 can be calculated by substituting K 33 = 14.8 pN and ∆ε = −5 into Eq. (1), yielding a threshold voltage of 1.82 V rms , above which the director of the MLC-6884 molecules will be tilted from the initially unperturbed vertical alignment, leading to molecular reorientation at a certain angle. As evident from electro-optical (or voltage-transmittance) measurements, the transmittance of MLC-6884 was 0% at voltages below the threshold but increased with increasing voltage beyond 1.82 V rms at 1 kHz ( Fig. 3(a)). The  corresponding optical texture of MLC-6884 remained dark at 0 and 1.7 V rms ( Fig. 3(b) and 3(c)), but the slight increase in light leakage at 1.9 V rms suggests that although the voltage was not high enough to realign MLC-6884 from fully perpendicular to fully parallel to the glass substrate, the LC was no longer aligned uniformly ( Fig. 3(d)).

Electric field-assisted protein biosensing
The Fréedericksz threshold voltage of MLC-6884 can be calculated by substituting K 33 = 14.8 pN and Δε = −5 into Eq. (1), yielding a threshold voltage of 1.82 V rms , above which the director of the MLC-6884 molecules will be tilted from the initially unperturbed vertical alignment, leading to molecular reorientation at a certain angle. As evident from electro-optical (or voltage-transmittance) measurements, the transmittance of MLC-6884 was 0% at voltages below the threshold but increased with increasing voltage beyond 1.82 V rms at 1 kHz ( Fig. 3(a)). The corresponding optical texture of MLC-6884 remained dark at 0 and 1.7 V rms (Fig. 3(b) and 3(c)), but the slight increase in light leakage at 1.9 V rms suggests that although the voltage was not high enough to realign MLC-6884 from fully perpendicular to fully parallel to the glass substrate, the LC was no longer aligned uniformly ( Fig. 3(d)). It is assumed that in an electric field slightly lower than the threshold voltage, the increase in the distortion free energy density of MLC-6884 was insufficient to overcome the anchoring strength of DMOAP so that MLC-6884 remained homeotropically aligned but were given a higher tendency to tilt in comparison with its configuration in the absence of the electric field. This may endow MLC-6884 with a higher propensity to respond more sensitively to the presence of biomolecules at the LC-glass interface. Indeed, when an AC It is assumed that in an electric field slightly lower than the threshold voltage, the increase in the distortion free energy density of MLC-6884 was insufficient to overcome the anchoring strength of DMOAP so that MLC-6884 remained homeotropically aligned but were given a higher tendency to tilt in comparison with its configuration in the absence of the electric field. This may endow MLC-6884 with a higher propensity to respond more sensitively to the presence of biomolecules at the LC-glass interface. Indeed, when an AC voltage of 1.7 V rms at 1 kHz was applied, the brightness of the optical texture of MLC-6884 was drastically enhanced at each BSA concentration (Fig. 2, lower panel). The LOD was further lowered at least an order of magnitude to 10 −12 g/ml BSA. It is therefore concluded that by loosening the LC molecules from the restraint of the alignment layer with an externally applied electric field slightly lower than the Fréedericksz's threshold voltage, the optical signal, resulting from the disruption in the homeotropic alignment of MLC-6884 by BSA, can be effectively amplified.
The cell gap of the LC cell is an important factor in determining the extent of the electricfield-assisted signal amplification. When exposed to an electric field of the same strength used in this study, signal amplification was not observed in a LC cell with a larger cell gap of 8.2 µm (data not shown). It is therefore suggested that there exists an optimal cell gap for maximal signal amplification, but at the present stage of method development, a cell gap around 4.7 µm was arbitrarily chosen for the ease of manual sample fabrication at which a significant enhancement in optical signal was observed. It should also be noted that the electric field-assisted signal amplification demonstrated here is feasible only when negative LC was applied as the biosensing mesogen. Positive and negative LCs are induced to reorient in different (orthogonal) directions in an electric field. When DMOAP was used as the alignment layer under the influence of a vertical electric field, MLC-6884 tends to be loosen from homeotropic alignment, but positive LCs such as 5CB will be more firmly anchored in the homeotropic state, making it more difficult for biomolecules to disrupt the orientation of 5CB. We have previously established that the nematic LC HDN with birefringence larger than 5CB gave rise to enhanced optical signal and lower LOD [15][16][17]. Here it is demonstrated that electric-field assistance may be considered a signal amplification strategy for negative LCs (with birefringence lower than 5CB).

Quantitation of optical texture images through binarization analysis
As demonstrated in Fig. 2, the brightness of the optical texture of MLC-6884 was dependent of the BSA concentration, but such qualitative observation was insufficient to fulfill the criteria for a desired LC-based biosensor with full functionality. Quantitative analysis of LC optical texture with an image processing software has been reported in our previous studies [17]. Relative intensity calculated from the total RGB value in the micrograph of the POM correlated with the brightness of LC optical texture as well as BSA concentration, but the nonlinearity of the resulting calibration curve rendered the method impractical for protein quantitation.
In this study, Digimizer, an image analysis software, was utilized in the binarization analysis of the optical texture images in an attempt to establish an improved quantitative procedure. For a LC cell at an external voltage of 1.7 V rms in the absence of immobilized biomolecules, the gray level intensity given by Digimizer was 10, which was defined as the threshold for background light leakage in our analysis. Any light leakage caused by the presence of biomolecules and thus the disturbance of homeotropic alignment of MLC-6884 should result in a grey level intensity higher than the threshold. In the binarization analysis, the pixels in the optical texture image with grey level intensity greater than 10 were defined as positive pixels, which were added up within a circular region slightly larger than the area occupied by immobilized BSA of 1,130 pixels in diameter. The circular area was confined by a manually added virtual mask to eliminate background noise coming from sections in the micrograph without BSA. Finally, the sum of the positive pixels was divided by the area of the defined circular region to obtain the bright area ratio (BAR), which was plotted against BSA concentration.
As shown in Fig. 4, the correlations between BAR and BSA concentration at 0 and 1.7 V rms was consistent with those observed in the optical texture (Fig. 2). In addition, the BAR at BSA concentrations from 10 −6 g/ml to 10 −12 g/ml was significantly higher in electric field-assisted protein detection compared with that obtained without field assistance. At 10 −5 -g/ml BSA, the BAR in the presence and absence of an externally applied voltage both reached a similar value near 90%, suggesting that light leakage at the LC-glass interface has achieved a maximum, at which the relatively large amount of BSA blocked the anchoring effect of DMOAP and gave rise to significant disruption in the orientation of MLC-6884 so that signal amplification contributed by the external voltage was rendered irrelevantly. On the other hand, the BAR at 10 −12 g/ml was 3% and 1% with and without electric-field assistance, respectively. The three-fold increase in BAR when an electric field was imposed confirmed the improved LOD as seen in Fig. 2. grey level intensity of 30 in the absence of BSA (Fig. 3(d)). By redefining the threshold for background light leakage at 30 and calculating BAR, we found that the BAR at each BSA concentration was significantly higher than that measured at 1.7 V rms or at null voltage. This indicates that the optical signal was further enhanced at 1.9 V rms but at the cost of a higher background. However, when the external AC voltage was increased to 2.0 V rms , at which the threshold for background light leakage was set at a gray level intensity of 80, the values for BAR at each BSA concentration was reduced drastically from those at 1.7 V rms to similar to those in the absence of an assisting electric field. This can be explained by the already distorted orientation of MLC-6884 at 2.0 V rms , leaving limited room for BSA disruption to take effect before maximum light leakage was reached, thus resulting in the relatively smaller values of BAR when the background noise was filtered. These results suggest that there exists an optimal voltage close to that of Fréedericksz's transition for an externally applied electric field to enhance the optical signal of LC-based biodetection.

Electric-field-assisted protein biosensing in the reflection mode of the polarizing optical microscope
By switching the detection mode of the POM from the transmission mode to the reflection mode, we expect to further enhance the optical signal from the electric field-assisted bioassay system through the increase in the optical path length. Although the" mean free path" remains the same, the optical path length in the reflection mode was twice longer than that in the transmission mode, resulting in two times more scattering centers encountered by the optical beam and hence the potential amplification of optical signals. When an electric field of 0, 1.7, 1.9 or 2.0 V rms at a frequency of at 1 kHz was applied across the LC cell, the BAR calculated from the optical textures obtained in the reflection mode increased with increasing concentrations of BSA in a similar fashion to that observed in the transmission mode, in which optical signal amplification was evident at 1.7 and 1.9 V rms , but was not observed at 2.0 When the intensity of the electric field was increased to 1.9 V rms at 1 kHz, which slightly exceeded the Fréedericksz threshold voltage of 1.82 V rms , distortion of the MLC-6884 director was commenced, bringing about a brighter background than that at 1.7 V rms with a grey level intensity of 30 in the absence of BSA (Fig. 3(d)). By redefining the threshold for background light leakage at 30 and calculating BAR, we found that the BAR at each BSA concentration was significantly higher than that measured at 1.7 V rms or at null voltage. This indicates that the optical signal was further enhanced at 1.9 V rms but at the cost of a higher background. However, when the external AC voltage was increased to 2.0 V rms , at which the threshold for background light leakage was set at a gray level intensity of 80, the values for BAR at each BSA concentration was reduced drastically from those at 1.7 V rms to similar to those in the absence of an assisting electric field. This can be explained by the already distorted orientation of MLC-6884 at 2.0 V rms , leaving limited room for BSA disruption to take effect before maximum light leakage was reached, thus resulting in the relatively smaller values of BAR when the background noise was filtered. These results suggest that there exists an optimal voltage close to that of Fréedericksz's transition for an externally applied electric field to enhance the optical signal of LC-based biodetection.

Electric-field-assisted protein biosensing in the reflection mode of the polarizing optical microscope
By switching the detection mode of the POM from the transmission mode to the reflection mode, we expect to further enhance the optical signal from the electric field-assisted bioassay system through the increase in the optical path length. Although the" mean free path" remains the same, the optical path length in the reflection mode was twice longer than that in the transmission mode, resulting in two times more scattering centers encountered by the optical beam and hence the potential amplification of optical signals. When an electric field of 0, 1.7, 1.9 or 2.0 V rms at a frequency of at 1 kHz was applied across the LC cell, the BAR calculated from the optical textures obtained in the reflection mode increased with increasing concentrations of BSA in a similar fashion to that observed in the transmission mode, in which optical signal amplification was evident at 1.7 and 1.9 V rms , but was not observed at 2.0 V rms (Fig. 5). Because BSA is a protein standard commonly used in constructing calibration curves for the quantitation of protein samples, linear regression was performed on the plots of BAR versus BSA concentration in both the transmission and reflection modes at 1.9 V rms within the concentration range between 10 −12 g/ml and 10 −6 g/ml for comparison (Fig. 6). There was no significant difference between the calibration curves obtained in the transmission and reflection modes, indicating that switching to the reflection mode did not further promote signal amplification as expected. Nevertheless, the relatively linear correlation between BAR and the logarithm of BSA concentration and the relatively small standard deviation for the BAR values from repeated experiments suggest that the binarization analysis demonstrated in this study provides an improved quantitative approach compared with the RGB analysis reported previously [17].
V rms (Fig. 5). Because BSA is a protein standard commonly used in constructing calibration curves for the quantitation of protein samples, linear regression was performed on the plots of BAR versus BSA concentration in both the transmission and reflection modes at 1.9 V rms within the concentration range between 10 −12 g/ml and 10 −6 g/ml for comparison (Fig. 6).
There was no significant difference between the calibration curves obtained in the transmission and reflection modes, indicating that switching to the reflection mode did not further promote signal amplification as expected. Nevertheless, the relatively linear correlation between BAR and the logarithm of BSA concentration and the relatively small standard deviation for the BAR values from repeated experiments suggest that the binarization analysis demonstrated in this study provides an improved quantitative approach compared with the RGB analysis reported previously [17].

Electric field-assisted DNA hybridization
To demonstrate the feasibility of the proposed signal amplification strategy in the detection of other biomolecules and biomolecular interactions, a label-free DNA hybridization assay using MLC-6884 as the sensing mesogen was manifested with electric-field assistance. The capture ssDNA was first immobilized on the DMOAP-coated glass substrate at various concentrations for the assessment of optimal hybridization conditions. A dark-to-bright transition in the optical texture of MLC-6884 was observed when ssDNA concentration increased from 10 −9 M to 10 -12 10 -11 10 -10 10 -9 10 -8 10 -7 10 -6 10 -5 0 BSA concentration (g/ml) B Fig. 5. Quantitative analysis of the optical texture of MLC-6884 obtained in the reflection mode of the POM in the field-assisted protein assay through binarization analysis. BAR calculated from the results of binarization analysis is plotted against BSA concentration with various applied voltages at 1 kHz. Error bars represent the standard deviation calculated from the BAR value of at least three independent experiments. Cell gap: 4.7 ± 0.5 µm. 10 −8 -M, suggesting that 10 −9 M was the maximal ssDNA concentration that can be allowed at the LC-glass interface without disrupting the homeotropic alignment of MLC-6884 (Fig. 7).  In the DNA hybridization process, we immobilized 10 −9 -M ssDNA on the DMOAP-coated glass substrate to establish a dark background, followed by reaction with various concentrations of complementary ssDNA so that any light leakage occurred should be attributed to the binding and DNA duplex formation between the immobilized and the complementary ssDNAs. Because DNA hybridization was completed prior to the assembly of LC cells and detection with the  mesogenic substance (MLC-6884 in this study), most LC biosensing platforms designed at the LC-glass interface as the one demonstrated here are end-point assays instead of real-time detection, which is currently more easily realized at the LC-aqueous interface. Nevertheless, by investigating biocompatible LC mesogen resembling the aqueous environment for biomolecules to be soluble in LCs while maintaining their biological activity and three-dimensional structure, it is possible to implement real-time biodetection at the LC-glass interface. As shown in Fig. 8, the relative weak signal produced by reacting 10 −9 -M capture ssDNA with complementary ssDNA at 10 nM to 10 µM can be enhanced by applying an external AC voltage of 1.7 V rms at 1 kHz with an LOD estimated to be 0.1 µM. Quantitative results from binarization analysis indicates that, although discernible in Fig. 8 under electric-field assistance, the change in the brightness of the optical texture with DNA concentration was statistically insignificant when expressed in BAR due to the relatively large standard deviation (Fig. 9). On the other hand, when 10 −8 -M complementary DNA was reacted with the immobilized capture DNA, the BAR calculated from the optical texture of MLC-6884 was close to 0% in the absence of the electric field, but was amplified 3-fold in the field-assisted DNA hybridization at 1.7 V rms (Fig. 9). As found previously in the electric-field-assisted protein assay, increasing the external voltage to 1.9 V rms gave rise to greater BAR values but accompanied by a higher background, and the effect of signal amplification diminished when the electric field was further elevated to 2.0 V rms , at which the MLC-6884 molecules were already triggered to be reoriented under the influence of the electric field instead of the formation of the DNA duplex. These results provide evidence for the applicability of the field assistance technique to other label-free LC-based bioassays for the purpose of signal amplification.

Conclusions
Strategies for signal amplification are indispensable for label-free LC-based biosensing but are seldom investigated. In this study, by stimulating the LC mesogen with an electric field slightly lower than the threshold voltage for Fréedericksz's transition, the LC molecules are loosened from the constraint of the homeotropic alignment layer and are more prone to rotate, leading to a more significant optical response and a lower LOD in the presence of biomolecules at the LC-glass interface. In addition, a quantitative approach for the LC optical texture was proposed utilizing binarization analysis, from which linear correlation can be derived between the optical signal, expressed by BAR, and the concentration of biomolecules. The field-assisted label-free LC-based biosensing technique was demonstrated in protein assay using BSA as the protein standard, as well as in biomolecular interaction (DNA hybridization in this work). It is concluded that signal amplification through electric-field assistance is potentially applicable to a wide variety of LC-based bioassays, including protein-protein or protein-peptide binding, immunoreactions, and enzymatic activity assays. To further extend the application of the current biosensing platform at the LC-glass interface to real-time detection, novel biocompatible negative LCs should be explored to ensure the solubility and activity of biomolecules in the LC phase while enabling electric field-assisted signal amplification.