1. Introduction

Image contrast is critical to many fields such as microbiology, which studies biological samples that can be as tiny and thin as a single cell. A major problem in visualizing cells is that they are nearly transparent (phase objects), which makes cells difficult to be observed using conventional microscopes. Approaches to image biological objects require the samples to be stained, and thereby converted to an amplitude object. However, staining has plenty of drawbacks: 1) it is invasive, because staining chemicals may alter the structure of the object being studied; 2) sample preparation is time-consuming and requires experienced personnel to perform it; 3) staining chemicals typically have a limited duration of efficacy (minutes), and 4) identifying the appropriate fluorophore to enhance a specific area of interest requires years of research. One more important requirement in microscopy of living organisms is that low illumination intensities are needed to avoid damaging the sample [1].

Current approaches on imaging use active optical elements such as spatial light modulators (SLMs) [2], piezoelectric transducers [3], and nonlinear optical materials to adjust the contrast in a Zernike configuration. In this manner, not only the adjustment of contrast is allowed, but the images can also be processed to enable quantitative phase microscopy (QPM). Some are based on the phase-contrast method reported by F. Zernike that modulates the zero-order of the spectrum at the Fourier plane [4]. That approach to QPM leverages phase-shifting techniques [5] from two-beam interferometry to provide quantitative phase information of living organisms with applications, including in-vitro fertilization [6] and live stem cell study [7]. QPM techniques in common-path configuration, such as the Zernike approach, have shown higher stability for phase measurements over a long time period [8,9].

The drawback in [2] is that it requires expensive and pixelated SLMs to enable such measurements. SLMs require careful calibration and provide discrete phase retardation, which reduces the dynamic range of image contrast, and the pixelation effect induces diffraction effects in the image.

The system proposed in Ref. [3], uses a piezoelectric transducer in a customized twin concentric mirror system. This is a fast and low cost-system (20% of the SLM cost) compared with the use of SLMs. However, the visibility of this system is limited because the mirror's dimensions are not optimized for common-path configuration (no plane reference wavefront). The accuracy and visibility of [2,3] are limited because these methods are constrained to the assumption of plane reference.

Another approach to visualizing phase objects consists of the use of nonlinear optical materials. Various nonlinear optical materials have been proposed as Zernike filters, including azo-dye doped liquid crystals [10], and other crystalline materials [11,12]. Through different mechanisms, the nonlinear material absorbs the intensity at the zero-order of the spectrum, which modulates the material's refractive index. Consequently, phase modulation of the zero-order converts, under certain circumstances, phase into intensity. In our previous work [13], we reported quantitative phase measurements using azo-dye-liquid crystal materials and, in this manner, avoiding the use of expensive SLMs. This approach enabled an optimum phase contrast system with high accuracy measurements [14]. However, the intensity required to induce the phase filter was 10 mW/cm². Even though this intensity is four orders of magnitude below the intensity required in confocal microscopy [15], it is still high for the study long time cell monitoring [2].

In this manuscript, we propose the use of an azobenzene liquid crystal (LC) to enable prolonged cell monitoring and real-time adjustment of the contrast. This material has recently reported an outstanding optical nonlinear response (n₂=0.2cm²/mW) [16]. The nonlinear coefficient of azobenzene LC is four orders of magnitude larger than the previously used azo-dye-doped liquid crystals [17,18]. Our research interest in the use of this LC material is that the nonlinear effect can be activated under very low illumination intensities (below 1 mW/cm²). That opens the door for applications in microbiology where very long time periods (days or weeks) of observation are necessary to study samples without damaging them. This will enable long time monitoring of biological samples that can potentially be implemented in personalized medicine. Besides that, we are also interested in the anisotropic behavior of the azobenzene liquid crystal. We show that the anisotropy of this material will enable to control the visibility of the contrasted images by rotating the polarization of the illumination. This capability can be leveraged in label-free microscopy not only to study samples with low visibility but also, to adjust contrast in real-time when necessary with no digital processing or preparation and staining new biosamples. Additionally, having the ability to obtain various phase-contrasted images makes this material a great candidate for future implementations of quantitative phase microscopy in common-path configuration.

This manuscript is organized as follows. In Section 2, a brief description of the azobenzene liquid crystal material and the extraordinary nonlinear optical phenomena due to photoisomerization is presented. Section 3 shows the experimental setup used to obtain adjustable contrast in an imaging system. Exemplary results of contrasted images and the calculation of contrast are shown in Section 4. The results are discussed and summarized in Section 5, and potential applications of this work are presented.

2. Azobenzene liquid crystal materials

Liquid crystal materials have been employed in various applications, including displays and SLMs [19]. In both cases, LC displays and SLMs are arrays of pixels where either amplitude or phase is electrically controlled. Besides the high cost of SLMs, dealing with pixelated optical devices represents many challenges—for instance, the loss of intensity illumination due to diffraction effects. Consequently, the sample in QPI systems must be illuminated with a higher intensity, which is not desirable in bioimaging.

Recent reports of the use of LC materials in optical elements that are continuous (non-pixelated) have gained the interest of the optics community. One of them is the LC cell reported in [20], where the wavefront is controlled by the use of spatially distributed transparent electrodes. Another example of the contribution of LCs is the construction of geometric phase elements, including lenses and gratings [21,22]. Thanks to the LC anisotropic behavior, it is possible to build optical elements where light can converge and diverge simultaneously depending on the polarization illumination. This remarkable performance has applications on the construction of compact optical systems, and it is impacting optical metrology [23,24].

However, LCs have also been used in imaging as self-induced phase filters in phase-contrast techniques [10,13]. In this approach, an azo-dye-doped liquid crystal cell was placed at the Fourier plane of a 4f system, as shown in Fig. 1. The phase modulation mechanism in this material is photoisomerization; when illuminated, the azo molecules are converted from trans- to a cis isomer that results in a torque applied to the adjacent LC molecules [18,25].

 

Fig. 1. a) Phase-contrast imaging using an azo-dye-doped LC cell as a phase filter to convert the object's phase information into intensity. b) Triggered by light, azo molecules transition from a trans- to a cis isomer (photoisomerization) to apply a torque to the LC molecules, and hence, modulate the refractive index of the zero-order.

Download Full Size | PDF

The generalized phase-contrast theory for common-path interferometry provides a mathematical framework to explain the conversion from phase information to intensity [26]. The intensity distribution I_n, at the camera plane is given by Eq. (1) [13].

In the approach proposed here, no SLM is required. We use a 25X21 mm² liquid crystal cell as a phase filter. We harness the optical nonlinearity of a liquid crystal material to self-modulate the undiffracted light at the Fourier plane. Furthermore, the birefringence can be used to adjust the contrast and obtain n intensity measurements. In other words, the phase-contrast filter is photoinduced, and the contrast is controlled via polarization. The nonlinear optical material in this experiment is the azobenzene LC, 4955 by Beam Co [17]. It possesses one azo component to each LC molecule, which increases its nonlinearity. The mechanism of phase modulation is also photoisomerization. However, the nonlinear effects are four orders of magnitude larger than azo-dye-doped-LC. The nonlinear coefficient of azobenzene LC material is n₂=2.1X10⁻¹cm²/W, for 532 nm wavelength, intensity = 4.4X10⁻⁷ W/cm², parallel light polarization respect to the molecular orientation, and cell gap of 10 mm. The optical anisotropy (the difference between the principal values of the refractive indices of the LC) reported by the manufacturer, is 0.20 at 633 nm wavelength. These specifications are the premise of accomplishing the goal to acquire phase contrasted images with low-intensity illumination and where contrast can be adjusted via polarization.

3. Real-time, label-free microscopy using azobenzene liquid crystals

The experimental setup is shown in Fig. 2. The illumination source is a vertically polarized He-Ne laser at 633 nm that is attenuated by a combination of neutral density filters. Polarization is controlled by means of a half-wave plate and monitored by a polarimeter. The light was collimated (3.7 mm beam diameter) to illuminate the phase object and to cover the microscope objective's aperture (2.5 mm diameter). As a phase object, we used a traceable AFM (atomic force microscopy) transparent microscopic binary phase mask to validate the experiments. The imaging system consists of a 20X microscope objective and a 10X eyepiece. The liquid crystal material was placed in a 25X21 mm² cell with a 3.3 microns gap. To properly leverage the optical anisotropy, the LC cell had a homogeneous alignment, meaning that the LC molecules are parallel to the substrate. It is important to emphasize that no voltage or temperature control was required. The phase filter was photo-induced at very low intensities, and the contrast was controlled via polarization due to the birefringence and molecular alignment. As mentioned earlier, the nonlinear material used was the azobenzene liquid crystal 4955 by Beam Co. The images were captured by a FLIR Blackfly CCD camera of 5.86µm pixel pitch (Model: BFLY-PGE-23S6M). The camera was placed at two times the focal length of the eyepiece. We used a kinematic mount and a linear stage to ensure the camera was positioned at 2f. The criterium was to focus the image of the phase object by minimizing visibility. The half-wave plate was rotated to control the polarization from 0 to 180° respect to the molecular alignment of the LC cell.

 

Fig. 2. Phase-contrast optical imaging using an azobenzene LC cell. The direction of polarization and molecular alignment used in the experiment is shown at the left top.

Download Full Size | PDF

4. Adjusting the contrast via polarization

According to Eq. (1), the intensity distribution I_n, in a phase-contrast system depends on the phase modulation of the undiffracted light θ_n, i.e., the contrast can be adjusted by θ_n. In this approach, we leverage the birefringence of the azobenzene LC material to obtain n intensity measurements at the camera plane by varying the direction of polarization. In other words, by rotating the polarization of illumination, the wave travels through a different refractive index and is therefore modulated by a different phase shift, hence modifying the contrast. Figure 3 shows examples of different contrasted images of the binary phase object (focus star to 400 nm pitch, 350 nm height) described in Fig. 3(a). Black values are considered the background (0 nm), and white values represent the object (350 nm respect to the background). Figures 3(b) and 3(c) were obtained at 0 and 90° polarization angle with respect to the liquid crystal cell alignment.

 

Fig. 3. a) Binary phase object of a focus star height to 400 nm pitch. Black values are considered the background (0 nm), and white values represent the object (350 nm respect to the background). b) and c) Phase information converted to the intensity at 0 and 90°, respectively, between the molecular alignment and the direction of polarization.

Download Full Size | PDF

These images were obtained in an imaging system that operates at very low-intensity levels, i.e., the intensity in the phase object was 0.95 mW/cm², which is below the maximum intensity required to study biosamples to avoid photodamage [1]. The intensity of illumination was calculated by measuring the power at the object plane (0.102 mW) and estimating the area of illumination by means of a camera at the object plane (diameter = 630 pixels • 5.86µm pixel pitch = 0.37 cm).

The contrast values C in the captured images were calculated using Eq. (2).

 

Fig. 4. a) Plot of the average grey level of the background and object $\overline {{I_{BG}}} $ and $\overline {{I_{OBJ}}} $, respectively, as a function of the direction of polarization used to illuminate the sample. $\overline {{I_{BG}}} $ is shown in red (diamond) and $\overline {{I_{OBJ}}} $ in blue (circle). b) Resulting contrast between $\overline {{I_{OBJ}}} $ and $\overline {{I_{BG}}} $. Notice the inversion in contrast going from negative to positive values. These results were observed using 0.95 mW/cm² illumination intensity and varying the direction of polarization from 0 to 180° respect to the molecular alignment.

Download Full Size | PDF

5. Summary and discussion

We investigated the use of an azobenzene liquid crystal material due to its optical anisotropy and very high nonlinear optical behavior. The large nonlinear optical properties of this material allow for very low-intensity illumination in the phase-contrast configuration. Table 1 provides a comparison of various phase-contrast imaging systems where active Zernike filtering is implemented.

Table 1. Comparison of active imaging systems in common-path configuration with controllable phase contrast.

View Table

The contrast can be adjusted via polarization due to the birefringence of azobenzene LC materials. Figure 4(b) confirms that its optical anisotropy provides inversion of contrast going from -0.2 up to 0.3. These remarkable results were obtained with a 3.3 microns cell gap, and the object was illuminated with 0.95 mW/cm², which is far below the intensity required to observe biological samples. At the current stage, the intensity in the SLM approach is lower, the intensity needed to illuminate the object can be further reduced, and the dynamic range of the contrast can be increased with an LC sample with a larger gap. The use of an azo-benzene LC cell is very inexpensive compared to other active optical elements (∼0.5% the cost of an SLM). The approach presented here does not require to apply electronic or temperature controls. It is also important to remark that the LC cell approach is one of the handful phase contrast systems with optimized η parameter [14].

The contrast inversion can be leveraged in various manners. The direction of polarization can be used to adjust the contrast in real-time. Even though the liquid crystal's time response is in the order of milliseconds, in the setup, once the phase has been modulated in the LC sample, the visibility can be adjusted via polarization due to the optical anisotropy, enabling real-time visualization. Another remarkable application is that this imaging system opens the possibility to enable a fast differential phase contrast to study the biological sample further. Moreover, this system can be implemented in an optimized phase contrast microscope to enable accurate quantitative phase imaging. The advantage of low intensity is twofold: (1) allows continuous monitoring of samples over prolonged time periods and (2) allows using additional channels (for tomography or fluorescence microscopy) without damaging the biosample.

In conclusion, the approach presented in this manuscript is significantly cheaper than state-of-the-art approaches; the contrast can be adjusted in real-time with a large and continuous dynamic range in the phase retardation. Notably, it works under very low-intensity illumination to avoid photodamage [1], and it does not require any post-processing. Our results confirm that azobenzene liquid crystal material is a potential candidate to be used in phase-contrast imaging, differential phase-contrast, quantitative phase imaging, and long-time cell monitoring required in personalized medicine.