Biomimetic optics: liquid-based optical elements imitating the eye functionality

Abstract

The optical systems mimicking the eye functions are of great importance in various applications including consumer electronics, medical equipment, machine vision systems and robotics. This optics offers advantages over traditional optical technologies such as the superior adaptation to changing conditions and the comprehensive range of functional characteristics at miniature sizes. This paper presents a review on the recent progress in the development of human eye-inspired optical systems. Liquid-based and elastomer-based tunable optical elements are discussed with the focus on the actuation mechanism, optical performance and the possibility of integration into artificial eye systems.

This article is part of the theme issue ‘Bioinspired materials and surfaces for green science and technology (part 3)’.

1. Introduction

Miniature optical elements have become an integral part of modern optical communication and navigation devices, sensors and modern imaging systems, including medical diagnostic instruments, laboratory on chips, smart cameras and optical systems for microsatellites. Miniature autonomous optical systems have recently gained great importance both for civilian applications and for the needs of the army. The main requirements for these systems are compactness with maintaining a high degree of functionality, the ability to reconfigure performance in real time to complete any task, easy and fast adaptation to the environmental changes by analogy with easily reprogrammable microelectronic devices as well as the ability to integrate with microelectromechanical systems.

However, current conventional adaptive optical systems are not yet as flexible, reliable and mobile as electronic counterparts. The miniaturization of solid-based optical systems inevitably leads to reducing performance quality and decreasing the range of functional characteristics. The mechanical approach, traditionally used to tune the performance of the optical systems, involving a movement of the fixed-focus lenses along the optical axis or a change in the relief of the piezo-membrane mirrors has drawbacks associated with low reliability of moving parts due to increasing surface friction on microscale, slow response, high power consumption and high prices.

The best solution to these challenges is the development of alternative optical systems mimicking functions of the organs of vision of living species such as the eyes of vertebrates (the human eye) [1–4] and the compound eyes of insects [5–7]. Evolution made biological vision systems compact, anatomically simple with a wide range of tunable functionality and the ability to adapt to changing environment using several distinctive mechanisms and reactions [8,9]. By imitating the accommodation function of the human eye, it becomes possible to create the eye-like single miniature lenses, which replace a complex system of solid-state lenses due to the possibility of focal length tuning by the lens reshaping [10–11]. The implementation of optical elements inspired by the insect's compound eyes allows for creating compact optical detectors with a wide field-of-view [7].

Despite small aberrations, high sensitivity and the wide field-of-view determined by the eye structural features, the compound eyes of insects do not exhibit adaptive properties peculiar to the human eye such as the continuous adaptation of the focal distance and the optical aperture, the eye movement (tracking moving objects) and higher angular resolution. In this regard, the human eye attracts more attention of researchers and engineers as the source of inspiration for the development of the optical biomimetic systems featured with superior functionality and adaptability.

This paper presents a review on the recent progress in the development of human eye-inspired optical systems with reference to pioneering works. The two types of tunable optical elements based on using liquids and elastomers are discussed with emphasis on structural characteristics, actuation mechanisms, optical performances and integration into an artificial eyeball system.

2. The eye-inspired liquid-based optical elements

(a). Structure of the human eye optical system and adaptive functions

The optical structure of the human eye consists of four refracting tissues that form the image on the retina located on the back of the eyeball [9] (figure 1). The cornea is responsible for the focusing power of the eye and has a fixed focal distance. To obtain a clear image, the pupil area and the crystalline lens tune the optical information passed through the cornea. The pupil is an aperture of the iris in the centre of the eye. The iris is the muscle adjusting the pupil diameter to control luminous flux coming to the light-sensitive tissue. This effect is known as the pupillary light response. At high intensity of light, the pupil shrinks in diameter to decrease the light flux for protecting the sensitive retina from damage. When the light intensity is low, the pupil expands allowing the eye to receive the maximum possible amount of light. The pupil diameter normally varies from 2 to 4 mm in bright light to 4 to 8 mm in the dark. The crystalline lens is a transparent bi-convex structure allowing the eye to adjust the focal distance on objects located at different distances. This function is known as accommodation. According to [12], the accommodation response to a focusing stimulus takes 0.283 s with a standard deviation of 0.098 s. A change in the curvature radii of the lens is controlled by contraction of the ciliary muscle via zonules tissue. Additionally, one of the important adaptive functions of the human vision system is an ability to track moving objects to stabilizing the image of the object in the area of maximum visual acuity on the retina [13]. This effect is achieved through voluntary eye movements alone in discrete and quick mode and in slow and smooth movement mode. Relying on the pupillary light reflex, accommodation and the eye movement reflex, the vision system demonstrates excellent adaptive functions and responses to dynamically changing environment.

Figure 1.

Schematic view of the human eye designating components of the optical system (cornea, pupil, iris, zonula, ciliary body, retina). (Online version in colour.)

Liquid-based optical components are promising candidates for the creation of the compact optical systems mimicking the eye's reflexes in contrast with traditional solid-state counterparts due to the possibility of flexible variation of optical characteristics and compatibility. A number of liquid properties such as fluidity, surface tension, and thermal and mass diffusivity allow for fast and easy reconfiguration of interfacial geometry, creation of smooth refractive surfaces on the molecular scale and dynamic change of the refractive index via mixing of liquid components.

The idea of using liquid as an optical element is not new and goes back to the work of S. Gray (1695) [14], who proposed the construction of a simple microscope in which water droplets placed in holes of different diameters in a plate served as bi-convex lenses with different curvatures and, accordingly, with different magnifications. Images obtained with the lenses were of high quality due to the ideally smooth free surface of the droplets. In 1850, the astronomer E. Capozzi proposed using a rotating vessel with mercury to create a parabolic mirror for telescope [15]. The adjustment of the focal length was carried out by changing the angular velocity of rotation of the vessel. The main problem in the operation of such mirrors was the influence of gravity, which limited their spatial orientation. However, this problem was later solved, and today the paraboloids are exploited for astronomical observations.

In recent years, the eye-inspired liquid-based optical elements have been applied in medicine [16–18] and consumer electronics [19–21]. The most common approach for manipulating the optical characteristics of the liquid-based optical elements is the change in the shape of the refractive interface between two immiscible phases (liquid–air and liquid–liquid) with different refractive indices. Less commonly, the variation of the refractive index of a bulk liquid caused by heating [22,23] or changing the concentration of components [24] is used.

The methods for controlling the shape of the refracting liquid interface are based on applying external stimuli such as electric fields, causing electrowetting [19–20,25–35] and dielectrophoresis [36–41] effects, thermal fields [42–44], pneumatic and/or hydraulic pressure [45–47], pH change [48], optical irradiation [49] and acoustic pressure [50,51]. All these methods have advantages and disadvantages associated with features of the optical element design, implementation of optical characteristics control, quality of obtained images and the response time of an optical system to applied external stimuli [10,11].

(b). Electrically actuated biomimetic liquid lenses and irises

Electrical effects such as electrowetting and dielectrophoresis [19–20,25–41] are the most widely used for tuning optical parameters of individual optical elements such as liquid lenses and apertures.

A reversible change in the contact angle is an effective way to control the shape and position of a liquid droplet on a dielectric substrate or an interface between two immiscible conductive and non-conductive liquids in contact with the dielectric wall of a cell. One of the powerful driving mechanisms for implementation of this approach is the effect of electrowetting on dielectric (EWOD), which is based on the action of surface tension forces [25]. The electric voltage applied between the substrate coated with a transparent electrode and a transparent insulating film and the droplet of electrically conductive liquid sitting on the substrate leads to a change in the surface energy at the liquid–solid interface; as a result, the contact angle reversibly changes (figure 2a). The relationship between the contact angle and the applied voltage is described by the well-known Lippmann–Young equation

$$\cos \theta (U)=\cos {\theta }_0+\frac{{\epsilon }_0\epsilon }{2\gamma \delta }{U}^2,$$

where ${\theta }_0$ and $\theta (U)$ are the contact angles in the voltage-off and the voltage-on states, respectively, U is the voltage applied across the insulating film, γ is the interfacial tension between the conductive liquid and the surrounding medium, ε₀ is the permittivity of vacuum, and ε and δ are the dielectric permittivity of insulating layer and its thickness, respectively. The change in the contact angle results in a change in the radius of curvature of the liquid interface and, as a consequence, the focal length and the base diameter, figure 2b.

Figure 2.

(a) Electrowetting principle of the droplet reshaping process on a hydrophobic insulating film. In electric voltage-off state, the droplet contact angle is θ₀ > 90°; in voltage-on state, electric charges change the liquid-surface free energy, resulting an increase in wettability, θ < θ₀. (b) An electrowetting-actuated liquid droplet as a tunable-focus liquid lens, ΔF is a range of the focal length variation. (Online version in colour.)

At present, there are two approaches to the designing of eye-inspired liquid lenses exploiting electrowetting: (i) the sessile droplet of electrically conductive liquid surrounded by air, figure 2, and (ii) the system of two immiscible liquids placed in the miniature housing, figure 3, where one of the liquids is polar and electrically conducting (aqueous salt solutions), the other is non-polar, electrically insulating (typically non-polar oils).

Figure 3.

(a) Cross-section view of the electrowetting-actuated liquid lens consisting of the two immiscible (electroconductive and insulating) liquids in a cylindrical hydrophobic tube. In the voltage-off state (θ > 90°)—the diverging lens; in the voltage-on state (θ < 90°)—the converging lens. (b) Images of the shape of the interface between water/oil in a 6 mm diameter glass housing at applied voltage equals 0 (a convex interface) and 120 V (a concave interface) taken from [19]. (Online version in colour.)

In the first case, the droplet serves as a plano-convex microlens with a variable focal length in a small range. According to [26] for a 6 µl droplet of aqueous solution of potassium sulfate (0.1 M), the focal length tuning achieved about 20% of its initial value. However, it should be noted that the use of a network of transparent electrodes coating the substrate allows for displacement of the droplet lens in the plane perpendicular to its optical axis by 35% of the droplet base diameter [26]. The effect of the droplet shifting in this case mimics the eye movement reflex, which is a significant advantage of this approach. However, this design has a disadvantage caused by sensitivity to gravity, which limits the spatial orientation of the optical system and the size of the lens. Since electrowetting relies on surface forces, the droplet size is limited by the capillary length, $a=(\gamma /\mathrm{\Delta }\rho g{)}^{1/2}$, which is in the range of 1–3 mm for most liquids. Here, $\mathrm{\Delta }\rho$ is the density difference between the liquid lens and the surrounding medium, and g is the gravitational acceleration. In addition, the control electrodes deposited on the substrate in the light path may cause light scattering and lower the image quality.

The design of the liquid lens consisting of two immiscible liquids [19,27–32] avoids the drawbacks mentioned above. To enable the lens operation in any orientation relative to the direction of gravity force, the liquids are chosen to have the same density. The electrodes are located in the wall of the housing and do not influence the light flux. In this case, the interface between two liquids with different refractive indices serves as a refracting surface. This liquid lens provides the ability to operate as both the converging and the diverging lens by changing the sign of the interface curvature when applying the control voltage between the hydrophobic wall and the electrically conductive liquid (figure 3). The focusing range achieved in this case provides optical characteristics that are close to those of the eye. Note that the refractive interface is a spherical lens, which usually produces optical aberrations. In order to eliminate the image distortion, the position-dependent profile of the electric field is applied for creating an aspherical profile of the interface between two liquids [31]. To increase the field-of-view of such lenses, combined approaches inspired both by the adaptive capabilities of the human eye and by the structural features of the insect eye have been proposed [28,29]. In this case, the electrowetting-based lens is located on a flexible membrane, the curvature of which is changed by applying external pressure [28].

The dielectrophoresis is used as a driving force to tune the optical parameters of systems consisting of two immiscible dielectric liquids having different values of the dielectric constants and the refractive indices [36–40]. Contrary to the electrowetting effect, the dielectrophoresis results in the volume forces and does not require the use of electroconductive liquids and providing the contact of the electrode with the liquid. As in the case of EWOD, dielectrophoresis allows for two configurations of liquid lenses: (i) a droplet of dielectric liquid on a glass substrate coated with concentric transparent electrodes, and (ii) a droplet of a dielectric liquid (ε₁) surrounded by another liquid with a different dielectric constant $({\epsilon }_1\ne {\epsilon }_2)$ sealed in a space between two glasses with concentric electrodes. In the voltage-off state, the droplet takes on an equilibrium shape with an initial contact angle. By applying the electric voltage, the droplet shrinks or expands depending on the direction of the dielectric force, which is determined by the ratio between ε₁ and ε₂. According to [52], the contact angle changing can be estimated using the following equation

$$\cos \theta (U)=\cos {\theta }_0+\frac{{\epsilon }_0({\epsilon }_1-{\epsilon }_2)}{2\gamma \delta }{U}^2,$$

where ε₁ and ε₂ are the dielectric constants of the droplet and the surrounding liquid, respectively, and δ is the penetration depth of the electrical field.

A model of a single-liquid lens actuated by dielectric forces was first shown in [36]. The droplet (2 µl) of a high dielectric constant (ε₁ = 16.5) liquid crystal MDA2625 (Merck) surrounded by air (ε₂ = 1) was placed on a substrate covered with transparent electrodes insulated with a hydrophobic film. By applying electrical voltage, the focal length increased from 1.6 to 2.6 mm that corresponded to the contact angle decrease of 24°. However, the use of liquid crystals as lenses is restricted by the birefringence effect related to the temperature-dependent phase state of material. Meanwhile, this approach has great potential, since any low-volatile liquid with a dielectric constant higher than that of air can be used as a liquid lens. In [38], a two-liquid variable focus lens actuated by the dielectric force was reported. The lens consists of a low dielectric constant droplet (transparent oil SL-5267, ε₁ = 4.7) and a high dielectric constant surrounding liquid (glycerol, ε₂ = 47). The focal length of the liquid lens (the base diameter of 3 mm) was changed from 34 mm in the voltage-off state to 12 mm at the maximum applied voltage of 200 V. A common drawback of both configurations is the hysteresis of the contact angle, which can be significantly reduced by an optimization of the electrode parameters and the characteristics of the electric voltage [37].

Electrically driven compact iris apertures imitating the pupillary reflex of the eye [32–35,40,41] consist of a circular cell filled with the two immiscible liquids, which are typically a transparent oil surrounded by aqueous solution of opaque ink [33,35,40] or by liquid crystal doped with ink [41]. Both liquids have closely matched densities to avoid disturbances of the shape due to gravitational and vibrational impacts. The opaque water solution takes on a ring (a doughnut-like) shape due to adherence to the wall of the cell. Lids of the cell are coated with a system of radial or concentric transparent electrodes, providing the ability to control the aperture (figure 4). When the amplitude of electric voltage changes, the diameter of the opaque liquid ring is tuned due to the wetting/dewetting mechanism for the case of electrowetting or due to the displacement of liquids under the action of volume forces for the case of dielectrophoresis. It is worth noting that in most cases the tuning performance of irises, $T=(\mathrm{\Delta }D/D_{\text{max}})\times 100\mathrm{\%},$ where $\mathrm{\Delta }D=D_{\text{max}}-D_{\text{min}}$ is the difference between maximal and minimal diameters of the aperture, ranges from approximately 60% [32,33,35,40] to 80% [34]. This means that the iris aperture remains open regardless of states (the voltage-off or the voltage-on) and does not allow for completely shutting off the light flux. Although the latter is a restriction in terms of technical applications, nevertheless the values of the tuning ratio of liquid irises are close to that of the human eye (around 50%).

Figure 4.

Electrowetting-tuned liquid iris consisting of water solution of ink/transparent silicon oil. Cross-sectional schematic view on the left, top view images of the iris on the right. (a) Minimum aperture (voltage-off state, 2.3 mm in diam.); and (b) maximum aperture (65 volts applied, 6.1 mm in diam.). Rearranged from [35]. (Online version in colour.)

One of the advantages of the tunable optical elements actuated by electrical forces when compared with elements actuated by other mechanisms, e.g. temperature or pH-sensitive hydrogels [42,48,49], is fast response ranging between 1 and 100 ms [28,30], which is significantly faster than the accommodative response of the eye. In addition, the use of electric fields is a simple in implementation and highly precise method to control the optical characteristics of the liquid elements.

(c). The eye-inspired liquid optical elements actuated by laser-induced Marangoni forces

Despite a large number of studies on eye-inspired liquid optics, the vast majority of them focus on studying either the adjustment of the focal length or the aperture size. At the same time, some works [53–56] demonstrate miniature liquid optical elements that are capable of performing adaptation functions peculiar to biological vision systems such as the accommodation, the pupillary reflex and the eye pursuit movement.

The actuation of such elements is based on the generation of thermocapillary and solutocapillary forces (also known as Marangoni forces) on the liquid interface, induced by the thermal action of the laser irradiation [57,58].

The Marangoni effect [59] is the liquid flow caused by a gradient of surface tension, which in turn can be induced by a non-uniform distribution of temperature and/or concentration of components on the liquid interface. If the characteristic scale of the liquid system is smaller than the capillary length, the flow can lead to a deformation of the liquid interface. When non-uniform heating of a thin layer of a pure liquid, thermocapillary flow occurs [60], which transfers the liquid from the heated to cold area; as a result, the layer becomes deformed. In 1958, Block & Harwit [61] first showed that a free surface of a thin liquid layer arbitrary curved by thermocapillary forces can be used as an optical element. Later, Loulergue & Xu [62] suggested using a two-layer system of immiscible liquids deformed with a spatially modulated infrared (IR) laser beam as a prototype of sensitive IR visible image converter.

In the case of local heating of the layer with the laser beam, the thermocapillary concave deformation in the layer is formed in the heating area, which can serve as a concave liquid mirror [63,64]. It was shown in [64] that the focal length of the thermocapillary mirror, induced by the laser beam in the thin layer on the light-absorbing substrate, can be tuned over a wide range for up to five times by changing the laser power or the layer thickness.

Significant possibilities appear when using the two-layer system of immiscible liquids actuated with the thermal action of the pump laser beam. A bottom layer is used to supply thermal energy into the system by absorbing irradiation of a pump laser and to form an ideally circular and easily managed aperture owing to the slip condition and the absence of roughness at the liquid–liquid interface. To absorb the pump laser beam, the bottom layer is doped with the proper dye, which is transparent for an optical signal. A top layer is coloured in such a way to absorb completely the optical signal. The operating principle is based on creating a circular thermocapillary rupture of the top layer by thermocapillary forces induced with the pump laser beam. The thermocapillary rupture serves as an iris aperture [53] (figure 5a). A distinctive feature of the system is the convex deformation of the bottom layer in the rupture area that is caused by the combined action of the surface tension forces and hydrostatics (figure 5b). This two-layer system mimics functions specific to the eye such as the change in the aperture diameter, the accommodation of the focal length and the smooth pursuit eye movement. Comparing with the structure of the optical part of the eye (figure 5b), it becomes apparent that the rupture area in the top layer serves as the pupil controlling the incoming light flux, the top layer itself is the iris muscle and the convex deformation of the bottom layer as an analogue of the crystalline lens. We showed in [53] that at a fixed power of the pump laser beam, the rupture diameter and the focal length of the convex deformation are varied with the thickness of the top layer. In the experiments [53], the two-layer system was created in a transparent polystyrene dish (60 mm in diameter) using two immiscible liquids. Given that the thermocapillary rupture occurs in the layer with the free surface, the system was open at the top (figure 5a). Such configuration of the optical element is sensitive to horizontal tilt and vibrations; therefore, it can only operate in a horizontal position. To reduce the influence of vibrations, a highly viscous and heavy liquid was chosen to form the bottom layer. An important requirement of the liquids used is low vapour pressure. Glycerol coloured with Brilliant green to absorb the pump laser beam (He–Ne laser, 17 mW, wavelength of 632 nm, beam diameter of 1.5 mm) was used as the bottom layer of 2 mm thick. Hexadecane dyed with Oil Red O to absorb the optical signal (the solid-state laser, wavelength 532 nm, the beam diameter expanded to 50 mm) was used as the top layer of thickness varying from 600 to 300 µm. Transmittance of the glycerol layer at a wavelength of 532 nm was measured to be approximately 90%. In the heating area, the temperature increased by 20°C. The two-layer system operates in the long focal length range from 4 to 25 cm. However, the aperture diameter takes on values from approximately 1 to 11 mm that precisely fit the eye pupil performance. Note, that unlike most liquid-based irises operating by electrowetting and dielectrophoresis mechanisms, the laser-induced thermocapillary mechanism allows a 100% tuning ratio.

Figure 5.

(a) Schematic sectional view of the two-layer laser-induced thermocapillary iris: (i) the pump laser beam-off state—the iris closes and (ii) the laser beam-on state—the opaque layer ruptures, the iris passes the optical signal. (b) A comparison between the structure of the tunable part of the human eye and the bioinspired two-layer optical system. (c) Images of the aperture adjustment to a new position of the pump laser beam (shown by a circle) shifted for the aperture radius. From [53]. The top layer is 500 µm thick. A dashed circle in the last image shows the previous position of the aperture. Scale bar, 3 mm. (Online version in colour.)

A competitive feature of the proposed two-layer optical system is an ability of the pupil/crystalline lens area to follow closely a moving, in the horizontal plane, laser beam. This effect allows for replicating the smooth pursuit eye movement reflex. Figure 5c demonstrates a series of snapshots of the aperture moving to the pump laser beam, which was shifted on the distance equal to the radius of the rupture. When the pump beam is shifted, the aperture starts to move to a new position of the beam and self-centres onto the beam axis. The mechanism behind this displacement is similar to that of the thermocapillary motion of air plugs in thin gaps described in our previous work [65]. The pupil/crystalline lens area adjustment time decreases with an increasing top layer thickness. In this case, the pupil has the smallest size and moves without distorting its circle shape, while the large one elongates in the motion trajectory and, hence, takes longer to adjust because of the time used to restore its circular shape. The distortion occurs due to decreasing the Laplace pressure that saves the shape of the contact liquid–liquid line.

Despite the superior capabilities of the two-layer optical system tuned by thermocapillary forces, the response time to thermal stimulus is hardly comparable with typical response times of the eye and varies from 1 to 500 s depending on the thickness of the top layer and the power of the pump laser beam. Moreover, the variation range of the refracting surface shape is limited by deflection from the flat to the convex, by analogy with the thermocapillary mirror [64].

To expand the focal range tuning in the study [54], the thermocapillary effect induced by thermal action of the laser beam was applied for changing the shape of the liquid–air interface of the low-volatile liquid droplet with the pinned contact line (figure 6a). The droplet was placed in a hermetic cell with an aperture of 6 mm confined by two glass substrates. As the lens material, low-volatile liquids which formed sessile droplets on a glass substrate with relatively small contact angles of 10–20° were used. This initial droplet geometry served as a plano-convex lens and made it possible for most effective control of its shape by laser-induced thermocapillary forces. Ethylene glycol and benzyl alcohol were chosen as such liquids, the surface tension of which allowed for creating the surface deformation [54]. Note that low surface tension liquids (e.g. silicone oils) wet the surface and formed thin films, but high surface tension liquids (e.g. glycerol) did not allow for inducing the thermocapillary deformation. To enable the absorption of the laser irradiation (wavelength of 532 nm, beam diameter 0.8 mm), the liquid was dyed with Crystal violet. The droplet (less than 1 µl, the aperture diameter ≈ 3 mm) serves as a simple model of the crystalline lens with a wide range of reversible variation of the focal length in both convergent and divergent modes depending on the light intensity. Figure 6b shows the focal distance variation of the benzyl alcohol droplets of different volumes (0.3–0.9 µl) in response to the power of the laser beam. It has to be noted that time of the focal distance accommodation varies with the laser power in the range from 0.2 to 1 s, which is much faster in comparison with thermal sensitive hydrogel lenses (20–25 s) changing the focal distance sign in response to the thermal stimulus [48].

Figure 6.

(a) Cross-sectional view of the thermocapillary-actuated droplet lens of low-volatile liquid. The laser beam-off state—the plano-convex lens (i); the laser beam-on state—the plano-concave lens (ii). (b) Focal length variation in response to the laser beam power for the benzyl alcohol droplets of different volumes. Vertical lines are the asymptotes. The graph is taken from [54]. (Online version in colour.)

Although the thermocapillary liquid lens demonstrates very competitive adjustment of the focal length characteristic, it does not allow for tuning the aperture and tracking a moving optical signal as in the case of the two-layer system [53].

A droplet lens actuated by the laser-induced solutocapillary forces was proposed in our studies [55,56]. Such a lens can adapt not only optical parameters (the focal length and the base diameter) in response to the intensity of laser irradiation, but mimics the eye movement reflex by tracking the position of the laser beam. As a base for the liquid lens, a thin layer of binary liquid mixture located in a hermetic cell is used. One of the liquids in the mixture has higher surface tension and lower vapour pressure compared to the other liquid. To absorb the laser radiation, a dye is added to the mixture. The local heating of the binary mixture with the laser beam results in a local decrease in the concentration of volatile liquid due to its evaporation that in its turn leads to concentration-related surface tension gradient along the interface. As a result, the solutocapillary flows directed to the irradiated area arise, which eventually accumulate the liquid mixture as a sessile droplet in the laser spot [55–58] (figure 7a). Under continuous irradiation, the droplet saves its shape due to equilibration with its vapour, but turning off the laser beam leads to relaxing the droplet to the thin layer. The stepwise increase/decrease in the power of laser beam changes reversibly the shape of the droplet lens and, as a consequence, the focal distance and the aperture.

Figure 7.

(a) Simplified schema of the laser-induced solutocapillary actuation principle of the liquid lens. In the laser-off state—the flat layer with an infinite focal distance; in the laser-on state—the droplet serving as the plano-convex lens. (b) Side-view images of the droplet reshaping versus the power of laser beam (50%wt ethanol–ethylene glycol mixture), from [56]. (c) A reversible variation of the droplet aperture (50%wt. mixture) in response to increasing/decreasing the laser beam power, from [56]. (Online version in colour.)

Figure 7b represents side-view images of the droplet reshaping caused by the changing the power of laser beam. The reversible change to the base diameter of the droplet lens in response to increasing and decreasing the power of the laser beam for three cycles is shown in figure 7c. The working liquid is 50% mixture of ethylene glycol and ethyl alcohol (1.5 µl of volume) slightly dyed with Crystal violet to absorb the laser irradiation (532 nm, beam diameter 0.8 mm). The range of the focal length tuning depends on the concentration of components. In the case of a mixture consisting of 70% of ethyl alcohol, the focal length varied from 39 to 14 mm, but for a 50% mixture, it takes on values from 35 to 17 mm when the power changes from 2 to 30 mW. Figure 8a shows optical images of the test-object (a grid with 400 µm size cells) obtained with the droplet lens at different powers of the laser beam. Aberrations of the images are insignificant. Interestingly the focal length variation diapason obtained with this liquid lens covers the accommodation range of the eye. Note also that in our earlier study [55] of the water–ethanol mixture controlled by a high power UV source, the focal length changed 10 times. The response time of the solutocapillary lenses takes values from 5 to 30 s depending on the power of the laser and the concentration components [56]. This response is slower in comparison with that of the liquid lens tuned by the laser-induced thermocapillary forces [54] and the electrically actuated lenses [28,39]. However, this drawback can be overcome by decreasing the initial size of the droplet and selecting liquids with low viscosity and with a larger difference in surface tension between them.

Figure 8.

(a) Images of the test-object (the metal grid) obtained with the droplet lens (50%wt mixture) at powers of the laser beam of 2, 14 and 30 mW. The side of the square cell is 400 µm. Images from [56]. (b) The droplet motion towards the shifted laser beam (50%wt mixture; the power is 18 mW). Total time is 20 s. An arrow shows the motion direction. The dashed circle in the last image shows the previous droplet position. Images from [56]. (Online version in colour.)

Another feature of the liquid lens controlled by the laser-induced solutocapillary forces is the ability to adjust its position relative to the laser beam on plane of the substrate (figure 8b). This effect replicates the smooth pursuit eye movement reflex. The mechanism behind the droplet lateral adjustment is the action of solutocapillary forces between the irradiated edge of the droplet and its apex. The response time of the droplet lens, required to adjust a shifting equal to its radius, varies in the range of 10–40 s depending on the power of laser and concentration of the mixture components.

The light-actuated liquid elements mimicking the adaptive functionality of the eye are the perspective basis for developing miniature smart optical systems for various applications including medical imaging systems, robotic machine vision systems, microscopy and portable electronic devices.

3. The hybrid optical elements mimic the eye's optical system

Besides the liquid-based optical elements similarly adapting to the changing environments like the human eye, hybrid optical systems can imitate not only the basic eye reflexes but also the architecture of the eye and the actuation principles. An important application of such hybrid systems is the development of biomorphic vision organs (artificial eyes) which can be used in robotic machine vision systems for industry, real-sized eye models for ophthalmological research and humanoid or mobile robots. As a rule, in those systems, the tunable part consists of the lens and/or aperture, which is either a polymer chamber filled with a liquid or an elastomeric body, and the actuation principle in most cases relies on the use of artificial muscles, which server as analogues of ciliary muscles or iris muscles of the eye. The artificial muscles are chemo/thermal/electrosensitive elastomers that contract, strain or swell reversibly upon exposure to external stimuli.

The papers [2,3] present optical elements imitating the crystalline lens and ciliary muscles of the human eye. In contrast with liquid droplets sessile on a transparent substrate and two-layer columns of immiscible liquids discussed earlier, the lens in this case imitates closely the shape of the eye's crystalline lens. The lens is a bi-convex body made of polymer flexible membrane filled with liquid. The accommodation, i.e. the focal length tuning, is performed by applying strain to the membrane along the perimeter of the edge, which leads to a change in the curvature and diameter of the lens by analogy with the crystalline lens. The strain is realized either mechanically by using fastening metal clamps or an artificial muscle—dielectric elastomer actuator controlled by an electric field. It is worth noting that the lens [3] allows almost a twofold change in the focal distance to be achieved with insignificant values of strain, less than 1%. The use of dielectric elastomer for lens actuation [2] demonstrates a range of the focal length adjustment comparable to those of the eye and fast response of about 60 ms, which, according to the authors, is an undeniable advantage, since it allows focusing on fast-moving objects.

Another type of the optical device mimicking the crystalline lens and the ciliary muscles was proposed in [66]. The lens body consists of a layer of electrosensitive, optically transparent non-ionic polyvinyl chloride (PVC) gel sandwiched between the transparent substrate coated with the electrode and the ring copper electrode. In the voltage-off state, the PVC gel swells up in the aperture of the ring electrode forming a lens with a short focus. Applied electrical voltage between those electrodes leads to the appearance of an electrostatic adhesive force, resulting in a reversible decrease in the curvature of the gel body in the aperture of the ring electrode. This effect imitates the accommodation of the eye’s lens, and the ring electrode, according to the authors view, serves as an analogue of the ciliary muscle. The focal length variation (3.8–22.3 mm) and the response time of accommodation (0.64–0.68 s) of that eye-mimicking lens is comparable to that of the human eye. However, the operational values of the electric field are quite large and reach hundreds of volts per millimetre.

In [67], an eyeball-like optical system for a bionic robot eye was proposed. The eye model consists of the cornea and the crystalline lens, imitating the accommodation function and an artificial retina as an optical sensor. A plano-convex solid-state lens serves as the cornea followed by a water-filled flexible polymer chamber connected with an annular piezoelectric actuator. Tuning the focal distance is achieved by convex deformation of the membrane posterior side due to changing pressure in the liquid-filled chamber by displacing the actuator. According to the test results, it is shown that the proposed approach for focusing on objects located on different distances is similar to accommodation characteristics of the human eye, where the cornea possesses maximum optical power, but the shape-variable lens has low optical power. A similar architecture of an eye-like optical system based on the same operating principle was reported in [68]. However, the authors provided the system with an ability to mimic saccadic eye movement by means of mechanically tilting the system at small angles.

A compact biomimetic imaging system combining a tunable polymer lens and a liquid-based iris inspired by the human eye was reported in [69]. The system demonstrates the pupillary light reflex and accommodation by using two different mechanisms of actuation. The control of the aperture of the two-layer liquid iris is carried out by the electrowetting effect reported by the authors earlier [32]. As it was analogous to the crystalline lens, an elastomeric plano-convex body attached to a motorized actuator was used. Applying 10% strain allows for 5.36% focal length changing with an aperture of 5 mm. Later [3], the authors improved the eye-mimicking system by replacing the motorized actuator with an artificial muscle representing thermal sensitive liquid-crystal elastomer (LCE). A tunable part, including a liquid iris based on the electrowetting, the LCE iris and the elastomeric lens, was integrated with a set of solid-state lenses and an imaging sensor chip. This engineered eyeball [3] demonstrates functional capabilities and optical performance close to those of the human eye.

4. Conclusion

The optical systems mimicking the eye functions have drawn the attention of researches due to great potential in various applications including consumer electronics, microscopy, medical equipment and machine vision systems for industry. The advantages of eye mimetic optics are small sizes, simplicity of realization, high level of adaptation to changing environment and wide range of functional characteristics that are not available to traditional optical technologies.

Two main directions in the field of designing eye mimetic optics can be distinguished: the liquid optical elements, including the lenses and the iris apertures, and the elements based on soft and smart materials. The control of optical parameters, such as the focal length and the aperture, as well as the spatial position, is realized by applying external forces (electrical, thermal, optical).

The advantages of using liquid as the base for optical elements are fluidity, flexibility and ideally a smooth surface. In addition, liquids are cheap and easily available materials. The liquid-based elements allow for fast and easy reconfiguration of optical characteristics via modulation of curvature of the interfaces between two phases. As the literature shows, the eye-inspired liquid optics controlled by electrical effects are most widely used in various applications owing to their fastest response to external stimuli, the integrability with other devices and possibility to compensate optical aberrations. A weak point of this method is high control voltage.

Recently, we have demonstrated an alternative method for the manipulation of liquid-based optics using the thermo- and solutocapillary forces induced by thermal action of the laser beam along liquid interfaces. The optical elements based on liquid droplets and thin layers replicate eye adaptive functions such as the pupillary light reflex, the accommodation and the smooth pursuit eye movement, and have a great potential for applications in science, medicine and technology. However, development of multifunctional devices or integration in complex optical systems currently requires overcoming a number of challenges. These problems include (i) unavoidable evaporation of liquids, especially in the case of a solutocapillary lens, affecting the stability of the lens performance, (ii) the use of dyes absorbing in the optical spectral range that reduce image quality and the lens speed, (iii) high sensitivity of the system to tilt to the horizon, which drastically limits its applicability, (iv) the large response times to changes in the beam power that are associated with inertia of the thermal mechanism, and (v) different types of aberrations peculiar to all liquid optical elements.

To design the complex systems that reproduce the entire structure of the eye, for biomedical research and robots, optical components based on liquid-filled and bulk elastomers changing shape upon thermal, chemical or electrical effects are mostly used. Besides, the elastomers not only perform the function of an optical element but also operate as actuators imitating muscles in the eye. The main challenge of that system is sustaining stability and mechanical strength of elastomeric materials over long-term operation.

Data accessibility

This article has no additional data.

Competing interests

We declare we have no competing interests.

Funding

This study was supported by the Russian Foundation for Basic Research (grant no. 17-08-00291) and the Ministry of Education and Science of the Russian Federation (grant no. 3.4744.2017/6.7).