1 Introduction

The ability to control crystallization is a key goal in very different technological areas such as the production of relevant materials and structures, drug manufacturing in pharmaceutical industry, food processing, sensor/characterization techniques involving different biological solutes or in the water desalination or decontamination [1,2,3].

In the last years, photovoltaic optoelectronic tweezers (PVOT) have emerged as a versatile multi-functional tool for a large variety of applications in different fields, such as optical manipulation and trapping of nano-objects [4,5,6], optofluidics [7, 8], plasmonics [9, 10] or biotechnology/biomedicine [11,12,13], to cite a few.

The technique is based on the bulk photovoltaic effect. It is a singular phenomenon that appears in a few crystalline ferroelectric materials (LiNbO3 clearly standing out) when properly doped (mainly Fe). It allows the generation of remarkably high electric fields (1–3 × 105 V/cm) [14, 15] for moderate or low light excitation levels (~ mW/cm2). The photovoltaic electric field extends to the vicinity of the crystal, and this evanescent electric field is responsible for the capabilities of the technique.

However, the action of PVOT in aqueous, or in general in ionic liquids, has remained a challenge because the photoinduced electric fields were rapidly screened by water nearly inhibiting the platform capabilities. Only during the last years, a few approaches have been proposed to extend the applications of PVOT for operation in aqueous media. Most works propose the manipulation of aqueous droplets inside a non-polar medium [8, 16,17,18,19,20] although in a few cases they deal with the direct manipulation of micro-/nano-objects in aqueous solutions [11, 21,22,23] involving the interaction between the objects and the photovoltaic surface. In this latter context, as many physicochemical or biochemical and pharmaceutical processes occur on aqueous liquid–solid interfaces [24, 25], the action on aqueous ionic solutions could be an interesting function for PVOT. For instance, all-optical control of the solubility, material synthesis and/or crystallization processes could enrich the range of PVOT capabilities and its technological impact.

In this work, we have addressed this challenge, investigating the feasibility of achieving optical control of the spatial distribution of salt crystallization on the PVOT substrate from an aqueous dissolution. This would be an innovative function for photovoltaic ferroelectric platforms in aqueous media. Light-assisted spatial micro-patterning of salt crystallization on lithium niobate substrates is demonstrated and characterized. Two methodological approaches/protocols, the so-called sequential and simultaneous operation, are evaluated and compared. The technique is successfully applied to a variety of salts and the viability and flexibility of drawing the patterns by using a moving laser beam is explored. A discussion on the possible physical mechanisms is also included.

2 Physical basis of PVOT operation

As previously mentioned, PVOT are based on the bulk photovoltaic effect, present in ferroelectric crystals such as LiNbO3:Fe (LN:Fe). This effect arises from directional electron photoexcitation along the ferroelectric axis from localized states of impurities, such as Fe2+/Fe3+ or Cu+/Cu2+, which leads to a photovoltaic current. Once displaced, the electrons are trapped in other defects and generate a spatial charge distribution and the corresponding electric field that extends to the vicinity of the crystal [26,27,28]. Figure 1a shows a schematic of this charge distribution and the corresponding evanescent electric field for the crystal orientation used in this paper, usually called z-cut configuration.

Fig. 1
figure 1

Schematics of a the light-induced charge distribution and the corresponding evanescent electric field, b the illumination setup used in the experiments, c the cuvette containing the LN:Fe crystal and filled with water, d the drying process: the crystal covered with a film of saline water is put in contact with an absorbent paper for about 1 s

Full size image

Usually, PVOT operate in air or inside a non-polar liquid medium so that the generated bulk photovoltaic fields are not screened by the surrounding medium. Under these conditions, the buildup of the photovoltaic field EPV is roughly exponential with the time under illumination t and the time constant τPV is inversely proportional to the light intensity I:

$$E_{{{\text{PV}}}} = E_{{{\text{sat}}}} \left[ {1 - \exp \left( { - \frac{t}{{\tau_{{{\text{PV}}}} }}} \right)} \right]$$
(1)

with:

$$E_{{{\text{sat}}}} = \frac{{\gamma [{\text{Fe}}^{3 + } ]l_{{{\text{PV}}}} }}{\mu }$$
(2)

and

$$\tau_{{{\text{PV}}}}^{ - 1} = \frac{\mu e}{{\varepsilon \varepsilon_{0} }}\frac{{sI[{\text{Fe}}^{2 + } ]}}{{\gamma h\nu [{\text{Fe}}^{3 + } ]}}$$
(3)

where γ is the trapping coefficient of acceptors (Fe3+), lPV is the bulk photovoltaic drift length (~ 1–5 Å), μ is the electron mobility and s is the photoionization cross section of donors (Fe2+), h is the Planck constant and ν is the light frequency [29].

However, when PVOT operate inside an aqueous environment the photovoltaic fields are partially or totally screened. A few years ago, a simple theoretical model to describe the time evolution of the photovoltaic electric field when the crystal is surrounded by a leaky medium with a non-negligible electrical conductivity was developed [30]. In that model, the buildup of the photovoltaic electric field inside the crystal is given by:

$$E_{{{\text{PV}}}} (t) = E_{{{\text{sat}}}} \frac{{\tau_{{\text{s}}} }}{{\tau_{{{\text{PV}}}} + \tau_{{\text{s}}} }}\left[ {1 - \exp \left( { - \frac{t}{{\tau_{0} }}} \right)} \right]$$
(4)

where t is the time under illumination, τs is the screening relaxation time of the surrounding medium, τPV is the photovoltaic response time (inversely proportional to the light intensity) defined by Eq. (3), τ0 = τsτPV/(τs + τPV) is the effective time constant and Esat is the saturation photovoltaic field when the surrounding medium is not leaky (i.e., τs → ∞). The relaxation time τs is related to the electrical conductivity of the surrounding medium through the Maxwell–Wagner time constant: τs = εsε0/σs, where σs is the electrical conductivity and εsε0 the dielectric permittivity. In our case with aqueous solutions, σs will be strongly dependent on the concentration of ions in the solution, which are responsible for the screening by forming a Debye double layer at the water-LN:Fe interface. Note that both the saturation photovoltaic field and the effective time constant τ0 depend on the light intensity through τPV. Namely, the higher the intensity, the higher the saturation photovoltaic field and the faster the time evolution. On the other hand, when light is switched off, the evanescent field rapidly fades away.

3 Experimental details

The experiments have been carried out using monodomain 1-mm-thick z-cut congruent LiNbO3 crystals, both sides optical quality polished (typical roughness ~ 1 nm) and highly doped with iron (0.25% mol) in order to have a strong photovoltaic effect. They are purchased from Photon Lines Optica Ltd. Both the + c and –c crystal faces have been used as active surfaces of the photovoltaic substrate because the stored charge generated under illumination has opposite sign as illustrated in Fig. 1a.

A single beam from a frequency-doubled Nd:YAG cw laser (λ = 532 nm) has been used to illuminate the substrates. The gaussian light beam was focused on the substrate with a long working distance microscope objective (magnification 20 × , numerical aperture NA = 0.42). The light spot has a diameter 4σ = 150 μm (where σ is the standard deviation of the gaussian beam profile). A schematic of the illumination setup is shown in Fig. 1b.

Two different methods are currently used in the operation of PVOT for particle patterning from non-polar suspensions. The simultaneous method is a real-time procedure in which the substrate is in contact with the solution during illumination. The substrate is placed on the bottom of a cuvette filled with aqueous solution (see Fig. 1c) where it is illuminated. In turn, the sequential method is a two-step procedure. First, light induces photovoltaic fields in the ferroelectric substrate. Then, once illumination is switched off, the solution containing the material to be patterned (micro-/nanoparticles, salt) is put in contact with the substrate. When the materials are in aqueous solution, typically the simultaneous method operating with high light intensities are previously used to avoid full photovoltaic field screening [21, 23].

In the present application and in both, the sequential and the simultaneous method, right after being in contact with the salt solution, the illuminated crystal face is placed on an absorbent paper (Fig. 1d) for about one second to facilitate substrate drying. The result of the complete process is visualized and recorded with a Nikon (Eclipse 80i) brightfield optical microscope.

4 Results and discussion

4.1 Real time operation of PVOT

First results were obtained by real time operation, i.e., illuminating the photovoltaic crystal inside the saline water before putting it in contact with the absorbent paper. This is the simultaneous method, described in the previous section. High light intensities were employed (I ~1 kW/cm2) in order to achieve a short photovoltaic response time that avoids full screening of evanescent photovoltaic fields (see Sect. 2). The saline solution filling the cuvette is NaCl in deionized water. The substrate is illuminated on the –c face with two light spots during 2 and 3 min and the salt concentration was 10 g/L. Right after illumination, the crystal is removed from the cuvette and dried on the absorbent paper. Inspection of the results by optical microscopy shows a successful light control of the spatial distribution of salt. As shown in Fig. 2a, two spots of crystal salt can be observed exactly on the illuminated regions, whereas all the rest of the image is free of salt. The structure of the salt crystals in the right spot is magnified in the inset. Similar results (Fig. 2b) are obtained if the + c face is exposed to the light. In this case, the exposure times are 1 and 4 min for the left and the right spot, respectively.

Fig. 2
figure 2

Microphotographs showing the salt precipitation on the position of light spots (diameter d ~150 µm). For a and b, the light intensity is 1 kW/cm2 and the exposure times are a 2 and 3 min for the left and right spots, respectively, in the –c face and b 1 and 4 min for the + c face. In c, four light spots with I = 1.5 kW/cm2 and exposure time of 1 min are recorded following the order indicated by the numbers. The dotted white lines indicate the illuminated regions (diameter 4σ = 150 µm)

Full size image

A 2D arrangement has been also attempted in one experiment in which we record 4 light spots successively on the four vertices of a square, each one during 1 min, with 1.5 kW/cm2, illuminating the –c crystal face. A microphotograph of the result is presented in Fig. 2c where the numbers appearing in the figure indicate the order of light exposure. Again, salt crystallization is clearly apparent only inside the four light spots.

4.2 NaCl crystallization using the sequential method

The previous results with the simultaneous method demonstrate the capabilities of PVOT to control the position of the crystallization spots and to organize them in 1D and 2D patterns. However, in PVOT operation with the other cited method (sequential method) much lower intensities are usually applied (W/cm2 or even less) with the advantages of energy saving and avoiding any possible crystal heating or damage. Thus, a sequential method for salt crystallization control and patterning has been also designed and attempted, although this latter method has failed in previous work when operating in water solution [8, 16, 23]. In the sequential protocol, the substrate is first illuminated in air. Then, the crystal is introduced for about 5 s in a cuvette filled with the saline solution, before contacting it with the absorbent paper.

In Fig. 3, we show the results when focusing the laser gaussian beam on the substrate surface. The averaged light intensity is 55 W/cm2 (~20 times lower than in the simultaneous method), and the light spot diameter is about 150 µm. Results of Fig. 3a and b correspond to experiments performed in the –c face with one light spot and in the + c face with two light spots, respectively. Unexpectedly, successful salt crystallization on the illuminated regions has been achieved showing that this second strategy is also efficient for light-assisted salt patterning. The structure of salt spots is similar to that of the simultaneous method although the light intensity used is significantly lower. In the salt spot of Fig. 3a, two different structures can be also distinguished: a region with dot-like salt deposition and another one, more complex, with branched spikes. In fact, these two kinds of salt deposits can be distinguished in the previous figures: Fig. 2a presents complex structures, whereas Fig. 2b and c shows spots essentially formed by salt points.

Fig. 3
figure 3

Microphotographs of the salt crystallization spots deposited on a the –c face, and b the + c face. The light intensity is I = 55 W/cm2, and the exposure times are indicated in the photograph. The solution salt concentration is 10 g/L. The dotted white lines indicate the illuminated regions (diameter 4σ = 150 µm)

Full size image

4.3 Role of the light exposure time

The influence of the exposure time t on the final crystallized region was investigated to further characterize the process using the sequential method. To this end, the + c face of the LN:Fe crystal has been illuminated with four consecutive light spots with increasing exposure times. The light beam parameters are the same as in Fig. 3. The global experiment has been repeated two times, and the corresponding salt spot micro-photographs are shown in Fig. 4a and b, respectively. The salt crystalizes on the four illuminated regions, but the amount of deposited salt increases, as expected, with the exposure time. Moreover, as shown in Fig. 4c where d versus texp has been plotted, the diameter d of the salty region also increases with t, saturating for d ~200 µm. It is worthwhile noting that this kind of dependence with light exposure has been already observed in other kind of applications of PVOT [6, 31]. In these works, it was well correlated with the predicted recording kinetics of the photovoltaic surface charge generated under gaussian light illumination [28]. Therefore, this dependence arises from the recording kinetics of bulk photovoltaic effect.

Fig. 4
figure 4

NaCl salt spots on the + c LiNbO3 face for different exposure times. The sets of images a y b show the results of two series of experiments performed under the same conditions. The dotted white line indicates the illuminated region (diameter 4σ = 150 µm). c Dependence of the salt spot diameter on the light exposure time. The light intensity is I = 55 W/cm2 and solution salt concentration is 10 g/L. The scale bar represents 200 μm

Full size image

4.4 Role of salt concentration

The influence of salt concentration in the solution has also been investigated using the sequential procedure for the concentration range 1–100 g/L. The light intensity was I = 55 W/cm2 and the exposure time 2 min. The microphotographs of the obtained salt spots for decreasing salt concentrations are shown in Fig. 5a–e. One can see that the size of the crystals forming the crystallized region increases with the concentration. On the other hand, excepting for the highest concentration (Fig. 5a), all saline spots present a halo constituted by very small salt points (see Fig. SI-1 in the supplementary information that is given in Online Resource 1).

Fig. 5
figure 5

a–e Microphotographs of the salt spots for five decreasing salt concentrations indicated in the figures. The light intensity is I = 55 W/cm2 and the exposure time 2 min. The scale bars represent 50 µm

Full size image

At first glance, the microphotograph corresponding to 1 g/L (Fig. 5e) shows only the ring halo and one can wonder whether inside it, there are nanometric crystals smaller than the resolution of the optical microscope. To investigate this point and to better visualize the halo, we have resorted to a scanning electron microscope that provides a higher spatial resolution. We have found that indeed there is salt covering the full circle. Three representative images at increasing magnifications are shown in Fig. 6. Figure 6a shows the whole spot, although at this magnification only the outer boundary (also appreciable with the optical microscope) can be distinguished. Figure 6b illustrates the nanocrystals inside the circle forming again branchy shapes probably precursor of the complicated branchy structures observed at higher concentrations (see Figs. 2, 3 and 5). In Fig. 6c, with the highest magnification, it can be clearly observed that the structures are formed by single roughly cubic nanocrystals that are separated from each other. Their sides are in the range 10–100 nm.

Fig. 6
figure 6

SEM images of the salt spot for a concentration of 1 g/L, for three increasing magnifications as can be inferred from the scale bars indicated in the image. Other parameters are the same of those of Fig. 5 (I = 55 W/cm2 and exposure time of 2 min). The images correspond to a the whole spot, b and c regions inside the spot

Full size image

4.5 Discussion on the physical mechanisms

Although it is the first report on salt crystallization and patterning by PVOT, it is worthwhile providing a discussion on the physical mechanisms most likely involved in this phenomenon.

The spots of salt precipitation are generated just on the illuminated regions of the substrate. Moreover, the diameter of the spot recorded with the sequential method follows the typical time dependence of the generation of the photovoltaic field under gaussian illumination [28, 31] (see Fig. 4). Thus, the photovoltaic field generated in the substrate, or in other words the action of PVOT, is the main mechanism responsible for the spatial patterning of salt precipitation. However, in previous work, PVOT have had no effects, or at most a residual effect, when acting on aqueous dispersions of particles or biomaterials due to water screening of the photovoltaic fields [13, 16, 21, 23]. Thus, one can wonder why this screening effect does not prevent salt precipitation. Our experiments suggest that the key point is the electrowetting effects on the water/LN:Fe interface, originated by the surface charge induced by the bulk photovoltaic effect. This wetting increase, induced by electric fields, is a well-known phenomenon [32,33,34,35,36]. In fact, it was observed in our experiments that, during drying by the absorbent paper, the salty water was removed from the non-illuminated substrate regions but remains partially on the illuminated ones, right where the salt precipitates. Moreover, the electrowetting mechanism is consistent with two experimental observations: (i) the possibility of operating with the sequential method despite water screening and (ii) the successful results in both, + c and –c faces, with no significant differences.

To further confirm light-induced electrowetting as the main mechanism, we have conducted a simple experiment to carefully visualize the whole procedure for salt crystallization patterning. The LiNbO3:Fe crystal has been illuminated with a single light spot, and the whole process for salt crystallization has been recorded on a video included in the supplementary information (Online Resource 2). Typical experimental parameters are used: I = 55 W/cm2, exposure time of 3 min and salt concentration of 5 g/l. Figure 7 shows a sequence of three video frames corresponding to the key steps of the process. Under each frame, there is a schematic diagram to facilitate understanding of the corresponding experimental conditions. Once the crystal is illuminated in air, it is covered with a NaCl dissolution film as shown in Fig. 7a where the dashed line circle indicates the illuminated region. No salt crystallization is observed at this step. Next, drying with the adsorbent paper is carried out. As a consequence, the saline water is removed from the crystal surface except right in the illuminated circle, as distinguished in Fig. 7b. In other words, light-induced electrowetting, indeed, acts retaining a small droplet. Finally, as water evaporates, the dissolved salt contained in it crystallizes (see Fig. 7c). Therefore, light-induced electrowetting is confirmed as the main mechanism of salt crystallization patterning. However, some contribution of direct salt ion trapping by photovoltaic electrical forces cannot be fully discarded. The possible influence of this latter effect will be addressed in future work.

Fig. 7
figure 7

a Microphotographs extracted from the video recorded after illumination (included in the supplementary information). b Schematic diagram illustrating the experimental conditions of each image. More details are self-explained in the figure. The contrast of the images has been enhanced for better visualization of the process

Full size image

5 Moving light beam: drawing the crystal patterns

So far, we have reported experiments in which we record one or a few static light spots. However, to enlarge the patterning capabilities of the technique we have checked the feasibility of applying the crystallization protocol with scanning light beams that follow an arbitrary trajectory. Quite successful results for a few simple light paths (lines and crosses) are shown in Fig. 8. Both the simultaneous (Fig. 8a and b) and the sequential methods (Fig. 8c–e) were used, and the results are shown in the first and second rows, respectively. As expected, the crystallization patterns coincide with the light paths. Figure 8c shows a particularly well-defined straight central pattern surrounded by the halo already found in Fig. 5.

Fig. 8
figure 8

Microphotographs of the crystallized patterns obtained under a moving light beam for a salt concentration in the solution of 10 g/L. The trajectory of the laser beam is indicated by a green dashed line. Images a and b correspond to experiments using the simultaneous method with an intensity I = 1 kW/cm2 and an exposure time of 30 and 60 s (30 s for each line of the cross), respectively. Images c, d and e correspond to experiments using the sequential method with light intensity I = 55 W/cm2, and the total exposure times are 3, 6 and 9 min, respectively. The scale bars represent 150 µm

Full size image

6 Other salts

A final point to be investigated is whether the spatial control of crystallization by light is specific of NaCl or a more general result. To this end, we have checked the sequential method with two rather different salt solutions: sodium bicarbonate in water and D-PBS (Dulbecco’s phosphate-buffered saline). D-PBS contains several salts and is widely used in biology, for instance, for cell culture. We used the usual green laser beam with light intensity of 55 W/cm2, and the exposure time was 3 min. The results are shown in Fig. 9. It is clearly seen well-defined salt spots on the previously illuminated regions for the two salt solutions, suggesting a general applicability of this optoelectronic method for light-assisted patterning of salt crystallization.

Fig. 9
figure 9

Microphotographs of the obtained salt spots from a sodium bicarbonate and b D-PBS solutions. I = 55 W/cm.2 and the exposure time 3 min. The concentration of sodium bicarbonate solution is 10 g/L. D-PBS is a solution of several salts: CaCl2 (0.05 g/L), KCl (0.1 g/L), MgCl2 + 6H2O (0.05 g/L), NaCl (4 g/L) and Na2HPO4 (1.08 g/L)

Full size image

7 Conclusions and outlook

In conclusion, spatial control and patterning of salt crystallization by a light beam on photovoltaic platforms have been demonstrated. This is a novel application of photovoltaic optoelectronic tweezers. The salt crystallizes only on the illuminated region of the photovoltaic platform. Several salts have been tested as well as static and scanning light beams have been used. In the latter case, the method allows drawing the crystallization pattern.

Two different protocols for the operation of PVOT, the simultaneous and sequential methods, have been successfully applied. The second one, which uses lower light intensity, is particularly interesting, but it was not efficient in previous applications of PVOT with water solutions. The results indicate that the success in the present application is due to the peculiar mechanism controlling light-assisted salt patterning, consisting in the spatial modulation of the salty water wettability on the LiNbO3:Fe substrate by photovoltaic fields.

The new functionality of photovoltaic platforms that allows a spatial modulation of the crystallization of salts should find relevant applications in different fields such as chemistry or biomedical microfluidics. For instance, one can envisage its relevance for LN-based biological laboratory on a chip [18] or for sensor applications to mark the precipitation regions just in correspondence of the sensitive areas [37]. On the other hand, the results suggest exploring the feasibility of additional applications of PVOT in water solution based on wettability patterning.