1. Introduction

During the last decade, Optofluidics, have gained significant attention from the photonic community, providing an alternative, yet efficient and functional way of controlling the propagation of light while utilizing fluids and fluidic actuations [1]. There have been several reports on the development of planar, fiber and free space optical components, where pneumatic, electrostatic, or magnetofluidic control of liquid mediums result to specific refractive, absorptive, and light scattering effects utilized for biosensing, lasing, actuating, and switching devices [1–7]. In most of the optofluidic devices presented, the actuation of a fluid is used for controlling the guidance of the light; accordingly the physical, spatial or optical properties of the fluid are tuned using external stimulations. However, there is little work until now on the inverse actuation, namely, the full control of fluidic actuation utilising light [8]. Therefore, a great challenge of Optofluidics is the control of the fluidic properties of photonic circuits and platforms by utilizing solely light sources, thus, turning fluidic operation and subsequent light manipulation to an all-optical process [9].

All-optical optofluidic actuation can introduce several advantages to the future development of optofluidic and microfluidic lab-on-a-chip devices: light driven fluidic actuations can reduce the need for moving or pneumatic parts and increase device simplicity and robustness. In the same time, rheological activity and fluidic immobilisation due to spatially confined changes of the surface wettability, may be carried out on surface areas with dimensions close to the optical resolution limit of the light source used. The last can improve integration features of the light driven optofluidic devices developed; it can may also used for cell immobilisation in biosensing chips, complementary to standard laser tweezing [10]. Further, future optically actuated optofluidic lab-on-a-chip and lab-in-a-fiber configurations may use non-laser radiation, i.e. solar light for tuning their performance or sustaining their operation, simultaneously covering emerging applications from biosensing [11] to energy conversion [12,13].

In this study, we report on the demonstration of an all-light controlled optofluidic operation, exemplified on a microstructured optical fiber (MOF) platform. To the best of our knowledge, this constitutes the first demonstration of controlling the flow of liquids within an optofluidic circuit by solely utilizing light, by means of laser beams. In order to achieve all-optical optofluidic actuation we demonstrate and exploit the reversible photoinduced surface wettability transduction of zinc oxide (ZnO) surfaces, grown either on flat glass substrates or inside the capillaries of the MOF used. The implementation of a light-driven ZnO wettability actuator in a MOF photonic device is a step towards the implementation of the Lab-in-a-Fiber approach [6,14–17]. Upon the fabrication method followed and conditions applied ZnO surfaces exhibit an inherent hydrophobic response. The wettability properties of ZnO have been investigated before, however using a combination of ultraviolet (UV) laser irradiation and subsequent thermal treatment for switching between the hydrophilic and hydrophobic states, respectively [18]. Here, we too employ UV radiation for turning the ZnO surfaces hydrophilic, followed by a continuous wave (cw) green laser irradiation, for back switching to hydrophobic state. For demonstrating the impact of this all-optical optofluidic actuation and investigating relevant operational dynamics, ZnO layers were grown into the capillaries of a MOF Fabry-Perot (FP), endface interferometer, utilizing a sol-gel method [19,20].

As it will be shown in the following sections of this manuscript, the above switching mechanism between the hydrophilic and hydrophobic character of the ZnO overlaid surfaces, in combination with the micro-meter diameter of the MOF capillaries and the intrinsically amplified surface tension effects occurring within those, prompt the straightforward fluid transfer inside and outside the MOF capillaries, thus, the demonstration of an efficient optufluidic actuation. The experimental results presented reveal a fully reversible optofluidic switching behaviour of the configured in-fiber system upon exposure to UV (248 nm excimer laser or 254 nm pencil lamp) and green (532 nm cw) laser beams, while using modest irradiation doses. Further investigations include the temporal characterization of this all-optical optofluidic actuation, occurring within the confined volumes of the MOF capillaries, providing critical information on the nature and speed of the fluidic actuation. Finally, photoluminescence spectroscopy is used for providing a first insight into the physical background underlying the optically driven wettability tuning of the ZnO nanolayers.

2. Experimental

For the development of the optical fiber FP interferometer a commercially available all-silica MOF (HNA-5, drawn by NKT Photonics Ltd.) is spliced to a SMF-28 fiber, as shown schematically in Fig. 1(a) [21]. Typical lengths of the designed HNA-5 micro-cavities are of the 700 to 1000 μm range. Upon splicing, the MOF capillaries were infiltrated with 0.152 M Zn-acetate (Zn(CH₃COO)₂•2H₂O)/methanol solution following appropriate annealing treatments for the formation of uniform ZnO nanolayers inside all six capillaries of the HNA-5 fiber. The growth of ZnO nanostructures inside the capillaries of MOFs by means of this sol-gel infiltration protocol has been extensively studied elsewhere, in order to identify optimum infiltration and annealing conditions for the formation of uniform ZnO nanolayers of outstanding crystalline quality [20]. For the present study, upon infiltration the MOF fiber cavity was heated at 120 °C for 20 min to allow slow evaporation of methanol, followed by final annealing at 300 °C for 2 h. The axial homogeneity of the so formed ZnO nanolayers was verified by field emission scanning electron microscopy (JEOL, JSM-7000F), i.e. performed at various cross-sections along the length of the HNA-5 fiber cavity.

 

Fig. 1 (a) Schematic representation of the optofluidic actuation cycle and (b) principle of the all-optical actuation based on the reversible transition between hydrophobic and hydrophilic states of the fabricated ZnO structures upon laser irradiation exposures.

Download Full Size | PDF

For the implementation of the actuation experiment, a droplet of ultrapure water (pH 7) was attached to the end face of the MOF so that the tip of the fiber remained always immersed within water as shown schematically in Fig. 1(a). The principle of the actuation switching relies on the control of the water flow within the FP cavity by the exploitation of the tunable wettability of the ZnO nanolayer by means of UV and green laser irradiations. In particular, a UV excimer laser emitting at 248 nm (10 ns pulse duration), or a UV cw pencil lamp emitting at 254 nm are employed to induce the ZnO hydrophilicity, whereas a cw green laser at 532 nm is used to revert ZnO back in its hydrophobic state [Fig. 1(b)]. In both cases the fiber was positioned perpendicular to the laser beams. Based on the above, the optical actuation principle can be described in three steps as summarized schematically in Fig. 1(a). Initially, in step 1, the hydrophobic character of the synthesized ZnO nanolayers within the HNA-5 capillaries prevents water infiltration of the MOF cavity, which would have been the expected outcome for a pristine MOF cavity due to capillarity effect. Once the HNA-5 ZnO nanolayer cavity is exposed to UV laser beam the ZnO structure becomes hydrophilic, and subsequently, capillarity effect takes place and the MOF is rapidly filled with water (step 2). Following step 2, MOF cavity is exposed to green laser irradiation, which reverses ZnO nanolayer back to its hydrophobic state. Once the ZnO structure turns hydrophobic, water is drained out of the HNA-5 capillaries and the FP cavity becomes empty of water (step 3), i.e. as it was prior to UV laser exposure in step 1.

Throughout the optofluidic cycle, the flow of water is examined by optical microscopy, whereas FP spectra were interrogated in reflection mode by employing a 50/50 coupler, a superluminescence source (Q Photonics), and an Optical Spectrum Analyzer (OSA, ANDO AQ6317B). For the in situ temporal FP spectra monitoring of this all-optical optofluidic actuation an IBSEN interrogator (I-MON 256 OEM) was employed, capable of spectra recording with frequencies up to 3 kHz within the 1525-1570 nm range, and a spectral resolution of 0.25 nm.

In addition to the optical fiber experiments, the wettability behavior of ZnO nanolayers deposited on flat silica glass substrates, using the same wet chemistry conditions, was characterized individually by means of contact angle (CA) measurements after exposures to the described laser sources at various irradiation conditions. Room temperature photoluminescence (PL) spectra of ZnO nanolayers grown within MOF capillaries were measured before and after laser exposures by using a cw He-Cd laser at 325 nm with a power of 25 mW. The PL spectra were recorder by employing a spectrometer with grating 600 grooves/mm blazed at 300 nm, and a sensitive liquid nitrogen cooled CCD camera. Finally, optical density measurements were performed on ZnO samples grown on silica substrates, using a Perkin-Elmer spectrophotometer.

3. Results and discussion

3.1 Fabrication of the ZnO-overlaid MOF Fabry-Perot interferometer

Figure 2(a) displays an optical microscopy photo of the configuration of the fiber-endface SMF-28/HNA-5 FP interferometer. Due to the difference between the core diameter and the corresponding modal profile of the two fibers, i.e. d_SMF-28 = 8.2 μm and d_HNA-5 = 5.1 μm, the splicing interface discontinuity acts as the one side mirror of the Fabry-Perot cavity, wherein light propagating into the HNA-5 fiber is back-reflected at the larger SMF-28 interface. The open and perpendicularly cleaved endface of the HNA-5 fiber, forms the other reflection interface of the Fabry-Perot resonator. A typical FP reflection spectrum over 10 nm of a SMF-28/HNA-5 interferometer with a MOF cavity of 780 ± 2 μm is shown in Fig. 2(b). For this specific cavity length we obtain a free spectral range (FSR) of 0.95 nm, which is in relatively good agreement with theoretical predictions [22]; collapsing of the MOF along the splicing point and errors into the estimation of the cavity length can largely affect the FSR. Figure 2(c) presents SEM scans of the cleaved end face of the HNA-5 fiber of the same FP interferometer after ZnO nanolayer addition, along with details of the ZnO overlaid capillaries. Inspection of SEM photos reveals the formation of uniform ZnO nanolayers with a typical thickness of ~25 nm on the inner surface of the capillary silica walls.

 

Fig. 2 (a) Configuration of the SMF-28/(ZnO nanolayer HNA-5) Fabry-Perot interferometer. (b) Fabry-Perot reflection spectra over 10 nm of pristine and ZnO overlaid HNA-5 cavity with length of 780 μm. (c) SEM scan of the cleaved end face of the HNA-5 MOF, along with details of the ZnO nanolayer formed on the surface of the capillaries.

Download Full Size | PDF

Upon ZnO nanolayer formation in the HNA-5 the effective index of the FP cavity is altered resulting to a red shift of 0.35 nm of the FP fringes when compared to that of the pristine cavity [Fig. 2(b)]. From FSR variations of the FP spectra we calculated an increase in the effective index of the ZnO overlaid HNA-5 fiber of the order of ~7x10⁻⁴.

3.2 Wettability tuning of ZnO nanolayers

Prior to investigating the ZnO overlaid FP MOF cavities, a study on the wettability of planar ZnO nanolayers grown on typical glass substrates by means of contact angle (CA) measurements versus UV and green laser exposures, was carried out. The ZnO nanolayers were synthesized by following an identical sol-gel route protocol as for the corresponding in-fiber grown structures [20]. Figure 3(a) presents typical shots of the shape of a water droplet on ZnO nanolayer substrate before and after 4 minutes of 248 nm excimer laser irradiation that delivers a total energy dose of 102 J/cm², while Fig. 3(b) shows how the CA varies throughout the exposure process using both pulsed and cw ultraviolet radiation sources. In general, prior to UV exposure a CA of ~85° is obtained for the ZnO film; this is in agreement with previously reported values for equivalent ZnO structures [18]. It becomes apparent that CA values decrease upon exposure to UV radiation, demonstrating that the ZnO nanolayer moves progressively towards the super-hydrophilic regime. Such behavior is directrly attributed to the introduction of surface oxygen defects on the ZnO surface [18]. Notably, after ~3 minutes of 248 nm laser irradiation, i.e. corresponding to an energy dose of 76.5 J/cm², a sudden drop of the CA value is noticed indicative of the super-hydrophilic regime threshold, i.e. CA < 25°.

 

Fig. 3 (a) Photographs of the shape of a 5 μL water droplet on ZnO nanolayer before and after irradiation with 248 nm excimer laser. (b) Contact angle variation against total energy dose of exposures using 248 nm excimer laser (black filled points) and 254 nm pencil lamp (semi-filled black points) radiation. (c) Photographs of the shape of a 5 μL water droplet on ZnO nanolayer pre-exposed to 248 nm excimer laser radiation before and after irradiation with green laser. (d) Contact angle variation against total energy dose of exposure of the ZnO UV pre-irradiated samples after reversing irradiation using a cw green laser at 532 nm.

Download Full Size | PDF

A point of particular interest refers to the efficiency of hydrophilicity actuation using pulsed 248 nm and cw 254 nm ultraviolet radiation. The data of Fig. 3(b) reveal that 254 nm cw pencil lamp radiation reverses the CA of ZnO thin films to super-hydrophilic at low doses of ~10 J/cm²; contrary, the same wettability transition takes place for ~6-7 times greater radiation doses when 248 nm excimer laser pulsed radiation is applied. ZnO is almost opaque to radiation for wavelengths lower than ~370 nm; a 25 nm thin ZnO film deposited with the method presented herein on fused silica substrates, exhibited high absorption at 300 nm, namely ~0.53 μm⁻¹. Finite difference spatio-temporal simulations utilising a 10 ns pulse duration such as the one of the 248 nm excimer laser, the 300 nm absorption figure of the ZnO film and the irradiation conditions of Fig. 3(a), yielded a local increase of temperature within the oxide layer greater than 300 °C [23]. Corresponding values for the actual, higher absorption exhibited at 248 nm will lead to even greater temperature increase leading to photo-thermal annealing effects. Such effects are known to eliminate the oxygen surface defects responsible for hydroxyl group absorption [18], and consequently counteract the desired UV induced hydrophilicity of the ZnO films when high intensity pulsed laser sources are used; and thus, switching to hydrophilic state requires greater energy density doses as appearing in Fig. 3(b). Another hydrophilicity counteracting element when the UV pulsed laser is employed, lies on the possibility that some electron holes generated by the irradiation, and being responsible for the introduction of defects that induce hydrophilicity, recombine back to their initial state between each consecutive pulse, i.e. before initiating defect forming reaction in the lattice. It is worth to note that equivalent simulations for the cw UV lamp exposure conditions revealed a temperature increase on the ZnO surface of ~35 °C, i.e. insufficient to cause any thermal annealing effects.

Following UV laser exposure, similar experiments were performed on the same ZnO nanolayer substrates irradiated with the 532 nm green laser. Figure 3(c) presents shots of the water droplet shape before and after 8 minutes of green laser exposure, i.e. corresponding to a total energy dose of 68.8 J/cm², while Fig. 3(d) illustrates the progressive recovery of the ZnO hydrophobic character. In particular, after 8 minutes of irradiation CA values of ~75° are observed, i.e. indicative of ~90% recovery of the initial hydrophobic state. The reversing curve rate and threshold from hydrophilic to hydrophobic behaviour appeared independent from the UV irradiation conditions, namely, the use of pulsed or cw radiation sources.

3.3 Fabry-Perot spectra and temporal dynamics of the in-fiber optical actuation

After the wettability studies on planar ZnO samples, investigations were focused on the ZnO overlaid FP HNA-5 microcavity. Initially, the ZnO overlaid FP HNA-5 microcavity was immersed inside a drop of de-ionised water (see step 1 in Fig. 1). Due to the hydrophobic character of the ZnO overlayer, no water was penetrating inside the capillaries of the MOF section; i.e. which would have been the expected outcome due to capillarity forces. Τhe visibility of the FP fringes as those measured in reflection mode was reduced from 8.3 dB [Fig. 2(b)] to ~1.34 dB [Figs. 4(a) and 4(b)] due to the presence of water on the end face of the cavity that alters its reflectivity.

 

Fig. 4 Fabry-Perot reflection spectra of SMF-28/(ZnO nanolayer HNA-5) interferometer throughout the optofluidic cycle, with the employment of the 248 nm laser (a) and the 254 nm pencil lamp (b) as UV sources. The MOF cavity length is 780 μm. (c) Fabry-Perot reflection spectra of the same cavity corresponding to a three full cycles operation (see text).

Download Full Size | PDF

Following this, the FP HNA-microcavity was exposed using the laser UV source, described in previous sections, for rendering the ZnO in-fiber overlayers hydrophilic; while spectra before and after exposure to UV radiation are presented in Fig. 4. The exposure to UV radiation allowed the complete infiltration of the HNA-5 MOF cavity (step 2 in Fig. 1), resulting in a shift of the FP spectral fringes by 0.36 nm. The radiation dose utilised for inducing this hydrophilic infiltration was 76.5 J/cm² approximately. This spectral shift is translated to a change of the effective refractive index of the MOF of n_eff~3.47x10⁻⁴ [22]. In the final step of the cycle (step 3 in Fig. 1), the hydrophobic nature of the ZnO nanolayers is regained by applying green laser irradiation, i.e. 8 minutes with total energy dose of 68.8 J/cm², forcing water out of the MOF capillaries; while the FP spectral fringes returned to their original position. After exposure to the 532 nm laser radiation the visibility of the fringes of the MOF FP is slightly increased. We expect that ZnO overlaid onto the MOF endface will repel water, thus, endface refractive index contrast will increase. The spectral results of the hydrophobic behaviour switched-on after the exposure to the 532 nm radiation were also confirmed by optical microscopy inspection. Notably, the optofluidic switching cycle (steps 1-3 in Fig. 1) can be repeated several times with the same outcome and with only negligible hysteresis effects in both fringe contrast and shift of the FP reflection spectra. Statistically, a three full cycles operation of the same FP cavity resulted in forward and reversing spectral shifts with an absolute hysteresis of ~15 pm, i.e. revealing good repetability performance. The actual FP reflection spectra of the three full cycles operation are shown in Fig. 4(c), while the UV 248 nm laser was employed to induce hydrophilicity of the ZnO layers (step 1 of each cycle).

To exclude material damage or extensive photosensitivity effects induced by the exposure of the ZnO overlaid MOF to the UV and green laser radiation, a second set of experiments was carried out with the ZnO overlaid FP MOF cavity being empty of any water, i.e. absence of water drop. The exposure conditions were kept identical to those used for switching between different wettability stages onto the ZnO. These experiments showed no significant shift (less than 5 pm) of the FP fringes, thus, no damage was induced into the ZnO film; while these findings were also cross-checked by inspecting scanning electron microscopy pictures of the exposed ZnO overlaid MOF where no surface damage or corrosion was traced.

Furthermore, prompted from the need to design smaller and more compact optofluidic switching in-fiber devices, the optical actuation experiment was repeated by employing a cw UV pencil lamp emitting at 254 nm, instead of the 248 nm excimer laser for triggering the hydrophilic nature of the synthesized ZnO structures. In order to reach the ZnO-nanolayer super-hydrophilic regime that allows the cavity filling with water, i.e. contact angles smaller than 25°, MOF cavity is exposed to the UV pencil lamp light for 5 minutes. Those exposure conditions resulted to a total energy dose of ~10 J/cm². Figure 4(b) shows FP reflection spectra for the three steps of an optofluidic cycle where the UV pencil lamp was used. Following exposure, the cavity fills with water and the ~0.36 nm red shift of the FP fringes is observed. Reverting ZnO to the hydrophobic condition and the FP spectra recovery to its original position, is identical to that described earlier when the UV pulsed laser was employed for inducing hydrophilicity. Confirming the CA results of Fig. 3(b), the two different UV radiation sources produce the same spectral effect into the hydrophilic infiltration of the FP MOF cavity, while identical green laser irradiation conditions were required for switching back to the hydrophobic state (step 3 of the cycle).

We consider now the temporal dynamics of the optofluidic actuation, as they emerge from ‘in situ’ measurements of the FP spectra during water filling and draining periods. Experiments were performed on a SMF-28/(ZnO nanolayer HNA-5) FP interferometer with a cavity length of 700 μm. For step 1 of the cycle (water filling), the MOF cavity was treated with UV laser irradiation by using the aforementioned exposure conditions, and thus, making the MOF ZnO nanolayers super-hydrophilic. Following this, the end face of the MOF was attached on a water drop, while FP spectra were recorded with an acquisition rate of 2.5 kHz and the filling process was simultaneously monitored by optical microscopy. Figure 5(a) shows the FP spectrum prior to water addition and the final spectrum after the cavity is filled with water. The expected red shift is obtained, accompanied by a significant reduction of the fringes visibility.

 

Fig. 5 (a) Fabry-Perot reflection spectra of SMF-28/(ZnO nanolayer HNA-5) interferometer that has been treated with UV laser exposure, i.e. 3 minutes with total energy dose of 76.5 J/cm², before and after immersion of the MOF end face in water. Temporal variation of fringe intensity during MOF cavity water filling –see Visualization 1- (b) and draining (c) for a randomly selected fringe (see text). The inset in (b) corresponds to the signal measured for a hydrophobic cavity. (d) Snapshot of the MOF cavity in the middle of draining process following incomplete green laser irradiation (see text). The MOF cavity length is 700 μm.

Download Full Size | PDF

More interestingly, Fig. 5(b) presents the periodic change in the intensity of a randomly selected fringe versus time throughout the filling process. A polynomial data fitting analysis reveals a temporal length of ~10 ms, corresponding to the time required for the empty MOF cavity to be totally filled with water, and thus, the reflected FP spectrum to be settled to its new red-shifted position and reduced intensity. Similar data analysis was performed in several FP spectra fringes, and of various MOF cavities, confirming an average filling speed of ~70 μm/msec. Notably, when the same measurement was performed on a MOF cavity with hydrophobic ZnO layers, i.e. without UV treatment, the expected decrease of the fringe intensity is noticed due to water drop scattering effects at the MOF end face, however without the presence of the filling time period as water infiltration in the MOF cavity is prevented [inset of Fig. 5(b)]. A high definition video of the cavity filling was captured at 1200 fps and it is submitted as supplementary material (Vid. 1). The video shows the progressive movement of water within the capillaries of the HNA-5 fiber, while digital analysis allows the calculation of cavity water filling speed, which is found to be ~65 μm/msec, i.e. in good agreement with the value obtained from the temporal analysis of the FP spectra.

A similar temporal investigation was performed for the water draining process (step 3 of the cycle), namely, when the FP MOF cavity was hydrophilic and fully filled with water. To speed up the draining process so that it can be completed within the resolution/time limits of the employed interrogator for FP spectra acquisition, higher intensities of green laser irradiation were used, namely 2.29 W/cm², leading to draining occurrence within ~30 sec. Figure 5(c) shows real-time changes in the intensity of a selected fringe of the recorded FP spectra. From the beginning of green laser irradiation at 0 seconds to 19.7 seconds only negligible variations occur. At 19.7 seconds a significant fluctuation of the fringe intensity starts, denoting the initiation of the capillary draining process. After a characteristic temporal slot of 3.1 sec approximately, the fringe intensity settles again to a relatively constant value indicating the end of the optofluidic cavity draining process. Draining times were measured for several cavities of similar lengths, resulting to draining speeds of ~0.23 μm/msec, substantially slower than the infiltration speed which is prompted by the ZnO hydrophilicity.

While the presence of the intense green laser beam does not allow continuous video capturing of the draining process, several snapshots of the MOF cavity were taken during the draining period and a typical one is shown in Fig. 5(d). Intermediate snapshots of the draining process like the one presented in Fig. 5(d), revealed that the draining of the MOF cavity is not homogeneous: there are capillary sections where water is drained, while other sections are still filled. Such a type of inhomogeneous infiltration can lead to the generation of secondary Fabry-Perot oscillations within the infiltrated MOF, leading to the modulated temporal behaviour observed in Fig. 5(c). However, once sufficient energy dose has been transferred onto the sample, ZnO fully reverts back to hydrophobic state, eventually forcing the water out of the fiber capillaries. Snapshots taken after the completion of the green laser irradiation, i.e. following the 3 seconds draining period, confirmed that all capillaries were totally empty and free of water residue.

From the temporal characterization of the optofluidic actuation and the determination of filling and draining speeds, it also emerges that draining is considerably slower compared to filling process. The possible reasons behind this behaviour can be several. One explanation is related with the length of the wettability meniscus that is formed for the case of hydrophilic capillarity effects; from the contact angle measurements we estimated that the length of meniscus can be up to ~56 μm for a capillary diameter of ~10 μm. Such a type of hydrophilic meniscus can prompt an easier penetration of the water inside the MOF capillaries, leading to increased infiltration speeds. Another possible reason relies on the fact, that in the case of cavity filling the water is driven-in the MOF capillaries also by means of hydrostatic pressure exerted by the water drop; whereas on the contrary, for water draining upon ZnO transition to hydrophobic state the water has to be forced-out of the cavity against capillarity forces and hydrostatic pressure applied from the water drop that is always maintained at the fiber end face [24]. The above possible cases may explain why the draining progression is not as homogeneous as the corresponding filling process.

3.4 Investigations of the optically driven wettability transitionof ZnO thin films

Wettability behaviour is intrinsically driven by surface related defects and topology. There have been investigations before on the wettability behaviour of ZnO under thermal and/or ultraviolet radiation stimulation [18], utilizing X-ray photoelectron and diffraction spectroscopy [25], as well as, photoluminescence (PL) spectroscopy [26–28], the presence and nature of ZnO surface defects responsible for its wettability characteristics. We employ room temperature PL spectroscopy and spectrophotometric measurements for investigating the fingerprint of the wettability mechanism of ZnO films which are used as actuation transducers [see Fig. (6)]. All irradiation conditions used for these PL experiments were identical to the ones employed throughout the optofluidic cycle presented above.

 

Fig. 6 (a) Room temperature photoluminescence (PL) spectra of ZnO nanolayers synthesized within MOF cavities before and after laser exposures. (b) Spectrophotmetric measurements of ZnO films on silica substrates subjected to the optofluidic cycle UV and green laser radiation exposure conditions.

Download Full Size | PDF

Prior to any irradiation treatment when ZnO films exhibit a strongly hydrophobic behaviour, their PL spectrum is characterised by a strong emission band at ca. 390 nm, related to the material bandgap, and a weaker and spectrally broader band at ca. 550 nm. The spectrally extended PL band, centred at ~550 nm, is mostly associated with surface defect states; these defects can be particularly augmented in domains of high surface to volume ratio such as ZnO nanowires [29]. Accordingly, the spectrophotometric measurements [Fig. 6(b)] show that the pristine ZnO film, strongly absorbs below 390 nm, while longer wavelengths exhibit a good optical transparency.

Following UV exposure to either the pencil lamp or the excimer laser with high energy photons (~5 eV) for attaining hydrophilicity, several types of different defects are photoinduced irrespectively with their position within the ZnO bandgap. These photoinduced defects are located within the surface and the optical absorption length of the ZnO for the 248 and 254 nm wavelengths. In particular, for both PL spectral [purple and blue lines of Fig. 6(a)] the near band edge transition of the ZnO layers shifts to shorter wavelengths, while two distinct shoulders appear in the 420 nm vicinity; and the longer wavelength feature related to surface defects, peaking at ~550 nm is suppressed. On previous PL studies, assigned bands in the 420 nm regime are attributed to the existence of surface oxygen defects, as well as ZnO-SiO₂ interface defects [25,27]. The blue shift of the near band edge transition implies the possibility of size transformation of the ZnO grains under UV irradiation; while it can be also correlated with strain induced within the thin overlaid films inside the confined volume of the MOF capillaries [28,30]. The confined nature of the ZnO layers within the MOF micrometric size capillaries can amplify both the grain transformation and strain generation effects, which can affect the wettability of the films. The formation of ZnO grains of smaller sizes upon UV irradiation is prompted by the transformation of surface defects (possibly oxygen related), and the consequent structural rearrangements towards alleviating instabilities from an energetically point of view. We recall at this point that it is well known form previous studies that ZnO hydrophilicity strongly relies on the introduction of surface oxygen defects upon UV irradiation [18,31–33].

There is also a substantial difference between the ZnO overlayers PL spectra obtained using the two ultraviolet radiation sources. Namely, in the PL spectrum of the films irradiated using the cw 254 nm lamp source the ~390 nm bandgap peak has almost disappeared; whereas on the contrary this peak is shifted but remains strong in the PL spectrum corresponding to the ultraviolet exposure using the pulsed 248 nm excimer laser. Since the exposure of the ZnO overlaid MOF films using the 248 nm excimer laser results in enhanced thermal effects induced by the short pulse duration, the persistence of the strong bandgap peak at 390 nm of the PL spectrum for this exposure is related to laser annealing effects, which can partially restore structural (mainly strain) changes induced by ultraviolet photons into the thin film. On the contrary, in the low thermal load exposures using the 254 nm lamp, this bandgap peak is quite low in strength, further supporting this assertion. Simultaneously, the ~420 nm peak related to strain effects is stronger for the irradiation using the 254 nm lamp, depicting further the impact of the laser annealing during the 248 nm irradiation.

The optical absorption data for the UV irradiated ZnO films presented in Fig. 6(b), reveal that a strong absorption band is formed from ~700 nm down to 390 nm, with the short wavelength Urbach tail, being particularly increased; confirming structural changes induced within the material by UV radiation. The absorption at the 532 nm wavelength band is also manifold increased with respect to the pristine sample, which can promote photo-thermal effects.

Interestingly, the PL data of Fig. 6(a) (green line) reveal that the 532 nm laser irradiation cancels out the effects of the UV exposure regardless of the UV source employed for inducing hydrophilicity; these PL data reconfirm the recovery of the ZnO back to its original hydrophobic state upon green laser irradiation. In particular, the bandgap peak returns to its initial wavelength position, as prior to any irradiation, while the intermediate bands at ca. 420 nm diminish, and the surface defects related band located at ~550 nm reappears. Also, the spectrophotometric data denote a predominant recovery of the optical absorption of the ZnO films after the successive UV and green light irradiations, except a shift of the Urbach tail and an overall background loss increase.

After considering the above data, we support a light driven wettability switching mechanism for the ZnO films; especially in terms of the employed green light irradiation for the recovery of the ZnO hydrophobic nature. We believe that the green light mechanism has a twofold character: the green photons access shallow energy site defects which control wettability; whereas green laser radiation also accelerates photothermal changes.

While, the photon energy of the 532 nm green laser (E_green = 2.2 eV), cannot reverse structural modifications manifested at shorter wavelengths, instead it can access shallower energy level defects such as neutral oxygen vacancies, neutral zinc interstitials, and positively charged hydrogen [26,34]. The activation of the latter type of defects upon green laser irradiation can play a key role on the elimination of hydroxyl groups from the ZnO surface and the recovery of the hydrophobicity in a twofold manner. Firstly, the positively charged hydrogen reacts spontaneously with the negatively charged hydroxyl groups to re-form water, and secondly, it can create hydrogen bonds on the ZnO surface that block defective sites and shield them from other negatively charged groups (like hydroxyl) or atoms, and thus, diminishing surface adsorption. Additional evidence for this mechanism arise from the spectrophotometric measurements of Fig. 6(b), revealing that most changes induced by the UV radiation are reversed by the green laser, except those resting deeply into the bandgap of the material, i.e. near the absorption edge.

In other studies the above defects are thermally recovered, at slow recombination rates and temperatures up to 200 °C [18]. In our experiments the green laser irradiation can accelerate defect transformation, through photothermal interaction underlined by the increased optical absorption created by the prior UV irradiation [see blue line absorption data at Fig. 6(b)]. The photothermal interaction induced by the 532 nm laser irradiation was confirmed in numerical simulations performed using a heat transfer model applied in the regions occupied by silica, ZnO, and water assuming heat flux and temperature continuity on the interfaces while ressembling the capillary structure of the 25 nm ZnO overlaid HNA-5 MOF filled with water [23]. The optical absorption of ZnO at 532 nm assumed to be constant (not being photobleached) during laser irradiation; we expect that this will result in an overestimation of the final temperature increase. In these simulations the 532 nm laser irradiation intensity was kept at 2.29 W/cm² for a time slot of ~30 sec, while considering either beam focusing effects [35] (intensity amplification by a factor of ~1.5x) induced by the MOF cladding or plane wave beam. These photothermal simulations revealed that the maximum temperature increase in the ZnO layer was varying between 136 °C and 106 °C for the two beam focusing conditions aforementioned, with the corresponding temperature into the centre of the water filled capillary resting between 60 °C and 46 °C, respectively. The temperatures estimated using the above exposure conditions denote that there is an accountable photothermal component during 532 nm laser irradiation that can underline the reversing of the UV-switched hydrophilic ZnO into a hydrophobic state, which in turn prompts water drainage effects from the MOF capillaries. These photothermal results add to the direct defect recombination induced by the 532 nm photons, into a twofold wettability switching mechanism for the ZnO nanofilms.

4. Conclusion

In conclusion, the realization of a MOF Fabry-Perot optofluidic cavity with light-driven actuating characteristics was demonstrated, by utilizing the reversible and controllable wettability properties of ZnO overlayers synthesized within the MOF cavity. The actuation principle relies entirely on the use of light sources, and thus, findings presented herein appear highly promising towards the design of optically driven liquid pumps and actuating devices. Furthermore, the effect of the ZnO light actuated wettability properties is under investigation for the development of alcohol sensing probes [36,37] with tuneable sensitivity characteristics. Further spectrophotometric and structural characterisation investigations for consolidating this light driven wettability mechanism proposed for the synthesized ZnO thin films are underway.