1. Introduction

Integrating several functionalities in a miniaturized chip has favored the creation of compact and versatile lab-on-chip (LOC) devices capable of performing relevant and useful functions, comprising biological and genome studies, synthesis of drugs and environmental hazards detection [1–4]. In the attempt to realize micro total analysis systems (µ-TAS), i.e. miniaturized devices that include all necessary key-components for automated biochemical analysis [5,6], the on-chip implementation of more and more physical functionalities considerably increase the analytical capability of the devices. For example, the integration of conductive electrodes within the chip is highly demanded in order to provide electrical functionality. In fact, the ability to control and establish electric fields across fluids gives rise to novel applications in sensing and microanalysis, di-electrophoresis, electrokinetic phenomena, magneto-hydrodynamics and micro-nuclear magnetic resonance [7–12].

Furthermore, in recent years, optics and fluidics have been used synergistically to introduce novel functionalities and new measurement possibility in LOC platforms leading to a new research field named Optofluidics, which is now subject of intense scientific and technological interest [13].

Although on-chip optical and electrical stimuli and characterization represent an important improvement in the LOC field alone, their coupling as well as some elements that would benefit from their synergy are still missing. Succeeding in this endeavor could impart an impressive upgrade to the final device. A decisive and important advancement could come from the possibility of having embedded 3D electrodes cleverly positioned around the microfluidic channels without restriction of geometry and shape, capable of interacting with optical elements in the chip. Importantly, such integration would be even more powerful if it allowed, thanks to semi-transparency, the coupling of a probing light beam with the microfluidic part, thus realizing an electro-optofluidic platform for advanced LOC devices.

Different approaches have been proposed to realize on-chip microelectrodes, based on metals, either evaporated through shadow masks [14] or patterned with photolithographic processes [15,16], on reduced graphene oxide [17] and on laser modified diamond [18]. However, all the reported approaches require many fabrication steps and the frequent use of expensive processing masks. Such requirements slow down the technology prototyping and realization, increasing the devices costs without a substantial gain in performance compared to the most common architectures exploiting either Indium Thin Oxide (ITO) or Fluorine-doped tin oxide (FTO). In the latter cases, ITO or FTO electrodes are patterned on a transparent substrate exclusively with a planar geometry. Importantly, the limitation to co-planar 2D electrodes extends even to the most sophisticated process proposed so far. Unfortunately, this can be both detrimental to the versatility and compactness of the final devices, as well as reducing the range of applications. These limitations result in the possibility to apply electric field exclusively in a single direction, typically orthogonal to the substrate plane. Integration of 3D electrodes has instead obvious advantages, as it would allow in principle to apply directional electric field over the entire solid angle.

To overcome such strong restrictions and to achieve more flexible process flows, in the recent past great attention has been paid to the fabrication of microfluidic structures by femtosecond laser irradiation followed by chemical etching (FLICE), in transparent substrates like fused silica [19]. FLICE is a robust and completely mask-less technique where a focused femtosecond laser pulse is used to locally change the material properties. As a result, a permanent periodic redistribution of the highly localized material density is achieved in the laser spot volume only, defining the desired 3D structure in the bulk of fused silica. In this way, a chemical etching of the laser modified zone, by a strong acid or base (typically HF or KOH, respectively), removes the irradiated volume, producing the hollow structure. This manufacturing method is emerging as a flexible platform to obtain, in principle, any surface or 3D geometry buried in the substrate for microfluidics applications, such as, for example, microchannels, basins, micro chambers and micro coils, largely simplifying the fabrication process [20–24]. For all these reasons, FLICE has been more and more exploited as a powerful tool towards sophisticated lab-on-chips, and in the recent past it has also been adopted for the fabrication of micro-electric components thanks to coupling with conducting materials [25]. 3D electrodes [26] and micro-coils [27,28] embedded in the substrate with both solid and liquid conductors were demonstrated, but their applications are still limited because of the remarkable difficulties in inserting conductive materials into their final containment structure as well as their opaqueness. In fact, semi-transparent conductive electrodes would facilitate light coupling or probing enabling the developments of novel high-performance optoelectronic devices, both for pure real-time vision investigations and for spectroscopic analysis.

Overall, while different approaches have been exploited to integrate a semi-transparent electrodes in LOCs, including fine meshes [29], transparent oxides [30,31] and ultrathin metallic films [32], none has been coupled with a process allowing 3D patterning. In fact, the possibility to define an electric field along directions that are different from the orthogonal to the plane of growth, i.e. the only one available at the moment using 2D patterned electrodes, can expand the range of applications and push forward all those optofluidic researches in which in-plane electric stimuli are required. Furthermore, even in cases where there was an initial attempt to integrate 3D electrodes into a LOC platform, transparency was not reported and very complex and inflexible manufacturing processes were required, often compromising their actual use. For all these reasons, a solution to combine truly 3D opto-microfluidics devices with fully embedded 3D transparent microelectrodes is very desirable but has not yet been successful proposed. Here we demonstrate an effective methodology to realize 3D transparent microelectrodes integrated into a fused silica slab by combining FLICE with drop-on-demand inkjet printing, a very flexible, digital and additive printing technique [33,34]. In our all direct written approach, FLICE fabricated 3D microcavities are precisely filled by dispensing controlled volumes of the well-known poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) polymer conductor that combines good electrical conductivity with high transparency in the visible range [35]. In order to demonstrate the potential of this approach for integrated electro-opto-fluidic systems, we realize a first proof-of-concept LOC device where the polarization state of liquid crystal (LC) molecules contained in a microchannel is controlled by 3D electric field applied in two orthogonal directions. The possibility to couple electric field in 3D paves the way to a novel class of lab-on-chips equipped with 3D micro-opto-fluidics.

2. Materials and methods

2.1 Substrates

High-purity quartz-based substrates (fused silica, type III, JGS1, FOCtek), with dimensions of 20 mm X 20 mm X 2 mm, were adopted.

2.2 Liquid crystal

The microchannel was filled with a nematic liquid crystal 4′-Pentyl-4-biphenylcarbonitrile (5CB) by Sigma-Aldrich. A liquid crystal is an optically anisotropic material, i.e., the propagation of light waves in the medium depends on its direction and polarization and therefore shows an anisotropic refractive index (RI): ordinary (n₀) and extraordinary (n_e) RI can be distinguished. 5CB at 633nm is characterized by n₀ = 1.53 and n_e = 1.72. n_ISO = 1.59 (randomly oriented).

2.3 Laser and laser writing setup

We adopted a re-generatively amplified Yb:KGW system (Pharos, Light Conversion), with fundamental harmonic at 1030 nm, delivering 230 fs long pulses with a maximum repetition rate of 1 MHz. The pulses were frequency doubled to second harmonic using an external BBO crystal obtaining a 515 nm wavelength. These pulses were then focused inside the substrate using a 50X objective with a numerical aperture of 0.42. Computer-controlled, 3-axis motion stages (ABL-1000, Aerotech), interfaced by CAD-based software (ScaBase, Altechna) with an integrated acousto-optic modulator, were used to translate the sample with respect to the laser beam.

2.4 Chemical wet etching

The etching step is performed by inserting the laser modified substrate in a 20% HF aqueous solution, all immersed in an ultrasonic bath, maintaining a temperature of 35 °C.

2.5 Inkjet printing of conductors

After the realization of the microcavities and before the printing process, the substrates were cleaned by rinsing them with water, acetone and isopropanol and then by exposing them to oxygen plasma at 100 W for 10 minutes. Silver interconnections (InkTec Tec-IJ-060) and contacts between silver lines and microelectrodes (Heraeus Clevios P Jet 700) were inkjet printed with a Fujifilm Dimatix Materials Printer DMP-2831. After the deposition of the silver lines, a thermal sintering at 180 °C followed.

Transparent 3D electrodes were realized by dispensing a PEDOT:PSS ink (OrgaconTM ICP 1050) in the micro cavities with a Jetlab 4xl-A inkjet printing system, adopting a nozzle with a diameter of 50 µm. A thermal drying at 110 °C on a hotplate followed.

2.6 Final chip assembly

After the realization of the 3D microelectrodes and of the interconnections, the chip was completed by inserting a micro-tube, to fill the microchannel with the 5CB liquid crystal, and the optical fiber with the help of micro-precision translation stages. External wires were connected to the printed silver pads to allow the application of a voltage bias through a voltage supply.

2.7 Chip test setup

The chip was mounted between the cross polarizers of an optical microscope. Through a waveform generator, a voltage square waveform at 1 kHz with a peak-to-peak voltage of 20 V was applied to the electrodes. An optical fiber (P1-630A-FC-1 - Single Mode Fiber Patch Cable, 633-780 nm) was connected to a laser source (TLS001-635-T-Cube He-Ne) whose emission peak is centered at 635 nm.

3. Results and discussion

3.1 Design of the chip

In order to provide a proof-of-concept device demonstrating the functioning of the electrodes in applying electric fields in 3D and their applicability in an on-chip configuration, the design shown in Fig. 1 was conceived. The chosen substrate is transparent fused silica, a common and largely adopted platform [36,37]. The device consists of a 1.5 mm long microchannel sandwiched, in two different positions, between two pairs of vertical and planar electrodes. The central part of the microchannel has a regular square cross-section with a ∼ 60 µm side.

 

Fig. 1. A) Top-view sketch of the electro-opto-fluidic LOC configuration with a microchannel sandwiched between two pairs of vertical and planar electrodes and aligned with an external coaxial optical fiber. B) Prospective view sketch of the device and C) zoom of each electrodes pair. D) Microscope images of the top-view of the device after etching and E) from the substrate sidewall displaying various components. F) Microscope image of planar electrodes after inkjet filling with the semi-transparent conductive ink.

Download Full Size | PDF

This shape and size were chosen to simultaneously fulfil several requirements. First of all, the square section of the microchannel prevents any type of geometrical induced pre-alignment of the liquid crystal, which would instead occur in the case of a marked difference in size [38], and to ensure the application of a uniform electric field in both directions. Furthermore, the LC filled microchannel has a RI bigger than that of the substrate, becoming an optical waveguide with a core at least 10 times larger than that of the input optical fiber thanks to the chosen dimensions. Owing to this, the electromagnetic wave can cross a considerable volume of LC devoid of any point defect singularities, caused by possible mini turbulences or resonant interactions at wall surfaces, that would affect the travelling optical signal. The two cylindrical openings at two ends of the chip, perpendicular to the channel direction, were designed to insert micro-tubes and to enable the smooth filling of the 3D microchannel with the liquid crystal (Fig. 1).

Suitable slots buried in the substrates are designed to host conductive inks and therefore to realize the 3D electrodes. The pair of vertical electrodes are placed on both sides of the central microchannel along the y direction. The pair of planar electrodes are instead positioned on both sides of the microchannel along the z direction and provided with openings on the surface for bubble-free filling. A coaxial slot is foreseen in line with the microchannel, of about 1 mm long, and compatible for the insertion of an optical fiber with a 125 µm external diameter to allow the coupling of the probe He-Ne laser into the chip.

3.2 Fabrication — FLICE

The first step in the fabrication of the integrated monolithic chip is the creation of the microfluidic circuit and of the electrodes structures by FLICE. This technique consists of two steps: direct writing by ultra-short pulses and wet etching to remove the laser modified zones. As far as the direct writing is concerned, a femtosecond laser beam at 515 nm is focused inside the substrate, whose linear absorption is negligible at such wavelength. Only in the focal spot of the laser, thanks to the high intensity, a non-linear multi-photon absorption mechanism produces well localized and precise modification of the material, creating internal nanocracks and leaving unmodified all the rest of the glass invested by the beam. The irradiation patterns and parameters have a critical role in the final result, therefore they have been specifically optimized for each component. In fact, it is well known that the internal roughness left by the FLICE manufacturing technique is closely related to the fabrication parameters as a function of geometry, beam polarization, irradiation depth and energy density [39]. In order to obtain buried structures as smooth as possible, the writing energy density and - in particular - the writing paths used (laser lines crossed in two orthogonal directions on the same plane) contributed significantly to the reduction of internal roughness. In our case the irradiation pattern parameters are: scan speed of 10 mm/s, average power of 200 mW for vertical electrodes, 250 mW for top planar electrodes and microchannels, and 300 mW for bottom planar electrodes, respectively. The central square microchannel section has been obtained by making an optimized distribution of the laser writing lines in the volume (2 µm in y direction and 10 µm in z direction) along the whole length of the microchannel. The circular openings have been made with cylinders of variable radius, whose contour is made of straight lines, starting from the smallest one. Each edge within the microchannel can give rise to a defect in the LC bulk whose topological charge is not equal to zero. For this reason, exploiting the three-dimensional manufacturing capabilities of our fabrication technique, a special connection between the two different geometric sections (from circular to square) has been made during the writing process to make the duct as smooth as possible. The fabrication of the vertical electrodes was performed by keeping in mind the subsequent inkjet printing steps. For such reason, the structures hosting the conductive inks had to be realized as micro-trapezoidal slots with an opening on the surface. Indeed, one of the walls was skewed, as shown in the Figs. 1(B) and 1(C), to allow an easy access for the jetted ink droplets and avoid any source of air bubble. From the geometric point of view, the electrodes pair is created by forming 800 µm deep micro slots in the substrate, placed 150 µm away from the microchannel on both sides. The fabrication consisted in an array of lines along the z direction, with the pitch between lines being 2 µm in both x and y directions.

Planar electrodes, on the other hand, are buried in the substrate. In order to allow optical access to the microchannel (a clear vision), they have the additional feature of being semi-transparent for a suitable length (in our case about 350 µm). To obtain this peculiar and innovative characteristic in an electrode buried in the substrate, in addition to use a transparent conductive ink, an important role is played by the roughness of the internal surfaces left by the FLICE manufacturing process. To reduce the undesirable effects that excessive roughness has on vision, the planar electrodes consist of two rectangular slots written with a series of planes composed of crossed lines along the two directions. The pitch between the lines is of 2 µm both in x and y. The pair is placed 150 µm away from both sides of the central microchannel.

Given the distribution along z of the various components, the fabrication had to start with the bottom-most component, which is the bottom planar electrode. This was followed by the microchannel, the top planar electrode, the vertical electrodes and finally by the fiber slot. Following the laser irradiation, an etching step is performed. The light-induced nano-cracks have a higher etching rate with respect to the pristine fused silica. They also act as pathways for the penetration of the acid, which, supported by the action of ultrasound cavitation, allow to produce the embedded 3D micro-structures. Figure 1(D) shows a microscope image from the top of the fused silica substrate after the etching step. The 3D structure of the whole device, schematically represented in the Fig. 1(B), is clearly visible also in the microscope image of the device taken from the side of the chip (Fig. 1(E)).

3.3 Fabrication — inkjet printing of the conductors

Inkjet printing was adopted both for the realization of the embedded 3D microelectrodes and for the interconnections between the microelectrodes and external wiring pads. First, the interconnection lines have been printed with a silver ink, subsequently sintered, on both planar surfaces of the substrates. Also in the case of the vertical electrodes a silver ink was printed since the chosen geometry does not obscure the microchannel for subsequent viewing. In order to ensure the uniform coverage of the vertical surfaces, the filling and sintering process was repeated three times. Conversely, the planar electrodes had to guarantee optical access to the microchannel. This was obtained by adopting a PEDOT:PSS conductive ink, that has the additional feature of being highly transparent in the visible. The entire volume of the planar microcavities was filled and then subjected to a mild drying step. Figure 1(F) shows a microscope image of the planar electrodes following the inkjet printing of conductors: as it can be appreciated, these electrodes remain transparent even after filling with the conductive inks.

3.4 Electro-optical tests

In order to demonstrate the functionality of the 3D microelectrodes, and the possibility to apply electric fields in 3D across the microchannel filled with the liquid crystal 5CB, we first coupled the probe laser through the aligned fiber optics. Thanks to the higher effective refractive index of the LC, a wave-guiding effect confining the light within the channel occurs. Since the waveguide performances are strictly connected to the difference of the refractive index between the LC waveguide and the substrate, its efficiency can be controlled by the application of a voltage across the LC channel. Experiments to assess the possibility to manage the LC director with the electric field (the 5CB is a positive LC) were performed by applying a voltage square wave with variable r.m.s. (V_rms) to the electrodes. In particular, we performed experiments in 3 different steps. In the 1^st and the 2^nd step, the vertical and planar set of electrodes were controlled individually, to observe independently the effect of the two orthogonal applied fields on the LC molecules orientation inside the microchannel. In the 3^rd and final step, the test was repeated with both the sets of electrodes operating simultaneously. During such experiments, we monitored the effect of the applied fields on the LC alignment by cross polarized imaging analysis under an optical microscope.

The application of a potential difference across the microchannel forces an alignment of the randomly distributed LC director along the electric field vector, with the final angle of orientation depending on the value of the applied voltage. Upon reorientation of LC, the RI of the inside of the microchannel changes, while the outside RI is unchanged. As a result, the performance of the optical guide is affected. The fraction of the input power that is guided towards the end of the channel with respect the complementary part, which is either refracted, due to radiation phenomena in transverse direction, or lost due to scattering, will also change. A modulation of the intensity of the outgoing light captured by microscope objective in a direction orthogonal to the optical guide is therefore expected [40,41].

3.5 Test of the vertical pair of electrodes

A square wave with an increasing r.m.s. voltage value was applied to the vertical pair of electrodes, resulting in the increase of the electric field across the microchannel and thus the LC order. The region of interest (ROI), clearly indicated with a rectangular yellow box in Fig. 2, is restricted to the region of the microchannel that is under the influence of the electric field between the vertical electrodes, where all the channel area is displayed. Figure 2(A) shows the microchannel with V_rms = 0 V, therefore with a minimum of light intensity observed through the microscope. A gradual change in the orientation of the LC inside the microchannel can be optically detected with different applied V_rms. Figure 2(B) shows the channel under the application of 5 V_rms, where an increase in light intensity is noticed. Further increases of V_rms complete the orientation of the liquid crystal until no considerable changing in light intensity is detected. Such condition was achieved starting from the application of the 8 V_rms (see Fig. S1 in Supplement 1), while at 10 V_rms it remains well fixed to its plateau. Figure 2(C) shows the latter condition corresponding to an average electric field of 50 mV_rms/µm. At this point the voltage was gradually reduced until the LC returned to the original random orientation. In Fig. 2(D) V_rms is again 5 V, and light intensity is correspondingly similar to Fig. 2(B), while finally in Fig. 2(E) V_rms = 0 V, and light intensity reaches again its minimum as in Fig. 2(A). It is important to highlight that the change in orientation of the LC is restricted to the ROI, thus demonstrating the proper operation of the vertical electrodes. In fact, the increase in LC order caused by the alignment of its director in the transverse direction (TE) reduces its RI, which gets closer to the one of the surrounding silica glass. This leads to an increase in the amount of light reaching the observation camera, through the top analyzer, due to the increased loss of the waveguide in the ROI. The adjacent area instead, where the planar electrodes are positioned and the external stimulus is not applied, is clearly unaffected by the field modulation in the vertical electrodes. In this case, the RI shown by a volume of LC completely randomly oriented is higher than that of the aligned one, increasing the optical performance of the waveguide. Moreover, the polarized light passing through a volume of LC randomly oriented is split between all possible light polarization directions. Of all the polarizations that will reach the top analyzer, only one will be allowed to cross it. This is the reason for the lower intensity observed through the camera.

 

Fig. 2. Testing of vertical electrodes: (A, E) Images of the microchannel containing the LC with no field applied, i.e. at V_rms = 0 V; (B, D) at an intermediate applied voltage V_rms = 5 V; (C) at the applied voltage of V_rms = 10 V, over the threshold voltage where no further change in polarization of LCs is noticeable (see Supplement 1). The field is applied only in the area within the yellow box.

Download Full Size | PDF

3.6 Test of the planar pair of electrodes

The same experiment was repeated with the planar electrodes, where it is the TM polarization state of the travelling wave to be affected. Microscope images of the microchannel with various applied V_rms are shown in Fig. 3. A visual inspection of the light transmission through the electrodes is possible thanks to their high transparency. This is a key feature of our 3D embedded electrodes towards opto-electro-microfluidics systems. Figures 3(A) and 3(E) show the channel with V_rms = 0 V, where the light that crosses the polarizers and reach the camera is at its minimum. The light intensity increases when 5 V_rms are applied (Figs. 3(B) and (D)), and it remains well fixed to its maximum plateau at 10 V_rms, corresponding to an average electric field of 35 mV_rms/µm (Fig. 3(C)). Also in this case, at all voltages the modulation of the LC domains is restricted to the ROI (yellow box in Fig. 3), which here corresponds to the area of the microchannel below the planar electrodes. Our observation demonstrates that also through the highly transparent planar electrode it is possible to locally control the waveguide performance by managing the molecular orientation of the LC in the microchannel.

 

Fig. 3. Testing of planar electrodes: (A, E) Images of the microchannel containing the LC with V_rms = 0 V, (B, D) with V_rms = 5 V and (C) with a voltage (V_rms = 10 V) beyond the threshold which no further change can be recorded.

Download Full Size | PDF

3.7 Simultaneous test of both pairs of electrodes

Having verified that through the application of a voltage at the two pair of vertical and planar 3D microelectrodes, one at a time, it is possible to control the LC orientation, we tested the possibility to alter at will the polarization state of a light wave travelling through the microchannel by operating both electrodes at the same time. Figure 4 reports the microscope images of the channel through a polarizer at different applied voltages. This experiment clearly demonstrates a simultaneous change in the electromagnetic polarization state in a LC waveguide by an electric field applied through the vertical and planar electrodes at the same time, leading, in the corresponding ROIs, to an increase of the outcome light radiation from the LC waveguide to the microscope objective. We have therefore shown that by operating the two pair of electrodes a TE and a TM polarization state of the travelling wave can be achieved applying orthogonal electric fields at the same time in different sections of the microchannel. Interestingly, in between the two sections a peculiar twist condition of the LC can be induced. Such feature could be exploited and optimized for the creation of new active optical elements integrated in optofluidic platforms, such as a buried in-plane wave plate or polarizer, electrically tunable, which are still missing today.

 

Fig. 4. Simultaneous testing of both the electrodes: (A, E) Images of the microchannel containing the LC with V_rms = 0 V, (B, D) with V_rms = 5 V and (C) with a voltage of V_rms = 10 V, beyond the threshold voltage which no further change can be recorded.

Download Full Size | PDF

4. Conclusions

We have proposed an effective methodology to realize both opaque and semi-transparent 3D microelectrodes completely buried into a fused silica slab by combining two direct writing manufacturing technology: anisotropic wet etching controlled by femtosecond laser irradiation with drop-on-demand inkjet printing. Taking advantage of the fabrication capabilities of the FLICE technique to realize 3D buried microcavities, two main configurations of micro electrodes have been realized: vertical and planar. Such microcavities were then filled by dispensing controlled volumes of conductive inks. In particular, we adopted both a silver ink and a polymer based conductor, PEDOT:PSS, which combines high transparency in the visible range with good electrical conductivity. In order to demonstrate the potentiality of integrated electro-opto-fluidic systems enabled by our 3D electrodes, we realized a first proof-of-concept device where the polarization state of liquid crystal molecules contained in a microchannel is controlled by an electric field applied in two orthogonal directions in space. Operating with the two pair of electrodes at the same time, TE and a TM polarization state of the travelling wave can be achieved in different sections of the microchannel. Thanks to the tuning of the voltages applied to the electrodes from 0 to 10 V_rms, the intensity of the outgoing light from the liquid crystal waveguide oscillates from a minimum to a maximum. The possibility of coupling electric fields onto fluids in 3D has thus been demonstrated, overcoming the 2D limitation of current lab-on-chip technologies.

Such powerful combination of optical and electrical functionalities into a monolithic chip configuration paves the way for a vast range of applications requiring visual analysis and imaging in an electro-opto-fluidic device and/or exploiting electrically controlled light manipulation in microfluidic chips, such as on chip polarization rotators, active lenses, light intensity modulators and tunable laser sources.