On‐Demand Inkjet Printed Hydrophilic Coatings for Flow Control in 3D‐Printed Microfluidic Devices Embedded with Organic Electrochemical Transistors

Microfluidic surface chemistry can enable control of capillary‐driven flow without the need for bulky external instrumentation. A novel pondered nonhomogeneous coating defines regions with different wetting properties on the microchannel walls. It changes the curvature of the liquid–air meniscus at various channel cross‐sections and consequently leads to different capillary pressures, which is favorable in the strive toward automatic flow control. This is accomplished by the deposition of hydrophilic coatings on the surface of multilevel 3D‐printed (3DP) microfluidic devices via inkjet printing, thereby retaining the surface hydrophilicity for at least 6 months of storage. To the best of our knowledge, this is the first demonstration of capillary flow control in 3DP microfluidics enabled by inkjet printing. The method is used to create “stop” and “delay” valves to enable preprogrammed capillary flow for sequential release of fluids. To demonstrate further utilization in point‐of‐care sensing applications, screen printed organic electrochemical transistors are integrated within the microfluidic chips to sense, sequentially and independently from external actions, chloride anions in the (1–100) × 10−3 m range. The results present a cost‐effective fabrication method of compact, yet comprehensive, all‐printed sensing platforms that allow fast ion detection (<60 s), including the capability of automatic delivery of multiple test solutions.


Introduction
Capillary-driven microfluidic devices take advantage of capillary forces to move liquids without using any external pumps or valves, this occurs due to the intermolecular forces of the liquid and the surrounding solid surface within the microfluidic device. Such microfluidics has paved the way for implementing lab tasks into point-of-care (POC) devices, independent from complex peripheral equipment.
Geometrical changes, such as a sudden expansion in the microchannel cross-section, can modify the forefront meniscus of the liquid-air interface and, consequently, the driving force of the capillary flow. So, it is used to obtain the desired flow control. Autonomous capillary-driven microfluidic devices having microchannels with varying geometries have been proposed in the literature, to enable pre-programmed immunoassays based on sequential delivery of bioliquids and reagents. [1][2][3][4][5][6] Besides utilizing the geometry to manage the air-liquid meniscus and control the capillary flow, the use of surface chemistry within microfluidics can be exploited for spontaneous manipulation of liquids. Different microfluidic coatings lead to various liquid-solid contact angles. It changes the curvature of the liquid-air meniscus in various sections. Consequently, different capillary pressures are accessible at different stages. Zhao et al., formed virtual walls by creating different surface free energies through patterning of the microchannels using two methods based on self-assembled monolayer (SAM) chemistry. The first method created multistream liquid laminar flows (octadecyltrichlorosilane solution in hexadecane) to deposit SAMs on the microchannel sides. The areas where the silane solution stream passed were modified with SAMs and became hydrophobic, while areas where the nonsilane solution stream passed remained hydrophilic. This method requires pre-formed channels to drive multi-stream laminar flows, limiting the possible final patterns. The second method relied on a two-step process, where the whole surface of channels was first modified with photocleavable SAMs (2-nitrobenzyl), followed by ultraviolet (UV) irradiation of the microchannels to create hydrophilic patterns via photomasks. [7,8] UV irradiation through the masks on top of SAM-modified channels produced hydrophilic carboxylate groups in the exposed areas forming complex patterns. Furthermore, photolithography or dispensing (plotter pen) of hydrophilic patterns on hydrophobic surfaces as well as hydrophobic patterns on hydrophilic surfaces were performed to confine liquids into channels instead of limiting them by walls. [9,10] The dominant materials used for the fabrication of microfluidic devices, like resin in 3D-printing (3DP) and polydimethylsiloxane (PDMS) in soft lithography, are poorly hydrophilic or hydrophobic, while hydrophilic interfaces (water contact angles of ≤60°) are needed to satisfy the required surface energy of such capillary microfluidics within water-based applications. [1] By treating these polymers with solely oxygen/air plasma and additionally with polyethylene glycol (PEG) or polyvinyl alcohol (PVA), hydrophilic surfaces are created by the introduction of polar functional groups containing oxygen, etc. However, the surface characteristics are gradually lost toward the initial state within a few hours of aging. [11] On the other hand, the polar functional groups introduced by plasma treatment provide suitable adhesion of any subsequently deposited coating. [12] Therefore, different coatings were considered to increase and keep the stability of the hydrophilic surface for a longer period. Hydrophilic PDMS-based microfluidics can be obtained by the deposition of PVA and polyethylene glycol PEG before and after plasma treatment, at the cost of additional processing steps. [13,14] Digital printing technology, more precisely inkjet printing, has been employed for paper-based microfluidics by creating 2D structures (hydrophobic boundaries) on paper and by inkjetetching of hydrophobic substrates. Moreover, inkjet printing was used for level-by-level (several materials) fabrication of microfluidics with 3D features as well as printing of embedded sensors along with microfluidic channels in bilayer microfluidic chip platforms. [15][16][17][18][19][20] The inkjet printing technology ensures a low cost, scalable, and material/energy (2.4-10 pL ink drop volume) efficient manufacturing approach. The major advantage in comparison with other printing techniques is explained by the digital printing process featuring low ink consumption, high resolution deposition (≈30-40 μm), customizable on-demand shape and thickness of the printed structures including a large selection of materials on a variety of substrates. [21,22] To the best of our knowledge, inkjet printing of hydrophilic coatings onto microchannels of stereolithography (SLA) 3DP microfluidics has not been explored till now.
This study suggests a one-step surface modification method via fast and low-cost inkjet printing of a PSS − Na + -based hydrophilic coatings (several layers) on the surface of the multilevel 3DP resin to enable flow control in complex capillary microfluidic devices. On the contrary to plasma treatment, or other methods (in combination), this approach allows the creation of ondemand surface hydrophilicity at the place of interest (e.g., reservoirs, valves, etc.). The wetting properties of the inkjet printed coating inside the 3DP microfluidic channels retain for a long period of at least 6 months, see Figure S1a and Movies S1 and S2 in the Supporting Information.
To demonstrate the potential of our 3DP capillary microfluidics with inkjet printed hydrophilic coatings, we integrated fully screen printed organic electrochemical transistors (OECTs) within the microfluidic devices, thereby forming an all-printed sensing platform. The integration of OECTs with a surfacedirected microfluidic system, manufactured via photolithography and without the preprogrammed sequential release of liquids, has been previously reported. [23] In our work, the functionality of the coatings inside the 3DP multilevel microchannels was verified by sensing Cl − anions in the (1-100) × 10 −3 m range by the screen printed OECTs. The OECT technology, possibly integrated within microfluidic devices, shows a great potential for printed logic circuits and various biosensor applications, e.g., for quantitative testing of glucose, lactate, cholesterol, and label-free DNA biosensing. [24][25][26][27][28][29][30]

Microfluidic Function
Capillary flow is driven by the capillary pressure gradient caused by the arc meniscus front. This meniscus is formed by the surface tension applied on the liquid-air interface, which comes from the balance of the liquid-surface adhesion and the liquid intermolecular cohesion. Two-phase meniscuses across the capillary circuit provide driving and/or retention capillary pressure along or against the capillary flow. The capillary pressure can be calculated by employing the Young-Laplace equation based on channel size and liquid contact angle with the surface. [31] In a closed rectangular microchannel, the capillary pressure can be obtained according to Equation (1) [2,32] where and P is the surface tension of the liquid and the pressure, respectively. t , b , l , and r is the top, bottom, left, and right water contact angle on the respective channel wall. h is the channel height and w is the channel width. According to Hagen-Poiseuille's law, the pressure drop (ΔP) due to viscous forces on the walls can be described using the flow rate (Q) and the fluidic resistance (R). In the case of a rectangular cross-section channel, R can be calculated using the length (L) of the microchannel by Equation (2) Considering Equations (1) and (3), the direction and the flow rate of a capillary flow can be calculated, and the flow can be pre-programmed based on the parameters that change the pressure and the resistances. The schematics in Figure 1 illustrate the modified capillary-driven microfluidics with hydrophilic microchannels, including the detection site with an embedded OECT sensor device. Figure 2 shows 3DP capillary circuits, whose functionality are based on predesigned parameters such as channel dimensions and contact angles. Figure 2a shows the resulting force balance (capillary force-viscous force) on the liquid bulk. An arc meniscus front drives the liquid into a microchannel covered by a printed hydrophilic coating. Different deposition strategies can lead to different contact angles, and as a result, various flow patterns. Figure 2b depicts changes in contact angle when the PSS − Na + -based hydrophilic coating is deposited on a hydrophobic surface. These changes in contact angles lead to a different force balance. Figure 2c shows the force balance differences in two parallel channels, caused by the printed coating that leads to retention or capillary flow. In channel 1, reaching the noncoated surface completely stopped the flow, while in channel 2, the same aqueous liquid flowed into the channel by the capillary force.
Altering the channel geometry by 3DP and the contact angle by the deposition of the coating allows for flow control. In analogy with electronic circuits, each two-phase meniscus can be represented as a potential source and the channels as resistances (Figure 2d). By applying Equations (1) and (3) to calculate pressure (potential difference) and flow resistances, the sequential release of liquids can be pre-programmed ( Figure 2e). As a result, the liquids from each reservoir (from 1 to 3) are released downstream toward the detection site and the capillary pump, where the flow rate is controlled by the suction power of the capillary pump and the sum of the flow resistances.

Ink Formulation, Inkjet Printing Conditions, Wettability, and Stability of the Coatings
Herein, the ink for coating was prepared by dissolving 15 wt% poly(sodium-4-styrene sulfonate) (PSS − Na + ) in deionized water at ambient conditions, followed by the viscosity measurements to verify its suitability for inkjet printing. An ink viscosity of ≈6.7 mPa s was obtained, which falls within the range of the required viscosity for inkjet printing. It is well established that the ink and film properties can be improved by mixing a low boiling point solvent, being the main solvent, with a high boiling point solvent as the co-solvent. [33] Therefore, the prepared solution was further optimized by adding 10 wt% of poly(ethylene glycol).
Physical properties, such as viscosity, substrate surface free energy and contact angle are crucial for the film formation on various substrates. [34] The mixture of water and poly(ethylene glycol) (90:10 wt% ratio) was employed to prevent clogging of the nozzle and fast evaporation of the ink at the nozzle-air interface. [35] The inkjet printing process was performed on quartz glass substrates and 3DP microfluidic chips to assess and compare printability. The inkjet printing on quartz glass and 3DP substrate was carried out by several subsequently printed layers at a print resolution of ≈1000 dots per inch (DPI). A resulting coating thickness of 570 nm (and a minimum linewidth of 35 ± 5 μm) was obtained on quartz glass ( Figure S2, Supporting Information), while 3DP substrates resulted in nonhomogeneous coatings ( Figure S3, Supporting Information). This could be explained by the high surface roughness of the 3DP substrate (>6 μm, which exceeds the thickness of the PSS − Na + -based coating) and its poor hydrophilicity. It is well known that the quartz glass has good wettability and low surface roughness. As a result (Table 1), the designed widths and the actual widths of the printed patterns on 3DP substrates (hydrophobic) and quartz microscope glass slides (hydrophilic) are differing, which is an indication of different surface properties.
As already mentioned, the surface free energy, the contact angle of the substrate and the viscosity of the ink are important parameters that enable uniform and homogeneous film formation. To investigate the wettability of the 3DP resin and the PSS-Na+-based hydrophilic coating, the water contact angles of the two surfaces were measured. The deposited coating on the 3DP substrate resulted in reduction of the hydrophobicity (water contact angles changed from 69.7°± 0.57°to 40.1°± 7.85°), see Figure 3. The measurements indicated a hydrophilic surface (good wetting) on the PSS-Na+-coated resin and a hydrophobic feature on the 3DP resin. The coating is mainly envisioned for single use, applicable in disposable sensor systems. The PSS-Na+ is water soluble, which implies that the coating gradually dissolves over time, or absorbs moisture when stored at ambient conditions (22°C and 47RH%). The latter resulted in even lower water contact angle (≈31°) after 14 days of storage, see Figure 3. The suitability of using inkjet printed coatings inside 3DP microfluidics for sensor systems was successfully assessed after 6  months of storage in a plastic bag, i.e., the coating retained its surface hydrophilicity after long-term storage ( Figure S1a and Movies S1 and S2, Supporting Information).

Inkjet Printing of On-Demand Patterns
The drop-on-demand inkjet printing technolcogy has the potential to expand the frontiers of 3DP capillary-driven microfluidics to modulate, for example, the release of fluids or their flow via so-called "stop" and "delay" valves that will be introduced in this section.
The inkjet printed polyelectrolyte coating on the surface of the 3DP resin, reported herein, reduces the contact angle and changes, in this way, the wetting behavior of the resin-based substrate to hydrophilic, thereby making it suitable for capillary microfluidics. In Figure 4a, we show an on-demand inkjet printed pattern on quartz glass, including ≈400 μm wide stripes without the coatings. As a step further, the printing was adapted for the deposition of the same coating pattern on the 3DP resin of a microfluidic chip.
The deposited printed array was designed to manipulate (stop or delay) the flow of the liquid (a water solution) inside various channels of microfluidic chips. In particular, the pattern deposited inside the microfluidic channel (Figure 4b,d) comprised ≈600 and ≈800 μm coated and non-coated alternating areas. The areas free from coatings were expected to delay (Figure 4b) or stop ( Figure 4d) the flow inside the microfluidic system, thereby acting as "delay" (400-600 μm) or "stop" (>800 μm) valves. Figure 4c shows the implementation of the "stop" and "delay" valves in the 3DP microfluidic channels, where the respective valve is indicated by the brackets.
In Figure 4b,d, the red arrows ("Additional printed margin" or "No margin") indicate areas introduced to simplify the inkjet printing process, e.g., to avoid misalignment issues. Movies S3 and S4 (Supporting Information) show the flow behavior of the liquids in these microfluidic channels.

Inkjet Printing of Hydrophilic Coatings to Enable Capillary-Driven Microfluidics with Automatic Release of Test Solutions
Controlled release of test solutions has previously been demonstrated with capillary-driven microfluidics. [1,3,5] Figure 5a shows a schematic illustration and Figure 5b shows a photograph of a functional microfluidic device. For the sequential release of reagents, the device encompasses a reaction chamber, a capillary pump containing tissue paper, and three reservoirs (≈20 μL each) to keep solutions with different dilution of the reagents. The inkjet printed coating lowers the water contact angle inside the microchannels and enables the capillary flow, while the deliberated retention pressures at the reservoirs upstream take care of the sequential release of fluids (colored dyes).
In the work reported herein, the bottom channel walls of various crucial parts (including additional printed margin areas, see Figure 5c,d) of the 3DP microfluidic chip were patterned with the inkjet printed hydrophilic coating to enable autonomous delivery of test solutions to the so-called reaction chamber, which is equipped with a sensing device. Because of the hydrophilic coating and the channel geometry, the automatic and sequential release of three fluids with colored dyes has been realized. The purple fluid was released first, as shown in Figure 5b (already in the reaction chamber), followed by the consecutive release of the red and the blue colored fluids, see also Movie S5 in the Supporting Information.  Figure S4 in the Supporting Information. For this, current versus time measurements were performed. [19] More specifically, the drain current (I D ) was recorded (the drain voltage, V D , was kept constant at either −0.2 or −0.3 V) while the different solutions, with subsequently increased concentration of Cl − anions, were sequentially released and driven to the sensor. The I D current was modulated due to the increasing concentration of Cl − anions brought in contact with the sensing element (the Ag/AgCl gate electrode), as shown in Figure 6d,e; the recorded response was in agreement with Nernst equation. [36] Once the automatic release was initiated, the OECT sensor was exposed to a 100 × 10 −3 m NaNO 3 buffer solution that served as the baseline for the following stepwise delivery of the 1 × 10 −3 , 10 × 10 −3 , and 100 × 10 −3 m NaCl solutions to the sensing chamber. The solution of each one of the NaCl concentrations was premixed into a 100 × 10 −3 m NaNO 3 electrolyte buffer solution. To verify the results obtained with the microfluidic device, a set of experiments based on dropcasting of the analyte solutions on the sensor were performed, see Figure S5 in the Supporting Information.

Sensing of Chloride Anions by Using an OECT-Based Sensor Integrated in Modified Capillary-Driven Microfluidics
A linear variation (as expected from the Nernst equation) in the sensor response was recorded (Figure 6e), for both experimental   configurations, for increasing concentrations of Cl − anions, with slopes of ≈14 μA × log(mm) −1 and ≈20 μA × log(mm) −1 , respectively, for the manual and the fluidic calibrations ( Figure S5c, Supporting Information). Both linear fittings had R 2 >0.99, indicating a good correlation between concentration and sensor responses. To further evaluate the correlation between the two different experimental approaches, the obtained results for the respective analyte concentration were plotted against each other in Figure S5d (Supporting Information). The slope (0.76) of the linear fitting for this plot is close to 1 (slope for perfect correlation), indicating an acceptable agreement between the results obtained using the two experimental approaches. It should be noted that discrepancies between the results might be associated by the different equilibration times between the two experimental approaches, deviations caused by dilution (especially for 1 × 10 −3 m NaCl) and bubble formation that might occur in the microfluidic assisted assay.

Conclusions
Over the past decades, drop-on-demand inkjet printing of functional materials has allowed lab-on-a-chip (LoC) technology and point-of-care (PoC) diagnostics to evolve from conceptually promising into low cost, disposable, portable, and miniaturized tools for medical diagnostics.
The advantages of inkjet printing technology feature fast and low-cost deposition of on-demand shaped coatings in targeted areas, to match complex microfluidic circuits, which in turn modifies the surface chemistry and enables capillary flow in 3DP microfluidics with pre-programmed release of analytes. Additionally, the coatings at the bottom of the channels of the microflu-idic circuits also allow for a novel approach of obtaining "stop" and "delay" valves. The inkjet printed coating inside the 3DP microfluidic channels, in contrast with plasma treatment, [10] retains its surface hydrophilicity for at least 6 months, which indicates its suitability in, e.g., disposable sensor systems. The modified microfluidic circuits were successfully integrated with fully screen printed OECTs to sense and verify the automatic release of multiple test solutions, by current versus time measurements, inside the 3D-printed microfluidic chip. The initial results confirm the functionality of both the coating in the microfluidic device and the fully printed OECT through the detection of Cl − anions in the (1-100) × 10 −3 m range in a (100 × 10 −3 ) m NaNO 3 buffer solution.
The results achieved herein pave the way toward the use of printing techniques as a promising tool to enhance and create on-demand surface hydrophilicity in 3D-printed microfluidics, including various geometric patterns that could delay, or even terminate, the fluid flow within the microfluidics. A combination of these tools (inkjet printing and 3D-printing) brings the cost-effectiveness and precision that could shape new microfluidic designs and architectures as well as all-printed sensing platform solutions for PoC and LoC, for the automatic delivery of multiple test solutions and the detection of their ionic species.

Experimental Section
Fabrication of Microfluidics: The microfluidic devices were designed with SolidWorks 2021 (Dassault Systèmes SE, France) and printed using the stereolithography (SLA) 3D printer Form 3 (Formlabs, USA), laser wavelength of 405 nm, and a 3DP layer thickness of 50 μm. A monocure clear resin (Clear V4 Resin, Formlabs, USA) was used for microfluidic chip fabrication. Postprocessing involved cleaning (rinsing with isopropanol, using Formwash, Formlabs, USA) to wash away uncured resin, drying (under a stream of pressurized nitrogen gas), curing (10 min under a UV lamp at 60°C Formcure, Formlabs, USA). The microchannels were 3DP with one open side to avoid trapping resin and reduce over-polymerization. The inkjet printing of the coating was performed into the open side of microchannels, reservoirs, etc.
Fabrication of OECTs: The OECT devices reported here comprise coplanar channels and gate electrodes screen printed on flexible polyethylene terephthalate substrates (PET, 125 μm thick, Polifoil acquired from Policrom Screen). The PET substrates were pretreated at 120°C for 5 min prior fabrication. The manufacturing process was performed by using a DEK Horizon 03iX screen printing equipment. The device fabrication includes 6 consecutive screen printing steps.
Step 1-printing of Ag-based (Ag 5000 from DuPont) interconnects; step 2-printing of PEDOT:PSS-based (Clevios S V3, screen printing paste from Hereaus) electrochemically active transistor channels; step 3 -printing of carbonbased (7102 from DuPont) source and drain electrodes; step 4-printing of PEDOT:PSS-based gate electrodes (Clevios S V3); step 5-printing of Ag/AgCl electrodes (CI-4025, purchased from Engineered Materials Systems); step 6-printing of insulator (5018 from DuPont). The steps 1-5 require thermal curing at 120°C for at least 5 minutes, while step 6 requires curing by UV light. The ink/material processing steps are crucial to ensure device functionality.
Sensing of Ions (Cl − ) with Screen Printed OECTs: In the fully screen printed OECT device presented here, the channel and the Ag/AgCl electrode (further employed as the gate electrode) was used for the monitoring of anion (Cl − ) concentration. The respective 1 × 10 −3 , 10 × 10 −3 , and 100 × 10 −3 m NaCl solution was prepared in a 100 × 10 −3 m NaNO 3 buffer solution, to ensure selectivity upon increasing the anion (Cl − ) concentration.
The sensing of anions (Cl − ) was performed inside the microfluidic system (Figure 6d), thereby including automatic release and delivery of test solutions to the channel and gate electrodes of the OECT.
Coating Method: The inkjet printing process was performed by using a Dimatix inkjet printer (DMP-2800) in ambient conditions. Dimatix SAMBA 10 and 2.4 pL cartridges (purchased from Fujifilm) were used. The PSS − Na + -based (15 wt%) hydrophilic coating ink was prepared in deionized water and stirred for 2 h. Further, a high boiling point solvent (polyethylene glycol) was added to achieve a 90:10 wt% ratio. Prior to the inkjet printing process inside the microfluidic channels, a 0.45 μm syringe filter (Acrodisc, PVDF membrane) was used to filter the polyelectrolytebased ink formulation before filling the cartridge. The printing settings were set to achieve a resolution of 1016 DPI, and several layers were subsequently inkjet printed. The 2.4 and 10 pL cartridges resulted in different coating thicknesses, due to different volumes of the ejected drops. The thicker coatings ensure better wetting properties because of the high surface roughness of the 3DP resin. For the storage test of 14 days, the coating was stored at ambient conditions (22°C and 47RH%), while the microfluidic devices that were stored for 6 months were kept in a sealed plastic bag.
Embedding 3D-Printed Microfluidics with OECTs: A laser cutter (Trotec Speedy 300) was used to cut the adhesive tape, according to the shape of the reaction chamber, to ensure a proper contact with the gate electrode and the channel of the OECT, see Figure S6 (Supporting Information). A double-sided hydrophobic adhesive film (purchased from Paul & Co) was employed to seal the modified microchannels and bond the 3DP microfluidic circuit with the OECT device ( Figure S6, Supporting Information).
The absorbent pad was attached to the outlet of the microfluidic chip, thereby acting as the capillary pump. Detailed information on the assembly and the stepwise additive manufacturing process of the all-printed sensor system is described in Figures S6 and S7 in the Supporting Information. The conceptual illustration of the 3DP microfluidic circuit, also including the dimensions of the different parts, e.g., detection site and main channel, is shown in Figure S8 (Supporting Information). The overall stepwise manufacturing process is further illustrated in Movie S6 (Supporting Information).
Electrical Characterization: All measurements of OECTs embedded into the modified microfluidics were performed under controlled conditions: ≈22°C and ≈50%RH. Electrical characterization (current vs time measurements) of the OECT devices was carried out using a semiconductor parameter analyzer (HP/Agilent 4155B) to provide voltage to the drain and gate electrodes and to record the drain current versus time. The drain voltage (V D ) was set to −0.2 or −0.3 V, while various gate voltages were applied (0.2 and 0.4 V) to optimize the drain current modulation.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.