PolyJet 3D-Printed Microchip Devices with Integrated Carbon Electrodes for Neurotransmitter Analysis

Abstract

We present an approach for integrating carbon ink electrodes into PolyJet 3D-printed microfluidic devices for electrochemical detection. Devices for both microchip electrophoresis and microchip-based flow injection analysis can be created with this methodology. The fabrication involves printing two separate components, a channel layer and an electrode layer, which are thermally bonded to form the final device. For the electrode layer, carbon electrodes are first patterned onto glass substrates using a micromolding technique. A custom stencil is printed directly onto the PolyJet tray to guide precise alignment; the electrode layer is then printed directly over the glass substrate, transferring and embedding the electrodes accurately within the 3D-printed structure. The channel layer is produced by 3D printing either onto a pre-fabricated mold featuring a T-intersection (for microchip electrophoresis) or onto the printer tray along with solid support (for microchip-based flow injection analysis). This method yields devices with reliable electrode-channel alignment and minimal band broadening. For microchip electrophoresis experiments, the device effectively separated a mixture of neurotransmitters with theoretical plate counts up to 136,000 plates/m and a limit of detection for dopamine of 170 nM. Additionally, we demonstrate how to use 3D printing to integrate off-chip processes such as microdialysis sampling with microchip electrophoresis and electrochemical detection. Lastly, we show how microchip flow-based injection analysis devices featuring single or dual in-channel carbon electrodes can also be produced with this approach.

Introduction

Microchip-based electrophoresis (ME) has made significant advancements over the past three decades. Early demonstrations of this technique highlighted the many benefits of the microchip format, including rapid analysis times, the ability to apply high electric field strength and minimal solvent consumption [1–3]. The compact design of the chip, combined with its capability to integrate multiple processes into a single device, makes this approach highly appealing for a variety of applications [4, 5]. The small sample volumes (nL-to-pL) introduced into the separation channel, combined with the micron-sized separation bands, require highly sensitive detection techniques [6–8]. Initially, most studies employed laser-induced fluorescence (LIF) detection due to the simplicity of system construction, ease of focusing the laser beam within the channels, and the ability to achieve low limits of detection (LOD) [3]. However, a significant drawback of laser-induced fluorescence (LIF) detection in microchip electrophoresis is the need for analyte derivatization with a fluorophore, which adds complexity and time. Mass spectrometry detection offers extensive analytical data but is costly and is an off-chip detection technique [9]. Electrochemical (EC) detection is also a popular form of detection with ME, as it provides excellent sensitivity without requiring derivatization, and a good number of biomolecules including neurotransmitters are inherently electrochemically active. Typical LODs are in the mid-to-low nanomolar range [10]. Additionally, the technique is relatively simple and inexpensive, allowing miniaturization without compromising sensitivity [11].

A variety of substrates have been used to fabricate ME devices including glass, PDMS, PMMA, and paper [12, 13]. Glass remains popular choice for many applications due to similarities to fused silica used in traditional capillary electrophoresis, strong electroosmotic flow (EOF), minimal analyte adsorption, high separation efficiency and excellent optical transparency. PDMS is another popular material due to its low cost of production, simple fabrication, and the ease of integrating electrodes into the design. However, PDMS devices come with drawbacks such as weaker EOF, analyte adsorption, and the batch nature of making devices with soft lithography [14, 15]. Recently, additive manufacturing (commonly known as 3D printing) has emerged as a viable method for fabricating ME devices, with the time from initial design to final product achievable in just a few hours per iteration. Over the years, various 3D printing technologies have been employed to fabricate microfluidic devices for a wide range of applications [16–20]. Significant challenges remain in 3D-printing devices with narrow channels necessary for efficient electrophoretic separations, as most commercially available 3D printing tools are limited in their ability to print channels smaller than 100 μm. For ME applications, the Nordin and Woolley research groups have advanced Digital Light Processing (DLP) technology to produce microfluidic devices with channels as small as 18 × 20 μm using a custom DLP 3D printer [21–23]. One of the devices was fabricated with channel dimensions of 37 × 49 μm and was utilized for the electrophoretic separation of preterm biomarkers [24]. Previous work in our laboratory focused on using micromolds in combination with PolyJet 3D printing to fabricate microchip electrophoresis devices with T-intersections, achieving channel cross-sections as small as 42 x 12 μm [25]. We further advanced this work by developing a novel method to integrate metal electrodes (such as Au or Pt) of various sizes into microchip devices for amperometric detection [26].

It is known that, as compared to metal-based electrodes, carbon electrodes are desirable for neurotransmitter analysis. This is due to their large electrochemical potential window and being less prone to fouling [27, 28]. In addition, adsorption of catecholamine neurotransmitters such as dopamine onto carbon electrodes can lead to increased oxidation currents and improved sensitivity in electrochemical measurements [29, 30]. Various types of carbon electrodes, including carbon fiber, carbon paste, carbon ink, and pyrolyzed photoresist films have been integrated into PDMS substrates for ME-EC-based detection of catecholamines and related compounds [27, 28, 31]. In this paper, we describe how PolyJet-based 3D printing can be used to integrate carbon ink electrodes in microfluidic devices for the separation and electrochemical detection of catecholamine neurotransmitters. We show that these electrodes can be integrated with an end-channel configuration (for monitoring ME-based separations) or in-channel (for microchip-based flow injection analysis with either single or dual electrodes). In addition, we demonstrate how microdialysis sampling, a popular technique for monitoring in vivo and in vitro systems, can be integrated with ME-EC devices containing carbon ink electrodes. This approach should expand the use of 3D printing to create robust electroanalytical microfluidic devices for separations and fluidic applications.

Material and methods

Chemicals and materials

The following chemicals and materials were used as received: carbon ink (E3178) and solvent thinner (N-160) were purchased from Ercon Inc. (Wareham, MA, USA). VeroClear-RGD810 print material and SUP706B support were supplied by Stratasys Ltd. (Eden Prairie, MN, USA). Isopropyl alcohol and plain large microscope glass slides were acquired from Thermo Fisher Scientific (Waltham, MA, USA), Dopamine, pyrocatechol, tyrosine, homovanillic acid, 3.4-dihydroxy-phenylacetic acid, phosphate-buffered saline (PBS), and 2-[Tris-(hydroxymethyl) methylamino]-1-ethanesulfonic acid (TES sodium salt) were purchased from Millipore-Sigma (St. Louis, MO, USA). J-B weld epoxy and adhesive conductive copper tape were purchased from Amazon. Capillary tubing 150 μm ID was acquired from Polymicro Technologies (Phoenix, AZ, US).

Fabrication of Carbon ink electrodes

PDMS channels were used to pattern carbon ink electrodes on glass before 3D printing was started. Use of soft-lithography to produce PDMS molds is a well-established method and was carried out according to a protocol described elsewhere [32, 33]. The PDMS micromolding channel used in this study was created by mixing PDMS monomer and curing agent in a 10:1 ratio, then pouring the polymer mixture onto a silicon mold. The polymer mixture was cured at 60°C for 4 hours before peeling off from the silicon mold. The final PDMS micromolding channel was 32 mm long with dimensions of 60 μm x 70 μm (width x depth). Reservoir holes were created at each end of the channel using a 4 mm biopsy punch.

The steps involved in fabrication of carbon working electrode are depicted in Figure 1(A–C). The initial formation of the carbon ink electrode on a glass surface is based upon previously published methodologies [28, 34]. First, 0.2% (v/w) carbon ink solution was prepared by weighing 0.5 g of carbon ink and diluting it with 1 mL of solvent thinner. The PDMS mold was reversibly sealed on a clean glass slide. The enclosed micro-channel was filled with carbon ink by pipetting ink into one reservoir while using a vacuum pump to draw from the other reservoir. Following a drying process in the oven at 70°C for 4 hours, the PDMS was peeled off carefully leaving a solid dry carbon ink microstructure on the glass slide.

Figure 1.

Steps for fabricating and integrating carbon ink electrodes into a 3D printed microchip electrophoresis device (with a PolyJet printer). (A) A negative relief PDMS straight channel 60 μm x 70 μm cross section, 32 mm in length) is reversibly sealed to glass substrate and the channel is filled with carbon ink. (B) The ink is allowed to dry in an oven for 4 hours at 70°C. (C) The PDMS is delaminated from the glass substrate, leaving a carbon ink electrode. (D) A designed stencil is printed (2-layer thickness) directly on the tray to ensure alignment and positioning of the electrode for each print. The carbon electrode (on the glass substrate) is placed over the stencil with double-sided tape (location as defined by the stencil). (E) A CAD design model for the electrode layer is then printed on top of the glass piece. (F) Delamination of 3D-printed electrode layer from the glass piece, with the carbon electrode being transferred and embedded in the 3D printed layer. (G) An electrophoresis channel layer is made from a brass mold (also with 3D printing), and the two layers are bonded together with a thermal lab press to result in the final device (H)

PolyJet printing of Fluidic Devices with Integrated Carbon Electrodes

The process of integrating carbon ink electrodes in PolyJet 3D printing begins with the design of a 0.054 mm thick stencil, created using SolidWorks 2024 (Dassault Systèmes, Waltham, MA, USA) and printing directly onto the tray. This stencil design was of the same design of the eventual electrode layer, including reservoirs. The edge of the reservoir on the printed stencil served as a marker for placing the glass slide with the carbon ink electrode onto the tray, ensuring consistent alignment and positioning of the electrodes at the end of the separation channel reservoir for each print. The glass slide with carbon ink electrode was placed over the stencil using double-sided tape, ensuring the electrode was positioned according to the stencil design at the edge of the reservoir (see Figure 1D). Next, a CAD model for the electrode layer with reservoirs was printed on top of the glass slide and the electrode. After printing, the 3D-printed electrode layer was carefully removed from the glass slide by inserting a sharp blade between the printed layer and the glass around the edges. The printing process led to the carbon being transferred from the glass slide (with no residual carbon remaining) and embedded into the 3D-printed electrode layer, as shown in Figure 1F. A separate CAD design model was then printed over a T-positive relief microstructure (a brass mold fabricated using high-precision micro-milling at the Advanced Manufacturing Facility, Louisiana State University) to create microchannels [26]. After printing, the layer was delaminated from the brass mold and the microchannels were determined to ~ 40 x 40 μm (width x depth). Prior to bonding, the two layers were rinsed with isopropyl alcohol and placed in a UV plasma chamber (Harrick Plasma, Ithaca, NY, USA) for 90 seconds [35]. Copper tape was then used to establish connections between the carbon ink electrode and external components. To enhance conductivity at the interface between the carbon and the copper tape, the adhesive side of the tape was first cleaned with acetone, followed by the application of colloidal silver. The prepared tape was then positioned over the carbon electrode on the 3D-printed layer, ensuring both ends of the electrode were covered (Figure 1G). The final device (Figure 1H) was assembled by bonding the two layers using a thermal lab press (DABPRESSTM, Guangdong, China) under these optimized parameters: 54 °C and 350 psi applied for 6 minutes to ensure reliable sealing and structural integrity. Electrode alignment was examined using a stereomicroscope (Olympus SZ61) with a 10X objective (Microscope Central, Feasterville, PA, USA), and images were captured with an iPhone 15 Pro Max (camera specifications: 12MP ultra-wide lens with an f/2.4 aperture).

ME-EC experimental conditions

Prior to use, the microchip device was treated with a corona discharge unit fitted with a fine tip electrode (model ETP BD-20, Electro-technic Products, Inc., Chicago, IL) to increase EOF, as previously described [36]. All electrophoresis experiments were performed using a LabSmith HVS448 3000 V High Voltage Sequencer (LabSmith, Livermore, CA, US), which features eight independent high-voltage (HV) channels. TES buffer (25 mM, pH 7.4) was prepared with deionized water (18.0 MΩ-cm). Standard stock solutions of the catecholamine neurotransmitters (10 mM) were freshly prepared in the TES buffer each day and diluted to the required concentrations using the same buffer. A gated injection method was employed for sample introduction and separation in the simple T-device. The gate was created by applying +600 V to the sample reservoir, +640 V to the buffer reservoir, and grounding both the sample waste and buffer waste. Injections were carried out by setting the buffer voltage to 0 V for 2 seconds, then reapplying the separation voltages. The injected sample plug migrated down the 2.1 cm long separation channel under a field strength of 260 V/cm toward the carbon working electrode.

For ME experiments, the working electrode was placed at the channel outlet (Figure 2). This end-channel alignment has been employed previously in our lab with metal electrodes and yields minimal interference between separation and detection voltages [26]. The electrochemical detection was carried out using a two-electrode system consisting of carbon working set at +0.9 V and a reference electrode that was placed in the outlet reservoir. Note that a Ag/AgCl reference electrode (BASi, West Lafayette, IN, USA) was employed for the hydrodynamic voltammetry (HDV) experiments, whereas a Pt pseudo-reference electrode was used for all other electrochemical measurements described in this study. A model 9000-k7 handheld wireless electrically isolated potentiostat (Pinnacle Technology, lnc., Lawrence, KS, USA) was used to record the data. The device had a sampling rate of 13 Hz and data acquisition was performed through wireless transmission and visualized with Pinnacle Acquisition Laboratory (PAL 8400) software. The hydrodynamic voltammogram of 500 μM dopamine in 25 mM TES buffer (pH 7.4) was recorded in the presence of high voltage (field strength of 260 V/cm) as well as in the absence voltage. For studies without any electrophoresis voltage (for HDV comparison), fresh sample was introduced to the channel (via vacuum) and the current measured, with fresh sample being introduced between each potential change (0 V to +1.2 V).

Figure 2.

Simple-T microchip electrophoresis device integrated with carbon electrode for end-channel electrochemical detection. (A) Schematic of the device with applied voltages for gated injection as well as channel dimensions. (B) Image of the assembled device. Insert is a micrograph of the channel/electrode interface, with a 50 μm wide carbon electrode aligned at the end of the separation channel.

Fabrication of Device for Integrating Microdialysis Sampling with ME-EC

Two separate layers (channel and electrode) were designed using Autodesk Inventor Professional 2024 and 3D printed on a Stratasys J735 PolyJet printer with VeroClear material. The electrode layer featured four reservoirs (4.5 mm diameter) and a 0.415 mm hole to accommodate a 0.360 mm OD capillary tube. The electrode layer was printed following the method described in Figure 1. The channel layer was designed with a flow channel to enable continuous sample introduction from a microdialysis probe. This channel (which measured 145 x 90 μm) was printed on top of a brass mold with a positive relief structure (40 x 40 μm) to create electrophoretic channels. During printing, only the flow channel (not the electrophoresis channel) was printed with support material, which was later mechanically removed. A laser profiler (Keyence VK-9710K, Itasca, IL) was used to measure the dimensions of the flow channel after cleaning off the support material. After plasma treatment and placement of the copper tape connections, the electrode and channel layers were bonded together using a thermal lab press for 6 minutes at 54 °C and 350 psi of pressure. Following bonding, a 40 mm long (150 μm i.d.) fused silica capillary (360 μm o.d.) was inserted into the 0.415 mm hole, and J-B Weld epoxy was applied around the capillary to secure it to the device. The device was then left overnight to allow the epoxy to fully cure.

Microdialysis Sampling Experiments

A CMA 20 microdialysis (MD) probe with 4 mm membrane (Harvard Bioscience, Charlotte, NC) was inserted in sampling vial with the perfusate tubing connected to a syringe pump and the dialysate tubing connected to the capillary tube on inlet of the continuous flow/microchip device. The syringe pump operated at a flow rate of 2.0 μL/min to pump the buffer through the MD probe to the ME device. A high voltage (+500 V) was applied to the buffer reservoir, and ground at the buffer and sample waste reservoirs. A flow boundary was established between the sample and the buffer at junction [37]. A gated/flow injection was achieved when the high voltage on the buffer reservoir was switched to 0 V for 5 sec. The sample plug migrated down the separation channels (2.1 cm long) at a field strength of 190 V/cm. Amperometric detection was achieved in a similar manner to the gated injection device. The response time of the microdialysis/microchip electrophoresis system to changes in dopamine concentration was evaluated by initially injecting 200 μM dopamine in TES buffer (pH 7.4), then increasing the concentration to 460 μM. Injections were performed continuously until the peak heights stabilized, indicating that a steady-state concentration had been reached. The experiments were conducted at a MD probe flow rate of 2.0 μL/min and a 5 sec gated injection time.

PolyJet Printing of Microchip-based Flow Injection Analysis Devices with Carbon Ink Electrodes

Two separate layers, a straight flow channel layer and an electrode layer, were designed in SolidWorks 2024. The electrode layer was designed with a 4.5 mm diameter reservoir and a 0.415 mm hole to accommodate a fused silica capillary tube that can be connected to a syringe pump. The electrode layer was printed using the same method as the electrode layer in the microchip electrophoresis device but incorporated either a single or dual electrode within the channel. A stencil was used to ensure consistent electrode alignment across prints. The channel layer included a straight channel with a cross-section of 400 μm × 240 μm (width x height) and a length of 30 mm, and it was printed directly onto the build tray without molds. Solid support material was used during printing and was later removed mechanically. Both layers were treated with plasma, after which copper tape was applied to each end of the carbon electrode(s) before the bonding process. Finally, the layers were thermally bonded using the same conditions applied to the electrophoresis device. Following bonding, a fused silica capillary with a 360 μm outer diameter (150 μm id) was inserted into the small hole, and JB Weld epoxy was used to secure it in place. The epoxy was left to set and fully cure overnight.

Microchip-based Flow Injection Analysis Experimental Conditions

A syringe pump was used to pump 1X PBS (flow rate = 15 μL/min) through a four-port injector valve (1 μL rotor) to the 3D-printed flow device via the fused silica capillary tube. A sample plug was introduced into the flow system using a 4-port injection valve. The valve was initially set to the ‘load’ position, where the sample was introduced into the sample loop via a syringe. Once the loop was filled, the valve was switched to the ‘inject’ position, allowing the 1X PBS solution to flow the sample plug into the flow device channel for analysis. A stock solution of 10 mM catechol in 1x PBS was prepared daily and necessary dilutions with the same buffer were performed prior to use. Electrochemical detection was achieved using a CH Instrument potentiostat (Austin, TX, USA) in a 3-electrode format for the single electrode flow device. For dual electrode experiments, a similar setup was used, with the leading electrode being held at oxidative potential and the trailing electrode being held at a reductive potential. Both the single and dual carbon electrode(s) served as the working electrode(s), while an external platinum wire served as the auxiliary electrode and an external Ag/AgCl electrode (Bioanalytical Systems) functioned as the reference. The auxiliary and reference electrodes were placed in the buffer waste reservoir. Oxidation potentials were set at +0.90 V and the reduction potential in the dual electrode system was set at −0.40 V. Leakage studies were completed using fluorescein and a fluorescence microscope (IX71, Olympus America) equipped with 100 W Hg arc lamp, fluorescein filters, and a cooled 12-bit monochrome Qicam Fast digital CCD camera (QImaging, British Columbia, Canada). Images were captured with Streampix Digital Video Recording software (Norpix, British Columbia, Canada). Bright field micrographs showing the single and dual electrodes in-channel were obtained using Keyence VHX-500K digital microscope (Japan), and measurements were done with VHX Measurement Software.

Results and Discussion

We recently described a method for integrating microwire (Au or Pt) electrodes into 3D-printed microchip electrophoresis devices for electrochemical detection [26]. In this paper, we describe how PolyJet 3D printing can be used to produce microfluidic devices with carbon ink electrodes. This platform starts with producing a carbon ink electrode on a glass surface using a micromolding procedure [34]. In this approach, PDMS microchannels are used to define the geometry of the microelectrodes. The steps used to fabricate carbon microelectrodes are depicted in Figure 1A–C. First, a PDMS slab containing microchannels measuring 60 μm × 70 μm (width × depth) sealed onto a glass substrate and filled with carbon ink. After a controlled drying process in an oven for 4 hours, the PDMS mold was carefully removed, leaving behind carbon microelectrodes with final dimensions of approximately 50 × 55 μm. The slight reduction in size can be attributed to shrinkage during evaporation of the volatile thinner and other ink solvents. Figure 1D shows how these electrodes are integrated during the PolyJet printing process. Reproducible electrode alignment across prints was achieved using a stencil designed with the same lateral dimensions as the electrode layer, differing only in thickness. This stencil was printed directly onto the printing tray, and pre-fabricated carbon electrodes on glass substrates were secured onto it using double-sided tape, ensuring precise positioning at the end of the reservoir according to the stencil’s layout. Following PolyJet printing directly onto the carbon ink patterned on the glass substrate, delamination of the 3D-printed layer resulted in the carbon electrode being transferred and embedded within the printed material as shown in Figure 1F. Another CAD design (channel layer) was printed on a micromachined brass mold with a positive relief T-structure to create 40 x 40 μm microchannels, as previously described [25]. Copper tape was used to establish electrical connections between the carbon electrodes and external components. After the layers were bonded, either end of the tape could be used to connect to the potentiostat. The overall ME device dimensions are given in Figure 2A. The bonded devices were durable, maintaining their bond for up to 90 days of storage (longer durations were not tested). The microfluidic devices used in this study exhibit consistent performance across multiple experiments and operated reliably with continued use over a period of several weeks.

ME with EC detection

As can be seen in Figure 2, for ME devices the electrode placement was at the end of the separation channel (end-channel alignment), with no measurable gap between the separation channel and the electrode. To gain further insight into the channel/electrode interface, fluorescence microcopy was employed, along with a gated injection method to introduce a fluorescein plug into the device. A gated injection (see Figure 2A) was accomplished by applying +600 V to the sample reservoir and +640 V to the buffer reservoir, with both the sample waste and buffer waste grounded. To initiate injection, the buffer reservoir voltage was temporarily dropped to 0 V for 2 seconds, then restored to +640 V. This triggered the injection of a 0.20 nL plug of 1 mM fluorescein into the 2.1 cm long separation channel, under an electric field strength of 260 V/cm. Figure S1A shows the fluorescein plug as it passes the electrode (outlined with dotted white lines), exiting the separation channel and diffusing into the buffer waste reservoir. This configuration confirms that the electrode is accurately aligned with the end of the channel, resulting in minimal band broadening at the interface. Due to lack of a gap between the separation channel and the electrode, a shift in the half-wave potential can occur due to the applied microchip electrophoresis (ME) separation voltages [26, 38]. To evaluate this possibility and identify an appropriate detection potential, a hydrodynamic voltammetry experiment was conducted. Figure 3A presents the hydrodynamic voltammograms (HDVs) for dopamine under two conditions: one without any applied separation voltage (with fresh sample continuously flushed through the channel), and the other with the separation voltage applied using the gated injection conditions described above . A slight positive shift in the half-wave potential (~100 mV) was observed when the separation voltage was present, as has been seen previously with end-channel alignments [38].

Figure 3.

(A) Hydrodynamic voltammogram (HDV) for dopamine with separation voltage off (•) and separation voltage on (▪). Analyte concentration: 500 μM in 25 mM TES buffer (pH 7.4); field strength = 260 V/cm. (B) Sequential injections of 500 μM Dopamine (25 mM TES buffer, pH = 7.4) with the injection (inj.) sequence being denoted in the electropherogram (n=5). (C) Electropherogram for the separation and detection of various neurotransmitters using the carbon ink working electrode. Dopamine (DA, 100 μM), Tyrosine (Tyr, 500 μM), 3,4-dihydroxphenylacetic acid (Dopac, 100 μM), and Homovanillic acid (HVA, 400 μM). Gated injection time = 2 sec (inj. sequence denoted on each), field strength = 260 V/cm and separation buffer = 25 mM TES (pH = 7.4)

To further evaluate the device performance, sequential gated injections with a 500 μM dopamine solution (in TES buffer, pH 7.4) were performed. The data shown in Figure 3B shows good reproducibility, with an average peak height RSD of 4.6% (n=5). The average theoretical plates for these separations was calculated to be 127,000 plates/m. To evaluate the sensitivity of the microchip device, a calibration curve was generated using dopamine concentrations ranging from 5 μM to 100 μM. The resulting curve showed linearity, with an r² value of 0.996 (see Figure S1B). The limit of detection (LOD) for dopamine was determined to be approximately 170 nM. Our previous report using gold electrodes of the same size resulted in a LOD for catechol of 250 nM [26]. The separation efficiency and LOD are comparable to previously reported values for PDMS/glass with similar electrode alignments [33, 39]. Figure 3C demonstrates how these devices with gated injection can be used to separate catecholamine neurotransmitters, specifically a mixture of dopamine (DA), tyrosine (Tyr), 3,4-dihydroxyphenylacetic acid (DOPAC), and homovanillic acid (HVA). As can be seen in the electropherogram, an efficient separation with well-resolved peaks was obtained, with the average number of theoretical plate ranging between 49,000 (for DA) and 136,000 plates/m (for Tyr). The separation performance is comparable to previously reported values for similar analytes in PDMS-based microchip electrophoresis platforms with carbon electrodes [40]. Peak symmetry was assessed using peak skew, which is calculated from the ratio of the back peak width to the front peak width at 10% of the peak height. The average peak skew for the most abundant peak (Tyr) was 1.8, suggesting some minor peak tailing, as has been seen in end-channel alignments for ME [41].

Integrating MD-sampling with ME-EC

The use of a flow-gated approach to integrate microdialysis (MD) sampling with microchip electrophoresis (ME) was first demonstrated in an all-glass device [37]. The flow-gated interface enables the injection of discrete sample plugs from the continuously flowing MD perfusate into the ME separation portion of the device. Later work used a similar flow gated approach in PDMS/glass devices with integrated pyrolyzed carbon electrodes for both in vitro and in vivo monitoring of dopamine [42]. Previous work from our laboratory demonstrated that 3D printing can be used to create similar devices, allowing integration of continuous flow with on-chip gated injections [26]. Here we adapted that previous design to interface MD sampling, which typically operates at flow rates of 1–2 μL/min, with ME and EC detection. To achieve this, an electrode layer was designed with four reservoirs and a small pinhole (0.415 mm, for inserting a capillary) and printed over the carbon ink electrode as described for the gated microchip electrophoresis device. The channel layer was designed with an additional flow channel to interface the continuous flow from the MD probe with the ME device. This flow channel was printed with support material on the brass mold that created channels for ME. Only this flow channel has support material, which was mechanically removed after printing. Figure 4A shows a schematic of the final flow/microchip electrophoresis device, including the relevant dimensions. The final, printed flow channel was measured to be 145 × 90 μm. Figure 4B shows the integrated system designed to perform MD sampling coupled with analysis by ME separations and EC detection. To demonstrate the MD/ME/EC integration, a CMA 20 microdialysis (MD) probe with 4 mm membrane was used to sample a mixture of dopamine and tyrosine standards from a sample vial. The perfusate and electrophoresis buffers were matched (25 mM TES, pH 7.4). The perfusate buffer was pumped with the syringe pump at a flow rate of 2.0 μL/min. The analytes in the sample vial diffused across the microdialysis membrane and were carried out as dialysate, which was then pumped directly into the inlet of the flow channel. Most of the dialysate was directed to the outlet reservoir, while a fraction was diverted to the T-junction. Discrete plugs of the sample were introduced into the separation channel using a gated/flow injection scheme [26]. The analytes were separated within the channel and detected using a carbon ink electrode (see Figure 5A).

Figure 4.

(A) Schematic of microdialysis/microchip electrophoresis device with relevant dimensions. Samples from the MD probe were continuously pumped through the inlet into the flow channel (145 x 90 μm). The majority of the sample was directed to the outlet reservoir and a fraction was directed to the electrophoresis channels, where flow/gated injection was used to introduce a plug of sample into the separation channel. (B) Schematic of the microdialysis/microchip electrophoresis experimental setup. Perfusate (25 mM TES buffer, pH 7.4) was pumped through a 4-mm CMA microdialysis probe using a syringe pump. The probe was inserted through a small hole on the screw cap of a glass sample vial. Samples were continuously collected using the probe and directly introduced into the microchip device.

Figure 5.

(A) Separation and detection of dopamine (200 μM) and tyrosine (400 μM), sampled through the probe. Experimental conditions; separation field strength = 190 V/cm, 25 mM TES buffer (pH 7.4), syringe pump flow rate of 2.0 μL/min. (B) A step change experiment using dopamine as the analyte in TES buffer (pH 7.4) was conducted. The concentration of dopamine was changed from 200 μM to 460 μM; the red box represents the time at which the concentration increase was initiated. The microdialysis flow rate used was 2 μL/min, separation field strength = 190 V/cm and the flow/gated injection time was 5 sec.

In order to evaluate how the device responds to external changes in concentration, a concentration change experiment was conducted . This evaluation is particularly important for off-chip processes such as microdialysis sampling, where high temporal resolution is essential. It is typical to assess the rise time, which is defined as the time required for the signal to increase from 10% to 90% of its maximum intensity [43, 44]. Figure 5B shows the results of the concentration change experiment. Initially, the MD probe was placed in a sample vial containing 200 μM dopamine in TES buffer. The system was allowed to equilibrate for approximately 30 minutes prior to any sample injections. Following equilibration, several flow/gated injections of dopamine were performed and detected electrochemically using the carbon ink electrode. This process was repeated until steady-state dopamine peaks were observed. The dopamine concentration that the probe experiences was changed to 460 μM at a time denoted with the red box in Figure 5B. The resulting increase in peak height was monitored until the signal reached a stable plateau. Separations before the concentration change was initiated resulted in an average peak height of 6.01 ± 0.59 nA. For the region after the concentration change, where the signal reached a stable plateau, the average peak height was found to be 18.51 ± 0.78 nA. As can be seen in Figure 5B, the concentration change was noticed with an increase in signal after 1 injection/separation event. The rise time was 200 s, with the change being complete after four injection/separation events. These results are similar to previously reported PDMS-based approaches for integrating MD sampling with ME-EC [45].

Microchip-based Flow Injection Analysis with EC detection

Microchip-based flow injection analysis is also used for neurotransmitter analysis including systems involving on-chip cell culture [46, 47]. Here we show that single and dual (to monitor reversible redox couples) carbon ink electrodes can be integrated into microfluidic channels with an in-channel alignment. Since these devices rely on hydrodynamic flow rather than electroosmotic flow, larger flow channels were utilized (145 × 90 μm). The channel layers were printed using solid support material, which was mechanically removed after printing. The electrode layers were designed with a single (6 mm dia) reservoir and a 0.415 mm pinhole to accommodate a fused capillary. These layers were fabricated similarly to those used in the gated electrophoresis device, but with the electrode(s) positioned in-channel a few millimeters upstream from the buffer waste reservoir. Figure S2A shows a fully assembled device featuring dual electrodes integrated within the channel. A bright field micrograph of a single electrode in channel is shown in Figure 6A. To verify a complete seal at the electrode/channel interface, 10 μM of fluorescein solution was continuously pumped through the channel and fluorescence microscopy was used to take images at the channel electrode interface. The micrograph in Figure 6B shows no evidence of leakage at the interface. The device was connected to a syringe pump through a 4-port injector for continuous flow. The electrode sensitivity was determined by injecting (via a 1 μL, 4-port valve) catechol solutions that differ in concentration between 5 μM – 250 μM. A representative amperogram in shown in Figure 6C and the calibration curve showed a good linear correlation (r² = 0.999, see Figure S2B). The limit of detection was determined to be 150 nM for catechol.

Figure 6.

3D printed microchip device for flow injection analysis with electrode integrated in-channel. (A) Micrograph of the channel/electrode interface, with a 50 μm carbon electrode aligned in-channel, near the reservoir. B) Fluorescence micrograph showing a 10 μM solution of fluorescein being pumped through the channel (no leakage seen at the interface). (C) Amperogram showing discrete injection of catechol (1 μL injection volume from off-chip 4-port injector), with the peaks corresponding to different concentrations.

The use of dual-electrode detection in series format (as shown in Figure S3) can further enhance selectivity in EC detection by enabling the monitoring of reversible redox couples (as not all species have reservable electrochemistry) [48]. The micromolding/printing method was used to create dual 50 μm carbon electrodes with a spacing of 470 μm (see Figure S3A), with no leakage being seen at the channel/electrode interface (Figure S3B). The flow device was used to monitor the quasi-reversible redox reaction of catechol (see Figure S4). The first electrode was set at +0.90 V, and an anodic peak was obtained for the oxidation of catechol to ortho-benzoquinone. The second electrode was set at −0.40 V, and a cathodic peak was obtained due to the reduction of ortho-benzo-quinone back to catechol. While the performance could be improved by further minimizing the spacing between the electrodes, this still shows that the 3D printed approach is viable for producing devices with multiple working electrodes integrated into a printed channel.

Conclusions

In summary, this work demonstrates the versatility and effectiveness of using 3D printing to integrate carbon ink electrodes into microfluidic platforms for both electrophoretic and flow-based electrochemical detection. The use of micromolding for electrode fabrication combined with PolyJet printing enables precise electrode placement with reproducible device performance. The devices were fully characterized in terms of sensitivity, limit of detection, and separation efficiency, with results being comparable to traditional PDMS-based approaches, with the added benefit of rapid prototyping and design flexibility. We also showed successful integration with microdialysis sampling and as well as dual-electrode flow injection analysis. Overall, this approach provides a customizable pathway for developing advanced microfluidic devices for electrochemical sensing. Future work will focus on expanding electrode surface modification techniques to target a broader range of biomolecules, as well as further exploring the integration of microdialysis sampling with cell culture systems for near real-time monitoring of neurotransmitters.

Data availability:

The data supporting this study’s findings are available from the corresponding author, R.S.M, upon reasonable request.