Commercial 3D printers are beginning to challenge the resolutions achievable by soft lithography. Inkjet 3D printing is based on inkjet technology and can operate in continuous or drop-on-demand mode. To achieve high accuracy performance in i3DP, there are four critical elements that must be considered: ink material, substrate properties, printing platform, and droplet generation. Inconsistency or change in any of the parameters affects the process reliability. To maintain steady performance, i3DP needs to be used regularly; the clogging of apertures in a print head due to drying of the ink is a persistent problem. There is considerable cost involved in changing from one material to another, since the previous material must be flushed out by the new resin. One of the greatest advantages of i3DP is its ability to deliver multiple materials at the same time during the fabrication process, allowing for a wide range of material properties , and the utilization of inks of different color. Creating prototypes with smooth finishes and complex shapes is possible with i3DP; flow channels have been integrated with porous semi-permeable membranes supporting cell culture to study the transport and profiling of drugs. Hwang et al. utilized i3DP to print pillars with a diameter of 250 µm and found that the resolution of the process depended on the droplet size, the printer nozzle spacing, drying weed and the reflow of the material prior to UV curing. These factors affected the droplet spreading, changing the final dimensions of the printed devices.
Paydar et al. examined multi-material 3D printing for microfluidic interconnects. They fabricated a specialized interconnector part, composed of two rigid clamps for the mechanical attachment of a flexible elastomeric O-ring gasket to a microfluidic device. The parts were printed in a single step, eliminating the need for adhesives or additional assembly. While the cost of manufacturing was low, the interconnector was characterized by low maximum sealing pressure due to material fatigue.Two-photon polymerization is a laser-based technique that utilizes a femtosecond laser to create 3D structures in the bulk of photo curable epoxy resin, employing a highly localized process where two photons are absorbed simultaneously by the molecule being cured. The synergistic effects of optical, chemical, and material non-linearity make it possible to achieve reproducible resolution of tens of nanometers. 2PP has shown great potential for fabrication in microfluidics. Kumi et al. described the fabrication of a master for casting PDMS with rectangular microchannels of high aspect ratios by modifying SU-8 resist with a photoacid generator that then allows for the use of 2PP on SU-8 resist. By using the modified resin, the fabrication speed was also increased from 200 µm·s −1 to 10,000 µm·s −1 with a print time of 1 h. This technique required extensive resin preparation and had a slow build speed. Kawata et al. achieved resolutions of 120 nm, and other attempts included the use of new photo-initiators, a continuous scanning mode, a shorter wavelength, a longer exposure time, and confining the polymerization phenomenon using a quencher molecule. Sugioka et al. conducted a comprehensive review of the fundamentals and fabrication of 3D micro- and nano-components based on 2PP .
While 2PP currently produces the highest resolution for microfabricated three-dimensional structures, it is a very time-consuming process: the time required to fabricate a 1 mm3 volume microfluidic structure exceeds 104 days. The high cost of femtosecond lasers, positioning systems, optics, and the difficulty of working with multi-material systems are some factors hindering the utilization of 2PP for the mass-production of microfluidic devices. FDM is one of the most widely used additive manufacturing techniques. Many polymers used with FDM techniques are inherently biocompatible. In FDM, filament material is extruded through nozzles and deposited onto a heated substrate. Due to the inherent propensity of melted fiber to solidify as a line, there are limitations in the dimensional accuracy and the surface texture of the produced parts, resulting in a staircase pattern that is ubiquitous on many parts that are produced with FDM. Lee et al. printed microfluidic features via FDM and evaluated the printing resolution, accuracy, biocompatibility, and surface roughness of acrylonitrile butadiene styrene with P430 filament. They found that the accuracy of the printed features had an average deviation of 60.8 µm and 71.5 µm along the Y and X axis, respectively. The surface of the printed channels was rough with protruding filament strands. Kitson et al. fabricated polypropylene reaction ware with cylindrical channels 0.8 mm in diameter. The devices could be fabricated in a few hours and could avoid blockages due to the formation of precipitates. The potential of 3D manufacturing was demonstrated by stopping mid-point in the fabrication process to deposit solid reagents into a chamber, which was then sealed with the printer. This is a valuable feature not so easily realized with i3DP or stereolithography . Bishop et al. created a semi-transparent fluidic device using poly with threaded ports, enabling the integration of commercial tubing as well as specially designed 3D fittings. The transparent device included 800 µm × 800 µm square channels. A low-cost desktop Makerbot 3D printer was used for fabrication of the device. Prussian blue nanoparticles were synthesized in their lab and mixed in a 3D printed channel and applied to electrode surfaces for sensing of H2O2. Recently, Dolomite launched a production line that offered an FDM printer specifically designed for fabrication of microfluidic platforms.
The printer head design allowed for reliable printing of cyclic olefin copolymer . Their new software guided the production to guarantee smooth surface finish inside the channels. This is contrasted with conventional 3D printing, which emphasizes outside surface texture.SLA is an attractive option for microfluidics due to increasing availability of SLA printers utilizing inexpensive micromirror-based projectors, including some printer kits costing as little as $100. The performance of SLA printers is quantified by the dimensional accuracy and the surface roughness of the printed object. These factors are influenced by the fabrication settings: object orientation, layer thickness, resin properties, and build style. The minimum cross-sectional area of a microchannel made by SLA depends on the laser spot size and the resin viscosity. This resin must be drained post-print. Recent developments have expanded the SLA material selection for a single print to include elastomers and ceramics, while some printers are capable of using multiple resins. It is predicted that in the future it will be possible to use SLA to fabricate metallic sensors and actuators on flexible membranes. Comina et al. used a Miicraft 3D printer to fabricate a reusable mold for PDMS casting. The molds had structures of multiple feature sizes ranging from 50 µm to several mm. The resin had to be manually coated with protective ink to be properly used with PDMS. The Miicraft was also used to print a device with open fluidic channels that are subsequently sealed on top with the adhesive tape. In another study, Comina et al. printed a unibody lab-on-a-chip consisting of a separate microfluidic level, vertical growing systems and a layer with optical components. The integrated finger pump was used to initiate the preparatory sequence of mixing two reagents and three analytes. The colorimetric glucose sensing assay was read by a smart phone. Wang et al. demonstrated an effective approach to fabricating structural devices using 3D printing. A monomer initiator was added to the Miicraft resin to allow for modification of the surface properties such as hydrophobicity and hydrophilicity. Shallan et al. used a Miicraft printer for the fabrication of a transparent microfluidic device with enclosed 250 µm diameter channels. The dimension of the printed channels had a deviation of 50 to 100 µm from the designed dimensions, and the roof of the sealing channel was rough. This could be improved through changing the curing depth, intensity, exposure wavelength, and time. They argued that inexpensive Miicraft provides a sufficient resolution for most microfluidic devices. Patrick et al. printed fluidic open channels using a laser-rastering SLA printer. They found that the smallest achievable diameter of a circular channel was 900 µm, and the smallest channel side for a square channel was 650 µm. The surface topology was inspected using SEM and visible striations were discovered. Overall, the cost of the materials for each fluidic chip was around $6. A sample library of standardized microfluidic components was manufactured using SLA by Lee et al. and Bhargava et al.. These parts were then used to create a number of modular and reconfigurable microfluidic units. This approach allowed for the creation of complex microfluidic designs based on simple interlocking fluidic modules. The library allowed for the further miniaturization of microfluidic elements and materials. The Folch lab has printed diaphragm valves and a peristaltic pump integrated within a LOC device printed using SLA with bio-compatible Somos WaterShed XC 11122 resin. The valves were leakage free at a closing pressure of 6 psi , and they were operated over many cycles.
These valves are regarded as functional modules: two valves can be paired to build a switch, or three valves can be put together in a series to build a pump. However, the portability of these microfluidic devices was limited by the need to use peripherals such as gas canisters.Each of the techniques discussed thus far has its own set of advantages and disadvantages. For example, stereolithography, while allowing for higher resolution, is more expensive and less accessible than fused deposition modelling. Therefore, in this work, we applied a combination of fabrication techniques. In one approach when only an FDM printer is available, we developed a system based on disposable plastic syringes coupled to programmable servomotors. FDM was used to print the non-disposable frame for an automated colorimetric malaria detection test based on enzyme-linked immunosorbent assay . In another approach, when the end user has access to an SLA system, we developed a fabrication approach for SLA to produce elastomeric domes with integrated microfluidic channels. Flow of reagents was facilitated by servomotors compressing the elastomeric dome, and propelling the reagents through a microfluidic channel into a connected test chamber. In this approach, the frame was still printed using an FDM printer. We avoided fabricating microfluidic channels using FDM, because filament-based deposition often resulted in a “staircase effect” as discussed previously, causing fabricated structures to be prone to leakages. The colorimetric assay was designed to detect antibodies in subjects infected with four Plasmodium species that cause malaria. The most significant parasitic diseases in humans are: P. falciparum, P. vivax, P. ovale, and P. malariae. The ELISA kit contained clear polystyrene wells coated with recombinant antigens. When the test sample was added to the well, the specific antibodies in the sample combined with antigens in the well. Subsequently, a conjugate solution of recombinant antigens conjugated to horseradish peroxidase was added to the well, and these antigens reacted with the specific antibodies, if they were present. When the substrate solution of urea peroxide and tetramethyl benzidine was added to the well, it led to a change in solution color from colorless to blue , and finally to yellow when the stop solution was added. It is possible to determine the concentration of the antibodies in the sample by measuring the color intensity with spectral analysis, but in this proof-of-concept study, we limited our experiments to the positive and negative controls provided with the kit. The full details on the recommended volumes of reagents and incubation temperatures were provided by the kit manufacturers and they are summarized in Table 1. We performed incubation not at the recommended 37 ◦C, but at a room temperature, usually 20 ◦C to 23 ◦C. Where the kit instructions recommended draining the test tube and tapping it to make sure that all fluid exited the tube, we implemented automation where a plastic tube is placed within the test well to reach the bottom. This tube was connected to the aspiration syringe whose plunger was pulled out by the gear and rack hardware controlled by the servomotor. Pulling of the plunger created the suction necessary to aspirate solutions from the test well.Extending the IPA soak time of the SLA fabricated sample from 30 min to 6 h helped to clear out channels better. Further extending the soak time in IPA past 6 h did not result in any noticeable change, while the cross-linked resin started to deteriorate.