Microfluidics involves the precise manipulation of fluids at submillimeter scales leveraging fabrication technologies developed by the semiconductor and microelectromechanical system (MEMS) industries [1
]. A device built using microfluidic principles is commonly referred to as a micro total analysis system (µTAS) [2
] or a Lab-on-a-Chip (LoC) [1
]. There are certain distinct advantages to conducting experiments in a microfluidic setting rather than on the macroscale [1
]. Microfluidic chips need substantially smaller sample volumes that reduce the cost of reagents. They allow simplified analysis of multiple samples in parallel to generate maximum data per batch, while also providing greater control of spatiotemporal fluid dynamics. Additionally, multiple targets can be analyzed on the same sample. As these advantages translate well for use in biomedical research [3
], a microfluidic setting provides an ideal platform for portable, point-of-care diagnostic devices [5
Microfluidic systems almost always operate in the laminar flow regime leading to predictable fluid dynamics. In the absence of convective mixing, molecular transport is dominated by diffusion-based kinetics [1
]. While these properties are desirable, they necessitate careful selection of materials to be used in the fabrication of microfluidic chips. For example, a material that strongly absorbs solutes can quickly deplete solutions in the channels of a microfluidic chip [6
Polydimethylsiloxane (PDMS) is a silicon-based organic polymer that is most commonly used as a microfluidic material because of its optical clarity, biocompatibility, ease of molding, and flexibility [7
]. However, PDMS has several disadvantages. It is susceptible to channel deformation due to its softness [8
]. It is also known to leach uncured oligomers into the channel solution [11
]. This can lead to additional steps to negate such leaching [12
]. Hydrophobic interactions are a key driving force in biological phenomena and form the basis for drug design and discovery [15
]. PDMS is known to strongly absorb small hydrophobic molecules [17
]. Furthermore, the high vapor permeability of PDMS can lead to evaporation [19
]. This can be harmful to cells at the scale of a microfluidic experiment [20
]. Steps such as parylene coating [22
] can resolve this issue but are not ideal for cell biology applications [1
]. These drawbacks have resulted in a declining interest in PDMS biomicrofluidics [1
]. Common alternatives to using PDMS include glass, polycarbonate (PC), polymethylmethacrylate (PMMA), cyclic olefin copolymers (COC), polyimides (PI), and polyurethanes (PU) [24
PUs are a broad class of polymers that are most commonly formed by reacting a diisocyanate with a polyol [26
]. Depending on the proprietary polyol curative used, the hardness of the resultant PU can vary from Shore A [6
] through D. Unlike PDMS, PUs are compatible with organic solvents [25
] and several aqueous solutions under 0.5 M [26
]. This has led to many PU-based microfluidic studies [6
]. Xia et al. [32
] were the first to demonstrate that certain UV-curable PUs can be used in high-fidelity replica molding techniques past the micrometer scale with results similar to those for PDMS. Furthermore, the increased stiffness of PUs in comparison to PDMS allows fabrication of structures with greater aspect ratios that are less susceptible to ground and lateral collapse [33
]. PUs have also been used to create thermally-actuated microfluidic valves that are more stable than PDMS valves due to decreased evaporation [30
]. Finally, PUs have been used in biomicrofluidics, as soft bottom layers in hybrid microfluidic chips [31
] and to form whole PU elastomeric chips for cell culture [6
An increasing number of recent microfluidic studies involve chips made of hard plastics because of their suitability for use in modular microfluidics [34
]. A similar trend can be seen in biomicrofluidic studies with a move away from PDMS in favor of hard plastics like PS that have traditionally played a large and well understood role in cell experiments in vitro [35
]. Harder materials are also good candidates for surface modification techniques. One such technique—NanoAccel™ treatment in Accelerated Neutral Atom Beam (ANAB) mode [36
]—can physically roughen surfaces using neutral argon atoms. It has been used on polyether ether ketone (PEEK) to improve its bioactivity towards cell attachment and proliferation.
Another trend in microfluidic fabrication is the increasing use of 3D-printing-based replica molding strategies [38
] due to the low-cost of 3D printers, the reusability of 3D printed molds to create multiple microfluidic layers from the same mold and advancements in bonding rough layers to form sealed microfluidic chips.
This study stems from an ever-increasing body of research on hard plastics for microfluidics instead of PDMS due to its numerous disadvantages, the versatility and past success of softer PUs in microfluidics (including biomicrofluidics), and the relative ease with which PUs fit into simple 3D-printing-based replica molding and chip assembly strategies. To introduce a clear and castable alternative to PDMS in biomicrofluidic applications, we report the first characterization of a commercially available Shore D pour-and-cure-type, two-component PU resin (Ultraclear™ 480N with hardness 80D from Hapco, Inc.). We also demonstrate control over its hydrophilic behavior, by describing the first utilization and evaluation of NanoAccel™ ANAB treatment on PU surfaces.
2. Materials and Methods
2.1. General Recipe for PDMS, Blue Silicone R-2374 and PU
PDMS (Sylgard® 184, Dow Silicones Corporation, Midland, MI, USA) and Blue Silicone R-2374 (Silpak, Inc., Pomona, CA, USA) had similar preparations—mixing 10:1 wt % (polymer:cross-linker). PU (Ultraclear™ 480N-10 and 480N-60, Hapco, Inc., Hanover, MA, USA) was prepared by mixing 1:1 wt % (part A:part B). Part A consists of 10-20 wt % of proprietary polyether polyol prepolymer capped with 80-90 wt % of 4,4′-methylene dicyclohexyl diisocyanate (H12MDI). Part B consists of 95–100 wt % of a proprietary polyether polyol combination. Ultraclear™ 480N-10 and 480N-60 have gel times of 10 and 60 min, respectively. After mixing, all uncured polymers were degassed for 30 min and cured at 65 °C for 2 h using a vacuum oven (Heraeus D-6450, Heraeus Instruments GmbH, Hanau, Germany).
2.2. Optical Transmittance Studies
UV-Vis-NIR transmission spectra from 200 nm–850 nm were measured for three polymers—PDMS (Sylgard® 184), fast-gelling PU (Ultraclear™ 480N-10), and slow-gelling PU (Ultraclear™ 480N-60). PDMS and PU were prepared, as described in the general recipe section. 1 mL of uncured samples was cured in 1.5 mL PS cuvettes (Brand GmbH + Co. KG, Wertheim, Germany) with a path length of 1 cm. A Cary® 50 Spectrophotometer (Agilent Technologies, Santa Clara, CA, USA) was used to collect spectral data. Three independent experiments were conducted in triplicate.
2.3. Feature Range Characterization
SU-8 (MicroChem Corp., Westborough, MA, USA) patterns on Si wafers were used to create masters. To maintain a rigid-flexible-rigid replica molding strategy, masters with features from the millimeter to the nanometer scale were used to develop flexible PDMS stamps. 1 cm layers of uncured PDMS were poured onto the masters enclosed in 150 mm (diameter) Petri dishes (Fisherbrand™, Thermo Fisher Scientific, Waltham, MA, USA) and cured. Cured PDMS stamps were peeled off the molds and used to generate PU replicas in a similar manner. Finally, PDMS layers were peeled off to leave PU blocks with the same features as the SU-8 patterned Si master.
2.4. Scanning Electron Microscopy
PU samples were cut into 2 cm wide squares for Scanning Electron Microscopy (SEM). PU samples were sputtered with Au/Pd (60:40) in a Denton Vacuum Desk IV® (Denton Vacuum, LLC, Moorestown, NJ, USA) using 30 mA for 75 s to avoid excessive charging, and then mounted with carbon tape. SEM images were collected using a LEO 1550 (Carl Zeiss Microscopy, LLC, Thornwood, NY, USA) at accelerating voltages between 1 kV–3 kV and magnifications between 125 X-25 kX.
2.5. Surface Modification by Corona Treatment
5 mm layers of PU were cured in 60 mm (diameter) Petri dishes. A BD-20AC Laboratory Corona Treater (Electro-Technic Products, Inc., Chicago, IL, USA) with a field effect electrode was used to treat each sample for 60 s.
2.6. Surface Modification by NanoAccel™ Treatment in ANAB Mode
5 mm layers of PU were cured in 60 mm (diameter) Petri dishes. Briefly, large clusters (~1000–5000 atoms/cluster) of argon gas were ionized and accelerated to 30 keV in a vacuum (with base pressure of 6.5E-7 Torr). By promoting cluster dissociation and deflecting charged cluster-fragments away, a beam of neutral argon atoms impinged on the PU surfaces with average kinetic energies in the 10–100 eV range. To ensure that all samples were subjected to the same vacuum and the parameter to be measured was affected by ANAB treatment alone, untreated samples were also placed in the NanoAccel™ tool with the beam blocked by a Ni mask.
2.7. Water Contact Angle Measurement
Water contact angle was measured for three different surfaces—Untreated PU, ANAB-treated PU, and corona-treated PU. 5 mm layers of PU were cured in 60 mm (diameter) Petri dishes. Cured samples were cut into 20 mm wide squares. A Cam-Plus Micro contact angle meter (ChemInstruments, Fairfield, OH, USA) was used to measure the water contact angle for 2 µL drops at ten randomly chosen spots across each surface. Drops were allowed to stabilize on the surface for 90 s before measurement of contact angle by the Half-Angle method. Three independent experiments were conducted with measurement of ten drops for each sample.
2.8. Atomic Force Microscopy
Images for ANAB-treated PU and untreated PU were taken using a Dimension Icon (Bruker, Billerica, MA, USA) in tapping mode for three randomly chosen spots on each sample. 10 × 10 µm2 and 5 × 5 µm2 images were captured and RMS roughness (Rq) values were recorded. 2D isotropic power spectral density plots were generated using NanoScope Analysis 1.8 (Bruker).
2.9. Surface Energy Estimation
Surface energy of untreated PU and ANAB-treated PU was estimated by the Owens-Wendt method [40
] using three test liquids—water, formamide, and diiodomethane. Surface free energy parameters were taken from a PU study by Krol et al. and are listed in Table S1
]. Contact angle measurement was conducted as in Section 2.7
. Average contact angle values were used to calculate the polar and dispersive parts of surface energy from the Owens-Wendt equations. Three independent experiments were conducted with measurement of ten drops for each sample.
2.10. Fourier-Transform Infrared Spectroscopy in Attenuated Total Reflectance Mode
Attentuated total reflectance-Fourier-transform Infrared (ATR-FTIR) spectroscopy was used for chemical surface characterization of untreated PU and ANAB-treated PU. A Tensor 27 (Bruker, Billerica, MA, USA) with a PIKE MIRacle™ ATR accessory (PIKE Technologies, Madison, WI, USA) and a ZnSe crystal was used to collect spectra between 520–4000 cm−1 at a resolution of 4 cm−1. Each spectrum collected was an average of 128 scans. PU samples were 2 mm thick and clamped down to the crystal using the accessory. Baseline correction was performed in SpectraGryph 1.2.
2.11. Cell Viability Studies
1 cm layers of PU were cured in 35 mm (diameter) Petri dishes. After UV sterilization for 1 h in a cell culture hood, samples were rinsed with 1X PBS (Phosphate Buffered Saline from Gibco™, Life Technologies, Thermo Fisher Scientific, Carlsbad, CA, USA) three times. 75000 MDA-MB-231 cells/mL were seeded onto PU surfaces with Corning™ DMEM (Dulbecco’s Modified Eagle Media from Fisher Scientific, Hampton, NH, USA) + 10% FBS (Fetal Bovine Serum from Millipore Sigma, Burlington, MA, USA) + 1% Pen-Strep (Penicillin-Streptomycin from Millipore Sigma, Burlington, MA, USA) and grown in an incubator at 37 °C and 5% CO2.
Live-dead staining by 0.4% Trypan Blue Dye (Bio-Rad, Hercules, CA, USA) exclusion was used to quantify the percentage of cell death after 24 h. Cells were monitored over the next 3 days and passaged to verify trypsinization on PU surfaces. Cells on sample sets were imaged before and after passaging to visualize differences in cell adhesion. Four runs were conducted in duplicate.
2.12. Chip Fabrication
A 3D-printing-based replica molding strategy was used to fabricate PU microfluidic layers. Autodesk® Inventor® designs were 3D printed using a Form 1+ (FormLabs, Somerville, MA, USA). All parts were printed at an axis resolution of 25 µm. Printed parts were separated from supporting frameworks and cleaned with isopropanol to remove uncured resin. Parts were then coated with 3 mL of Sigmacote® (Millipore Sigma, Burlington, MA, USA) to facilitate the removal of stamps. Uncured Blue Silicone R-2374 was poured into the printed parts and cured. Once cured, stamps were peeled off and used as molds for PU. Blue Silicone R-2374 stamps were preferred over PDMS stamps because they were found to be easier to remove from the 3D-printed masters. Uncured PU was poured into Blue Silicone R-2374 stamps. Once cured, stamps were peeled away to leave PU microfluidic layers. Holes were drilled for the inlets and outlets. PS cut-outs of required size were made from Falcon® cell culture Petri dishes (Corning Inc., Corning, NY, USA).
PU microfluidic layers were bonded to PS sheets using UV-curable adhesive NOA-63 (Norland Products, Inc., Cranbury, NJ, USA) to form PU-PS chips. Since chip features in biomicrofluidics are much larger than regular microfluidics, stamping-based methods [45
] were unnecessary. A razor blade was used to level the adhesive on the microfluidic layer. The PS sheet was then gently placed on the NOA-63 coated-PU microfluidic layer and slight pressure was applied to remove excess adhesive and air bubbles while releasing boundary tension. Following a strategy used by Dang et al. [47
], smaller PU-PS chips were designed with a sacrificial channel to prevent clogging of the microfluidic channels with NOA-63.
A declining interest in using PDMS for microfluidics has led to the characterization of novel materials for rapid prototyping. This trend has continued in biomicrofluidic research. PU-based polymers are often used in biomicrofluidics because they do not absorb hydrophobic molecules and have better solvent compatibility and stiffness than PDMS. We report the first characterization of a commercially available Shore 80D PU for biomicrofluidic applications.
Ultraclear™ PU has an optical transmittance similar to PDMS in the Vis-NIR range. It can be used reliably, with replica molding strategies, to fabricate solid structures across the breadth of the microfluidic range. Unlike other PUs, Ultraclear™ PU is hydrophobic after curing. Corona treatment causes a temporary gain in hydrophilicity. To demonstrate applicability for biomicrofluidic studies, we report the first use of NanoAccel™ neutral atom beam surface modification of PU surfaces to permanently roughen the surface of Ultraclear™ PU and reduce its water contact angle. Surface energy measurements using Owens-Wendt equations demonstrate an increase in polar surface energy with increasing treatment flux. ATR-FTIR spectra prove that no new functional groups are introduced on the surface due to treatment. The improved surface roughness and hydrophilic behavior also favors MDA-MB-231 cell adhesion. Lastly, to demonstrate applicability in rapid prototyping, a 3D-printing-based replica molding strategy is utilized to create PU microfluidic layers that are sealed to PS using adhesive bonding. As a proof of concept, two versions of PU-PS chips were made. Overall, we demonstrate that Ultraclear™ PU is a clear, castable alternative to PDMS for use in rapid prototyping and biomicrofluidics. Future directions include testing the potential of NanoAccel™ treatment to pattern Ultraclear™ PU surfaces with specific hydrophilic and hydrophobic regions and incorporating such strategies into microfluidic chip usage.