One-Step Fabrication of Microchannels with Integrated Three Dimensional Features by Hot Intrusion Embossing

We build on the concept of hot intrusion embossing to develop a one-step fabrication method for thermoplastic microfluidic channels containing integrated three-dimensional features. This was accomplished with simple, rapid-to-fabricate imprint templates containing microcavities that locally control the intrusion of heated thermoplastic based on their cross-sectional geometries. The use of circular, rectangular and triangular cavity geometries was demonstrated for the purposes of forming posts, multi-focal length microlense arrays, walls, steps, tapered features and three-dimensional serpentine microchannels. Process variables, such as temperature and pressure, controlled feature dimensions without affecting the overall microchannel geometry. The approach was demonstrated for polycarbonate, cycloolefin copolymer and polystyrene, but in principle is applicable to any thermoplastic. The approach is a step forward towards rapid fabrication of complex, robust, microfluidic platforms with integrated multi-functional elements.


Imprint Templates with Cylindrical Microcavities
In this study, we used two imprint templates with cylindrical microcavities. The first (IT1) was used to create 3 × 3 arrays of round pillars with diameters of 80,70,60,50,40, and 30 μm, as shown in the main paper, describing Figures 1-4. Cavity height for IT1 was hc = 38.5 μm ± 0.2 μm. A second imprint template (IT2) contained cavities of cylindrical pillars and other shapes to demonstrate the versatility in fabricating micro features with different heights in the same channel. Cavity heights for IT2 were hc = 44 μm ± 0.2 μm.

Capillary Number
In general, flow through a capillary is either viscosity-driven or capillary-driven. Here we justify the claim in the main paper that viscous flow dominates capillary flow. In capillary-driven flow, the interfacial tension is a key value in determining the flow velocity, whereas it does not play an appreciable role in viscosity-driven flow. The capillary number (Ca) is a measure of the relative effect of viscous forces to capillary forces and is given in Equation (1): where ηT is the viscosity, v is the polymer melt flow velocity, and σ is the surface tension. A high Ca value means that the flow is dominated by viscous forces. Using the values ηT = 1.5 × 10 7 (Pa·s), v = 0.5 (μm·s −1 ), and σ = 0.033 (N·m −1 ), we see that Ca >> 1, and the flow of the polymer melt through the cylindrical micro cavities in the imprint template is determined to be dominated by viscous flow. Therefore the relationship of polymer flow into the cylindrical cavities in the imprint template given by Equation (1) (main text) is independent of interfacial tension between the polymer melt and the imprint template surface.

Control and Calibration of Hot Intrusion Embossing Using Te and Pe
Next, for the purposes of optimization of the technique, we examined the effect of Te and Pe on the heights of micro pillars in different thermoplastic polymers. In each case, the devices were fabricated using the same imprint template with cylindrical cavities of d = 80 μm. Figure S1 shows the resulting heights of the pillars formed in (a) polycarbonate (PC) and (b) cycloolefin polymer (COP) at different temperatures and pressures. For PC and COP, we identified a temperature and a pressure range that enabled control of h, over the range 10 μm ≤ h ≤ 38.5 μm. For PC, the choice of embossing temperature of 170 °C enables control over pillar height by changing pressures in the range 1.1 MPa ≤ Pe ≤ 1.7 MPa ( Figure S1a). Similarly, we used Te = 170 °C for COP to control the pillar height by changing pressures in the range 0.75 MPa ≤ Pe ≤ 1.1 MP (Figure 3b). For polystyrene (PS) we determined that in the range of 110 °C ≤ Te ≤ 115 °C control over pillar heights 2.5 μm ≤ h ≤ 38.5 μm could be altered by applying pressures 0.55 MPa ≤ Pe ≤ 1.7 MPa (data not shown). Outside these temperature ranges, control of pillar height was not possible for the values of Pe attempted in the ranges listed above for each material. We note that control over pillar heights could also be achieved by maintaining constant embossing pressure and varying the temperature. For example, Figure S1a shows that for Pe = 1.5 MPa, the variation of the temperature in the range 165 °C ≤ Te ≤ 175 °C yielded pillars with heights from 13 to 37 μm. (We verified that channel height remained 38.5 μm for each experiment). In each case the full height of the MF channel was 38.5 μm. Each data point was the result of averaging nine separate measurements, whereas error bars were generated from their standard deviation. Some error bars are smaller than the data points.

Control of Hot Intrusion Embossing Using d
A range of cylindrical cavity diameters used to generate micropillars with controllable heights were generated using 2 different imprint templates (IT1 and IT2, having heights being 38.5 μm and 44 μm, respectively). Figure S2a,c shows the two designs from each IT along with an example of the embossed result (b), (d) and (e). We showed the results of the normalized pillar heights in Figure 2d (main paper). We note that the largest dh, which resulted in the deepest cavity intrusion of heated polymer had tops which were flattened due to contact with the far side of the capillary.

Characterization of Physical Dimensions of Microlenses
Using low pressure and temperature, microlenses were produced by strongly limiting thermoplastic intrusion into the microcavities. We characterized microlens dimensions sing a combination of atomic force microscopy (AFM), optical profilometry and confocal laser scanning microscopy (CLSM). AFM yielded high resolution, continuous vertical cross-section profiles, but data acquisition was very slow and was limited to a 6 μm maximum vertical tip deflection ( Figure S3a). Optical profilometry, could give fast data acquisition over a relatively large surface area, but suffered from missing data points from the sides of the microlenses, which reflected light away from the instrument's objective lens ( Figure S4b). CLSM on the other hand produced continuous, threedimensional data sets with relative ease (e.g., Figure 5a in the main paper). Data files were transformed from three-dimensional image stacks to vertical cross-section profiles using the "reslice" and "find edge" commands in ImageJ and then saving as XY coordinates using the "wand" tool. Therefore, CLSM and AFM results were used to generate vertical cross-sections which were subsequently used for simulation of focal lengths.

Simulation of Microlens Focal Length
For building the 2D geometry of lenses, the interpolated curve of the coordinates extracted from CLSM and AFM measurements with relative tolerance less than 0.01 is used to avoid convergence problems during simulation. Refractive index of n = 1.58 was used for all lenses. All simulation were operated at an incident light wavelength of 500 nm and a free triangular mesh with maximum element size of 100 nm were used. A beam was introduced to the system from the bottom of the lens and a scattering boundary condition is used on the wall to avoid back reflectance of light. For analyzing the result, the normalized power flow, time average of the beam was plotted versus the Cartesian coordinates which for a two dimensional system can be acquired from the Equation (2):

=
( 2) where Px and Py are x and y components of power flow, time average respectively.

Temperature Dependence on Height of Partially-Occluding Straight Walls
As discussed in the main paper, thin trenches could be used to place partially blocking walls in microchannels in order to locally modify the channel depth. Figure S4 shows the results from a mask designed to produce a channel with thin trenches of 20 and 25 μm. As discussed in the main paper, the small difference in width resulted in different wall heights, depending on the embossing conditions. The heights of the walls were measured to be 28.7 (i) μm and 34.0 μm (ii). Using the same imprint template and embossing at the same pressure (Pe = 1.5 MPa), a second PS microfluidic device was fabricated with Te = 105 °C. The higher viscosity resulted in slower flow into the microtrenches leading to wall heights of 12.5 μm (i) and 16.4 μm (ii).  Figure S4 shows the array of cylindrical microcavities used to produce microlens arrays with different focal lengths. Colours blue and red showing the positions that were measured for Figure 5 in the main text.