*2.9. Biofunctionalization*

The immobilization of proteins inside of FDM printed microfluidic channels was done similar to a previously described protocol [28]. F5B (80 µM in PBS) was injected into the channel and exposed for 3 min to 5 min at 490 nm (exposure intensity: 7.7 mW·cm−<sup>2</sup> ). After exposure with a custom build lithography system based on a digital mirror device the channels were rinsed with distilled water, PBS and, again, distilled water [28]. The patterns were visualized by incubating the microchannel with STV-Cy3 (5 µg/mL) for 30 min. Afterwards the channels were again rinsed with distilled water and PBS. The biomolecule patterns were visualized using an inverted fluorescence microscope (DMi8, Leica, Germany). Images were analyzed using ImageJ 1.53a. Regions of interest (ROI) were selected for background and signals (triangle shapes) and were analyzed in terms of brightness. Mean values of all ROI of background and signals respectively were averaged and average values were used to calculate signal-to-noise ratio (S/N).

### **3. Results and Discussion**

We optimized the printing process to fabricate microfluidic chips in PMMA. Furthermore, several microfluidic devices were fabricated and assessed. The integration of high-resolution microstructures and biofunctionalization inside the channel was demonstrated.

### *3.1. Optimum Printing Parameters and Transparency Optimization* which is printed using 100% infill for high transparency. A minimum of two layers (0.15 mm) were deposited before printing the actual channel structures. This allows printing of embedded, leak-proof

*3.1. Optimum Printing Parameters and Transparency Optimization* 

To print PMMA microfluidic chips with an optical transparency in the region of interest, we studied two different printing strategies: (1) standard printing procedure with direct printing on the print bed and (2) printing onto a commercial PMMA substrate. Both strategies and the respective printed microfluidic channels are shown in Figure 1a,b. First, we printed PMMA microfluidic chips using the standard printing setup, where the PMMA is deposited directly on the print bed. To ensure a completely sealed microfluidic structure, the printed channels are embedded in the PMMA chip, which is printed using 100% infill for high transparency. A minimum of two layers (0.15 mm) were deposited before printing the actual channel structures. This allows printing of embedded, leak-proof and transparent microfluidic chips, as shown in Figure 1b. However, the deposited PMMA strands from the two bottom layers are clearly visible and therefore reduce the transparency of the printed microfluidic chips. In the second strategy, we printed an open channel structure directly onto a commercial, highly transparent PMMA substrate. The heated PMMA structure bonds to the commercial PMMA substrate upon deposition and therefore allows printing of fully functional, embedded and transparent microfluidic channels. The absence of FDM printed PMMA layers between the printed channel structures and the commercial PMMA substrate allows 3D printing of PMMA microfluidics with a higher transparency than with the standard printing procedure. and transparent microfluidic chips, as shown in Figure 1b. However, the deposited PMMA strands from the two bottom layers are clearly visible and therefore reduce the transparency of the printed microfluidic chips. In the second strategy, we printed an open channel structure directly onto a commercial, highly transparent PMMA substrate. The heated PMMA structure bonds to the commercial PMMA substrate upon deposition and therefore allows printing of fully functional, embedded and transparent microfluidic channels. The absence of FDM printed PMMA layers between the printed channel structures and the commercial PMMA substrate allows 3D printing of PMMA microfluidics with a higher transparency than with the standard printing procedure. We further optimized the printing process to improve the overall transparency of the FDM printed PMMA. It was found that the *z*-axis distance of the nozzle for printing the first layer has a significant impact on the observed transparency of the FDM printed PMMA. Printing with the standard parameters for *z*-axis distance, specified by the Prusa manual, PMMA with poor optical transparency is obtained. Reducing the first layer *z*-axis distance of the nozzle by 100 µm compared to the standard parameters (Prusa manual) results in PMMA microfluidic devices with significantly higher transparency (see Figure 1c,d). The reduced nozzle distance results in slightly broader deposited PMMA strands that fuse together without having gaps in between thus increasing the transparency of the FDM printed component significantly.

*Micromachines* **2020**, *11*, 873 5 of 13

a completely sealed microfluidic structure, the printed channels are embedded in the PMMA chip,

To print PMMA microfluidic chips with an optical transparency in the region of interest, we studied two different printing strategies: (1) standard printing procedure with direct printing on the print bed and (2) printing onto a commercial PMMA substrate. Both strategies and the respective printed microfluidic channels are shown in Figure 1a,b. First, we printed PMMA microfluidic chips

**Figure 1.** Fused deposition modeling of microfluidic chips in polymethylmethacrylate (PMMA). (**a**) Principle of printing microfluidic PMMA channels. PMMA was either directly printed on the print bed (**I**) or printed on top of a commercial PMMA slide (**II**). (**b**) Comparison of two identical microfluidic chips directly printed on the print bed and printed on a commercial PMMA substrate, respectively. As can be seen, the transparency in the region of interest is increased by printing on a commercial PMMA substrate (scale bars: 10 mm). (**c**) Increasing the optical transparency by reducing the nozzle distance: (**I**) standard configuration, (**II**) printing with a reduced nozzle distance which **Figure 1.** Fused deposition modeling of microfluidic chips in polymethylmethacrylate (PMMA). (**a**) Principle of printing microfluidic PMMA channels. PMMA was either directly printed on the print bed (**I**) or printed on top of a commercial PMMA slide (**II**). (**b**) Comparison of two identical microfluidic chips directly printed on the print bed and printed on a commercial PMMA substrate, respectively. As can be seen, the transparency in the region of interest is increased by printing on a commercial PMMA substrate (scale bars: 10 mm). (**c**) Increasing the optical transparency by reducing the nozzle distance: (**I**) standard configuration, (**II**) printing with a reduced nozzle distance which flattens the extruded filament. The nozzle is moving in the z-direction. The images show the 3D printed cross-section of the first layer calibration (scale bar: 100 µm). (**d**) Comparison of microfluidic channels printed with standard configuration and with reduced nozzle distance. As can be seen the transparency is increased for the printing setup with reduced nozzle distance (scale bar: 10 mm).

We further optimized the printing process to improve the overall transparency of the FDM printed PMMA. It was found that the *z*-axis distance of the nozzle for printing the first layer has a significant impact on the observed transparency of the FDM printed PMMA. Printing with the standard parameters for *z*-axis distance, specified by the Prusa manual, PMMA with poor optical transparency is obtained. Reducing the first layer *z*-axis distance of the nozzle by 100 µm compared to the standard parameters (Prusa manual) results in PMMA microfluidic devices with significantly higher transparency (see Figure 1c,d). The reduced nozzle distance results in slightly broader deposited PMMA strands that fuse together without having gaps in between thus increasing the transparency of the FDM printed component significantly.

We printed different channel cross-sections to evaluate the printing accuracy of these cross-sections and to determine the best cross-section for printing more complex embedded microfluidic channels. The CAD design and its respective printed PMMA chip can be seen in Figure 2a,b. FDM printing of a square cross-sections results in an excessive sagging of the top layer, which will clog the channel if small channel sizes are printed. The same effect was found for circular cross-sections. A major improvement was found when adding a triangular shaped roof on top of the square cross-section. The roof hereby compensates for the sagging effect and results in the square channel cross-sections. A similar optimization was found for elliptical and diamond shaped channels. Since the additional roof resulted in the most accurate microfluidic channels, all further structures were printed using this design. *Micromachines* **2020**, *11*, 873 7 of 13

**Figure 2.** Characterization of the channel cross-section of a printed PMMA chip. (**a**) Image of the design of the printed channel cross-sections. (**b**) Different channel cross-sections with a channel width/diameter of 600 µm printed in PMMA. (**c**) Fused deposition modeling (FDM) printed embedded channels with a square cross-section for analysis of the printed channel height. Due to sagging the channel height deviates strongly from the original CAD design. (**d**) FDM printed embedded channels with a roof shaped cross-section for analysis of printed channel height and width. The addition of the roof compensates the sagging resulting in square channels. (**e**) Open square crosssection channels for analysis of the channel height and width. (**f**) Comparison of channel heights of embedded FDM printed channels using square and roof shaped cross-sections shown in (**c**,**d**) and open channels shown in (**e**) with the designed CAD heights. (**g**) Channel width of open and embedded roof shaped FDM printed channels from (**d**,**e**) compared to the designed CAD widths. (Scale bars (**a**– **e**): 500 µm). **Figure 2.** Characterization of the channel cross-section of a printed PMMA chip. (**a**) Image of the design of the printed channel cross-sections. (**b**) Different channel cross-sections with a channel width/diameter of 600 µm printed in PMMA. (**c**) Fused deposition modeling (FDM) printed embedded channels with a square cross-section for analysis of the printed channel height. Due to sagging the channel height deviates strongly from the original CAD design. (**d**) FDM printed embedded channels with a roof shaped cross-section for analysis of printed channel height and width. The addition of the roof compensates the sagging resulting in square channels. (**e**) Open square cross-section channels for analysis of the channel height and width. (**f**) Comparison of channel heights of embedded FDM printed channels using square and roof shaped cross-sections shown in (**c**,**d**) and open channels shown in (**e**) with the designed CAD heights. (**g**) Channel width of open and embedded roof shaped FDM printed channels from (**d**,**e**) compared to the designed CAD widths. (Scale bars (**a**–**e**): 500 µm).

color gradient from yellow to blue (see Figure 3a). We further printed a simple microfluidic spiral with a channel width of 600 µm using the roof-shaped channel cross-section CAD design (see Figure 3b). We also show an enhanced mixer which consists of a 2 mm wide channel with 600 µm periodic geometries reaching into the channel (see Figure 3c). All three designs were printed using the roofshaped channel cross-section CAD design. The mixers in Figure 3a,c show effective mixing of blue and yellow colored water. To demonstrate the feasibility of truly three-dimensional geometries, we also printed a 3D microfluidic spiral with 1.2 mm cross-section, which is intertwining around a straight channel (see Figure 3d). To test their performance, the printed channels were flushed with liquid at a flow rate of up to 10 mL/min for several minutes. No leakage occurred during these

We printed several exemplary microfluidic devices to demonstrate the versatility of FDM 3D

*3.2. 3D Printed Microfluidic Devices* 

experiments and none of the channels showed clogging.

The accuracy of the printed channel dimensions was evaluated by printing open and embedded (roof and square cross-section) test channels of 1 mm, 800 µm, 600 µm, 400 µm and 300 µm channel width and height, respectively (Figure 2c–e). The measured height of the resulting channel cross-sections compared to the CAD dimensions are shown in Figure 2f. Printing embedded square cross-sections resulted in excessive sagging which reduced the accuracy of the printed height compared to the CAD design significantly. Adding a roof on top of the square cross-section compensated for the sagging and allowed 3D printing of square channels with their channel heights corresponding to the square CAD design without the added roof. The open channels showed nearly no deviation from the CAD design. Both the open and embedded roof cross-section showed a high accuracy down to a channel height of 300 µm.

We further compared the width of the resulting channel cross-sections to the CAD dimensions (see Figure 2g). The actual printed channel widths of the open and roof-shaped cross-section channels are only slightly smaller than their respective CAD dimensions. This small deviation can be explained by the employed printing setup. Due to the reduced nozzle distance, used for printing of high transparency structures, the deposited PMMA strands are pressed to a slightly broader size than estimated by the slicing software. Printing of channels with a width and height below 300 µm resulted in partial clogging and was therefore not further investigated.

### *3.2. 3D Printed Microfluidic Devices*

We printed several exemplary microfluidic devices to demonstrate the versatility of FDM 3D printing of PMMA (see Figure 3). An exemplary serpentine microfluidic mixer with a channel diameter of 600 µm was printed. It shows effective mixing along the channel cascade, resulting in a color gradient from yellow to blue (see Figure 3a). We further printed a simple microfluidic spiral with a channel width of 600 µm using the roof-shaped channel cross-section CAD design (see Figure 3b). We also show an enhanced mixer which consists of a 2 mm wide channel with 600 µm periodic geometries reaching into the channel (see Figure 3c). All three designs were printed using the roof-shaped channel cross-section CAD design. The mixers in Figure 3a,c show effective mixing of blue and yellow colored water. To demonstrate the feasibility of truly three-dimensional geometries, we also printed a 3D microfluidic spiral with 1.2 mm cross-section, which is intertwining around a straight channel (see Figure 3d). To test their performance, the printed channels were flushed with liquid at a flow rate of up to 10 mL/min for several minutes. No leakage occurred during these experiments and none of the channels showed clogging.
