micromachines

Article Development of a Multi-Material Stereolithography 3D Printing Device

Bilal Khatri 1, Marco Frey 1, Ahmed Raouf-Fahmy 1, Marc-Vincent Scharla 1 and Thomas Hanemann 1,2,*

1 Department of Microsystems Engineering, University of Freiburg, Georges-Koehler-Allee 102, D-79110 Freiburg, Germany; [email protected] (B.K.); [email protected] (M.F.); [email protected] (A.R.-F.); [email protected] (M.-V.S.) 2 Karlsruhe Institute of Technology, Institute for Applied Materials, Hermann-von-Helmholtz-Platz 1, D-76344 Eggenstein-Leopoldshafen, Germany * Correspondence: [email protected]



Received: 17 April 2020; Accepted: 21 May 2020; Published: 22 May 2020


Abstract: Additive manufacturing, or nowadays more popularly entitled as 3D printing, enables a fast realization of polymer, metal, ceramic or composite devices, which often cannot be fabricated with conventional methods. One critical issue for a continuation of this success story is the generation of multi-material devices. Whilst in fused filament fabrication or 3D InkJet printing, commercial solutions have been realized, in stereolithography only very few attempts have been seen. In this work, a comprehensive approach, covering the construction, material development, software control and multi-material printing is presented for the fabrication of structural details in the micrometer range. The work concludes with a critical evaluation and possible improvements.

Keywords: additive manufacturing; stereolithography; photopolymers; multi-material 3D printing

1. Introduction Starting with the brilliant invention described in “Apparatus for production of three-dimensional objects by stereolithography” in 1986 by Charles W. Hull [1], the development of different 3D printing techniques has happened at a brisk pace. This has made possible the realization of polymer, ceramic and metallic parts with geometrical features previously inconceivable due to the topological limitations of traditional fabrication methods, such as mechanical machining and injection molding. In addition to stereolithography (SLA), a wide variety of methods are now established under the umbrella of additive manufacturing, such as fused filament fabrication (FFF), 3D InkJet printing (also known as PolyJet®) or powder-based printing (binder jetting, laser or electron sintering or melting (SLS, SLM, EBM)). This speedy advancement of additive manufacturing technologies, as well as that of 3D printable materials, such as curable resins for SLA or PolyJet®, thermoplastics for FFF and fine metal powders for SLM or EBM, have led to a technological readiness for reliable rapid prototyping and, in some cases, even small-scale production [2–6]. However, and in contrast to established macroscale fabrication methods, there are still some open questions and fields, which require solving and answering for further development of this emerging field [7]. In addition to design and shape optimization, development of printing methodologies, digital material development, error control and modeling issues, a key challenge for the further dissemination of 3D printing into the industry is that of materials, in particular multi-material printing. In FFF, the printing of functional polymer matrix composites with ceramic and metal fillers is established [8–10], even for the realization of dense ceramic and metal parts, in an approach analogous to powder injection molding [11–13]. A first two-component FFF, combining the printing of zirconia and stainless steel filaments and subsequent thermal post-processing was

Micromachines 2020, 11, 532; doi:10.3390/mi11050532 www.mdpi.com/journal/micromachines Micromachines 2020, 11, 532 2 of 17 reported in 2019 [14]. Multi-material printing for FFF was first reported in 2002 [15], dealing with the modeling of the printing strategy optimization. Since then, the combination of electroplating and FFF has been investigated by Matsuzaki and recently by Ambrosi et al. [16,17]; a comprehensive review can be found in [18]. In the case of SLA, the composite printing of ceramic-filled resins, curing and thermal post-processing has been commercialized [19,20]. On the multi-material SLA side, an early work dealing with multi-material stereolithography was published in 2006 by Inamdar et al. [21] using a modified commercial SLA machine (3D Systems 250/50) and the implementation of a rotating vat carousel system. The vats had a volume around 9 L and a resin pump filling/leveling system. The resin polymerization was induced by a 355 nm solid state laser. The authors claimed a precision of the z-stage of 20 µm and a repeatability of ± 1 µm controlling the build layer thickness. Unfortunately, the printed test samples were not further ± characterized geometrically [21]. A more detailed description was given in [22], also introducing a rotary platform stage. The same research group developed an alternative system using only one vat; the resins can be exchanged using a syringe pump after manual cleaning and rinsing [23]. In the latter case, a layer thickness around 21 µm was possible. However, details about the x, y-resolutions were not discussed [23]. A comprehensive review summarizes both developments [24]. The authors extended the portfolio of usable curable materials to, e.g., polyethylenglycol (PEG) based hydrogels or conductive layers for electronic circuit fabrication [24]. In a hybrid approach, combining stereolithography with aerosol jet printing, a continuous mixing ratio of two different resins and subsequent UV curing was realized [25]. Roach et al. [26] implemented a combination of different 3D printing methods and related materials in 2019. Another early multi-material SLA system operating with two vats was presented by Zhou et al., in 2011, applying a digital micromirror device (DMD) serving as a digital light processing (DLP) modulator with a resolution of 1024 768 pixels [27]. The achieved structural × details were in the sub-millimeter range. Two different polyethylene diacrylate-based hydrogels have been used for the printing of micro-channels, including diffusion barriers also using a DMD projector and intermediate vat change. Using this system, micro-channels with dimensions around 200 µm could be realized [28]. Another hybrid approach, combining DLP–SLA with a drop-on-demand (DoD) Inkjet printing system, was presented by Muguruza et al. in 2017 [29]. As an upcoming trend, the combinations of different additive manufacturing techniques for the realization of multi-materials combining individual strengths is gaining more prominence [30]. For the additive manufacturing of polymeric multi-material components, three different methods should be considered: Extrusion-based techniques like FFF and the ones making use of curable resins, like 3D InkJet (PolyJet®) and SLA. Modern FFF two-component printers are commercially available below €10,000 and show resolutions of 20 µm in z- and 100 µm in x-, y-directions on paper. Realistically achievable values are at least double the given values for commercial thermoplastics and may not be repeatable for each printing position or shape. The realization of structural features below 100 µm is hindered by these issues. The use of functional composites, e.g., with dielectric or magnetic properties for the realization of ceramic or metallic parts, requires additional melt-compounding steps such as mixing–kneading and filament extrusion [8,9,11,31]. Due to the round filament shape and its deposition in a layer-by-layer manner, poor inter-layer adhesion within one component is one of the major drawbacks of single-material [31] and two-component FFF printing [32]. Commercial multi-component 3D InkJet printers are available at prices upwards of €30,000, enabling geometric resolutions better than 50 µm in x-, y- and z-directions. Unfortunately, the user is restricted to a set of fixed materials and closed software, both offered by the machine vendor. Material development for DoD printing is difficult due to its piezo-driven inkjet print-head, because the ink viscosity must be lower than 20 mPas at the printing temperature and all dispersed particles for the realization of functional composites must be smaller than 1 µm in order to avoid print-head nozzle clogging [33,34]. Individual layers cannot be detected after printing and post-curing, resulting in good mechanical properties. Currently, SLA is the best compromise, considering different aspects such as the material portfolio, potential for new material development, multi-material printing, geometric resolution, machine Micromachines 2020, 11, 532 3 of 17

Micromachines 2020, 11, x FOR PEER REVIEW 3 of 17 investment and consumable materials. The only disadvantage is the limitation of photocurable resins as theThe build aim material of this due work its operating is the principle.development of a new, versatile, multi-material micro- stereolithographyThe aim of this 3D work printing is the device development (MMSL), of adev new,eloped versatile, using multi-material low-cost commercial micro-stereolithography equipment and 3Dcomponents, printing device with the (MMSL), usage of, developed at present, using up low-costto three individual commercial curable equipment resins and without components, intermediate with thematerial usage exchange, of, at present, enabling up to threea more individual flexible curablematerial resins combination. without intermediate The functionality material of exchange,the new enablingMMSL device a more will flexible be shown material by combination.the combination The of functionality a variety of ofcurable the new acrylates MMSL devicewith different will be shownfluorescence by the markers. combination of a variety of curable acrylates with different fluorescence markers.

2. Materials and Methods Following earlier earlier investigations investigations on on the thephotocuring photocuring behavior behavior of diacrylates, of diacrylates, ethylene ethylene glycol glycoldimethacrylate dimethacrylate (EGDMA, (EGDMA, Merck, Darmstadt, Merck, Darmstadt, Germany, Germany,Figure 1a) Figureand bisphenol1a) and bisphenolA glycerolate A glycerolatedimethacrylate dimethacrylate (BAEDA, Merck, (BAEDA, Darmstadt, Merck, Darmstadt,Germany, Germany,Figure 1b) Figure were1 b)selected were selectedas suitable as suitablemonomers monomers for approval for as approval a curable as base a curablematerial base for MMSL material development for MMSL [35,36]. development Diphenyl(2,4,6- [35,36]. Diphenyl(2,4,6-trimethylbenzoyl)phosphinetrimethylbenzoyl)phosphine oxide (TPO, TCI, oxide Frankfurt, (TPO, Germany) TCI, Frankfurt, was selected Germany) as a photoinitiator was selected asdue a photoinitiatorto its good light due absorption to its good in lightthe range absorption between in the350 rangeand 430 between nm and 350 with and ratios 430 nm between and with 0.1 ratiosand 3 betweenwt.% (Figure 0.1 and 1c). 3 wt.%Different (Figure fluorescent1c). Di ff erentdyes fluorescentwere used dyesfor better were visualization used for better of visualization the printed ofMMSL the printed structures. MMSL Initial structures. experiments Initial used experiments an europium-based used an europium-based organic complex organic ((Tris complex(1,3-diphenyl- ((Tris (1,3-diphenyl-1,3-propanedionato)1,3-propanedionato) (1,10-phenanthroline)europium(III) (1,10-phenanthroline)europium(III) (TCI, Frankfurt, (TCI, Germany), Frankfurt, Germany),which is whichresponsive is responsive to near-UV to near-UV radiation. radiation. Subsequent Subsequent experiments experiments used used two two commercially commercially available fluorescentfluorescent dyes, the perylene-based Lumogen F305 (red) and the naphthalimide-based V570 (violet) (BASF, Ludwigshafen,Ludwigshafen, Germany),Germany), both at a concentrationconcentration of 0.05 wt.%wt.% inin thethe resinresin matrix.matrix. All resin mixtures were prepared using the Ultra-Turrax T-10 disperserdisperser (IKA,(IKA, Staufen, Germany) by mixing at 15,000–18,00015000–18000 rpm rpm for for 3–5 3–5 min min under under ambient ambient conditions. conditions.

(a) (b) (c)

Figure 1. Photocurable resin composition: ( a) EGDMA; ( b) BAEDA; ( c) TPO.

The viscositiesviscosities of of the the EGDMA-BAEDA EGDMA-BAEDA mixtures mixtur werees measuredwere measured using ausing cone anda cone plate rheometerand plate (CVOrheometer 50, Bohlin(CVO /50,Malvern Bohlin/Malvern Instruments, Instruments, Herrenberg, Herrenberg, Germany) Germany) at shear ratesat shear between rates between 10 and 500 10 1and/s and 500 in1/s the and temperature in the temperature range between range between 20–50 ◦20–50C. The °C. resin The suitabilityresin suitability for SLA for 3DSLA printing 3D printing was evaluatedwas evaluated using using the commercially the commercially available available B9Creator B9Creator SLA printer SLA printer (B9Creations (B9Creations LLC, Rapid LLC, City, Rapid IL, USA),City, IL, which USA), uses which DLP uses technology DLP technology and is capable and is of capable voxel resolutions of voxel resolutions down to 30 downµm. Ato post-print30 µm. A floodpost-print exposure flood was exposure performed was performed on all samples on all using samples the 600using mW the/cm 6002 Hönle mW/cm LED-Spot-1002 Hönle LED-Spot-100 UV lamp (Dr.UV lamp Hönle (Dr. AG, Hönle Gräfelfing, AG, Gräfelfing, Germany) atGermany) 365 nm forat 365 120 s.nm Tensile for 120 tests s. Tensile (five samples) tests (five according samples) to ASTMaccording D-638 to ASTM were performed D-638 were applying performed a Zwick applying/Roell a Z010 Zwick/Roell universal Z010 testing universal machine testing (Zwick machine/Roell, Ulm,(Zwick/Roell, Germany) Ulm, under Germany) ambient under conditions ambient (2.5 conditions kN load cell, (2.5 pull kN speeds load cell, of 2 pull mm /speedsmin and of 52mm mm/min/min). Inand the 5 casemm/min). of the In systematic the case investigationof the systematic of BE-5050, investigation doped of with BE-5050, the Lumogen doped with chromophors, the Lumogen five sampleschromophors, were investigatedfive samples for were each investigated selected layer for polymerizationeach selected layer time. polymerization The size of the time. specimen The wassize reducedof the specimen by a factor was of reduced 0.4 due by to vata factor volume of 0.4 restrictions. due to vat volume restrictions.

3. MMSL Printing Device 3.1. Design and Construction 3.1. Design and Construction Following the restrictions of earlier approaches dealing with multi-material SLA devices and Following the restrictions of earlier approaches dealing with multi-material SLA devices and potential consumer wish lists, the following set of requirements for the new device was defined: potential consumer wish lists, the following set of requirements for the new device was defined: Simple and low-cost construction and setup. • • Simple and low-cost construction and setup. • Integration of building blocks, like a DLP light source, from established commercial systems. • No material exchange during the printing procedure.

Micromachines 2020, 11, 532 4 of 17

Integration of building blocks, like a DLP light source, from established commercial systems. Micromachines• 2020, 11, x FOR PEER REVIEW 4 of 17 No material exchange during the printing procedure. • Realization• Realization of structural of structural features features below below 100 µ100m withµm with good good reproducibility. reproducibility. • • Simple, Simple, adaptable adaptable control control software. software. • ® Figure2 2 shows shows thethe initialinitial designdesign of of the the MMSL MMSL device device using using SolidWorks SolidWorks®. The The device’s housing was constructed from aluminum struts (Bosch Rexrot Rexroth,h, Stuttgart, Germany), with a milled aluminum sheet forming the bottom of the printing, an illumination chamber with openings for the attached projector and mounting points for the linear stages.stages. Transparent polymethylmethacrylate (PMMA) sheets, covered covered with with Kapton Kapton®® foil,foil, were were used used for for allall sidewalls, sidewalls, as aswell well as for as forthe thefront front door door of the of chamber.the chamber. An opaque An opaque yellow yellow PMMA PMMA sheet sheet was wasused used as the as chamber the chamber roof roofto re toduce reduce ambient ambient light. light. An An80 mm 80 mm fan fanhelped helped to toremove remove the the heat heat generated generated by by the the projector projector out out of of the the chamber. Individual stepper motors were applied for the three linear x-, y- and z-stages for the movement of the vats and the build-platform,build-platform, with with additional additional mechanical mechanical homing homing switches switches attached attached to each to stage. each Thesestage. motors,These motors,stages and stages switches and switches were reused were from reused an out-of-action from an out-of-action FFF printer FFF (MakerBot printer (MakerBot 2X, MakerBot 2X, Industries,MakerBot Industries,New York City,New NY, York USA) City, and NY, modified USA) and for modified our purposes. for our purposes.

FigureFigure 2.2. A SolidWorks®® -designed-designed 3D 3D model model of of the the MMSL MMSL device. device.

The laser cutting of PMMA sheets allowed for thethe prefabrication of the individual vat parts (Figure 33a),a), whichwhich werewere assembledassembled byby gluinggluing toto thethe finalfinal vatvat (Figure(Figure3 b).3b). The The vats vats were were designed designed to to 2 have an inner inner area area of of 80 80 × 9595 mm² mm andand were were designed designed to towork work with with small small resin resin volumes volumes of around of around 30 × mL.30 mL. They They were were equipped equipped with with a glass a glass illumination illumination window, window, covering covering around around 50% 50% of of the the vat’s vat’s inner ® surface area area (Figure (Figure 3b).3b). The The vat vat window window was was va varnishedrnished with with a non-stick a non-stick silicone silicone layer layer (Sylgard (Sylgard® 184 184optical-grade, optical-grade, Dow Dow Corning, Corning, Midland, Midland, MI, USA) MI, USA) to provide to provide a stable, a stable, optically optically transparent, transparent, yet soft yet ® innersoft inner surface. surface. For this, For this,around around 20 g 20of gliquid of liquid Sylgard Sylgard® 184 was184 evacuated was evacuated before before being being poured poured onto theonto vat the window vat window and cured and at cured 60 °C at for 60 up◦C to for 4 h. up All to vats 4 h. were All vatsfixed were on a fixedPMMA on platform a PMMA and platform placed 2 belowand placed the aluminum below the build-platform aluminum build-platform with a build-s withurface a build-surface of 64 × 36 mm² of 64 (Figure36 mm 3c). (FigureThe build-3c). × Theplatform build-platform assembly assemblywas equipped was equipped with four with tilt-adj fourustment tilt-adjustment points to points help tothe help build-surface the build-surface to be calibratedto be calibrated parallel parallel to the to vat the window. vat window. The Thebuild-platform build-platform was was connected connected to the to thez-axis z-axis arm arm by by a thumb-screw,a thumb-screw, which which in inturn turn was was mounted mounted on to on ato stepper-driven a stepper-driven 280 mm 280 mmlong longacme acme lead-screw lead-screw with witha 2 mm a 2 pitch. mm pitch. A modified A modified version version of the of Vivitek the Vivitek D912HD D912HD DLP DLPprojector projector (Vivitek, (Vivitek, Hoofddorp, Hoofddorp, The TheNetherlands), Netherlands), in its in 30 its µm 30 µ pixelm pixel mode, mode, and and with with a build a build area area of of 57.6 57.6 × 32.432.4 mm², mm2 ,was was used used as as an × illumination source. Due Due to to filters filters filtering filtering UV-B UV-B and UV-C radiation,radiation, thethe exploitableexploitable wavelengthswavelengths for photopolymerization are in the range between 375–405 nm. Figure 4 shows the mounted MMSL device after completion. The maximum size of printable parts is 57.6 × 32.4 × 50 mm³.

Micromachines 2020, 11, 532 5 of 17 for photopolymerization are in the range between 375–405 nm. Figure4 shows the mounted MMSL deviceMicromachines after 2020 completion., 11, x FOR PEER The maximumREVIEW size of printable parts is 57.6 32.4 50 mm3. 5 of 17 Micromachines 2020, 11, x FOR PEER REVIEW × × 5 of 17

(a) (b) (c)

Figure 3. MMSL resin vat: ( a)) Laser Laser cut PMMA sheets; ( b) Assembled vat; (c) Build platform.

Figure 4. Ready to use MMSL device setup. 3.2. Device Control 3.2. Device Control As mentioned above, three NEMA-17 stepper motors, reused from a dismantled FFF printer, As mentioned above, three NEMA-17 stepper motors, reused from a dismantled FFF printer, were connected to the printer stages and controlled by an Arduino UNO microcontroller mounted were connected to the printer stages and controlled by an Arduino UNO microcontroller mounted with a CNC shield overlay. The latter allows for up to four stepper motor drivers, three of which with a CNC shield overlay. The latter allows for up to four stepper motor drivers, three of which were connected to the stepper motors using Polulu stepper motor carriers (Polulu Corporation, Las were connected to the stepper motors using Polulu stepper motor carriers (Polulu Corporation, Las Vegas, NV, USA), each equipped with the DRV8825 high current stepper driver. The printer design Vegas, NV, USA), each equipped with the DRV8825 high current stepper driver. The printer design required three unique positions on the x-axis for vat selection and two on the y-axis for illumination required three unique positions on the x-axis for vat selection and two on the y-axis for illumination and layer delamination, while simultaneously establishing and breaking the line-of-sight between and layer delamination, while simultaneously establishing and breaking the line-of-sight between the projector and build-platform. These motors worked in the 1/8th microstepping mode to enable the projector and build-platform. These motors worked in the 1/8th microstepping mode to enable smooth operation, in contrast, the z-stage was configured at 1/16th microstep enabling small layer smooth operation, in contrast, the z-stage was configured at 1/16th microstep enabling small layer thicknesses from 100 µm down to 10 µm. The Arduino was connected to a desktop PC via USB and thicknesses from 100 µm down to 10 µm. The Arduino was connected to a desktop PC via USB and controlled by LabVIEW software (National Instruments, Austin, TX, USA) applying the Arduino controlled by LabVIEW software (National Instruments, Austin, TX, USA) applying the Arduino control interface, the same is valid for the projector using a serial RS-232 connection. control interface, the same is valid for the projector using a serial RS-232 connection. 3.3. Software: GUI and Printing Programs 3.3. Software: GUI and Printing Programs The homemade MMSL control software allows the import of slice files in the established JPG, PNG The homemade MMSL control software allows the import of slice files in the established JPG, or BMPThe formats. homemade These MMSL slice files control were software generated allows using the the import B9Creator of slice software, files whichin the wasestablished additionally JPG, PNG or BMP formats. These slice files were generated using the B9Creator software, which was usedPNG toor position, BMP formats. scale and These orient slice the files 3D models.were generated The MMSL using software the B9Creator is capable software, of working which directly was additionally used to position, scale and orient the 3D models. The MMSL software is capable of withadditionally this slice used data, to thereby position, reducing scale and the eorientffort forthe additional 3D models. GUI The programming. MMSL software Allrelevant is capable print of working directly with this slice data, thereby reducing the effort for additional GUI programming. parameters,working directly like illumination with this slice time, data, motor thereby movement reducing and the delays effort before for additional and after GUI exposure programming. could be All relevant print parameters, like illumination time, motor movement and delays before and after adjustedAll relevant in theprint software. parameters, The softwarelike illumination was programmed time, motor in movement a top-down and hierarchical delays before manner and using after exposure could be adjusted in the software. The software was programmed in a top-down hierarchical manner using LabVIEW state machines. The main user interface called subroutines for calibration sequences, printer functionality tests or the printing programs, allowing the user to define individual process parameters, like moving a motor or switching the projector on or off. When started, the program first undergoes a self-test by establishing and verifying communication with the

Micromachines 2020, 11, 532 6 of 17

LabVIEW state machines. The main user interface called subroutines for calibration sequences, printer functionality tests or the printing programs, allowing the user to define individual process parameters, likeMicromachines moving 2020 a motor, 11, x FOR or switching PEER REVIEW the projector on or off. When started, the program first undergoes6 of 17 a self-test by establishing and verifying communication with the Arduino and the projector, after whichArduino subroutines and the projector, to calibrate after the which build-platform subroutines and to calibrate projector, the as build-pla well as thosetform to and test projector, individual as motorwell as functions,those to test can individual be called. motor Figure functions,5 shows ca exemplarilyn be called. the Figure printer 5 shows functions exemplarily screen includingthe printer motorfunctions control. screen The including MMSL software motor control. includes The two MMSL calibration software programs, includes one two each calibration for the projector programs, and theone build-platform.each for the projector Both programsand the build-platform. follow a checklist-style Both programs interface follow to minimizea checklist-style unwanted interface motor to movements.minimize unwanted The projector motor calibration movement sequences. The projector first homes calibration all three sequence axes sequentially first homes and all switches three axes on thesequentially projector toand display switches a calibration on the projector grid. Coarse to display and fine a focalcalibration adjustments grid. Coarse can be manuallyand fine donefocal byadjustments positioning can the be twomanually focus-rings done by on positioning the projector. the Thetwo focus-rings sharpness of on the the calibration projector. The grid sharpness can then beof the viewed calibration on diff erentgrid can vats. then The be build-platform viewed on different calibration vats. programThe build-plat allowsform the build-platformcalibration program to be adjustedallows the parallel build-platform to the vat to window. be adjusted parallel to the vat window.

Figure 5. Printer functions screen with motor control.

The MMSLMMSL softwaresoftware has has three three sub-programs sub-programs each each for for single-material single-material and and multi-material multi-material print-jobs. print- Thejobs. first The is first used isto used select to andselect adjust and theadjust print the parameters, print parameters, followed followed by a pre-print by a pre-print checklist checklist and the print and programthe print displayingprogram displaying the status ofthe the status print-job of the being prin performed.t-job being performed. Both print-setup Both programsprint-setup first programs prompt thefirst user prompt to point the touser a folder to point with to the a folder slice files with in the JPG, slice PNG files or BMPin JPG, format PNG and or theBMP appropriate format and print the parameters.appropriate Theprint pre-print parameters. checklist The pre-print prompts checklis the usert toprompts sequentially the user position to sequentially the motors, position switch-on the themotors, projector switch-on and fill the the projector vats with and resin. fill the After vats the with print-job resin. After has begun, the print-jo no furtherb has interventionbegun, no further from theintervention user is necessary. from the Multi-material user is necessary. prints Multi-material involve the use prints of up involve to three the vats use during of up ato single three print vats job.during The a programsingle print requires job. The a printprogram recipe requires in the a form print of recipe a CSV in file, the whereform of the a CSV vat, file, the exposurewhere the time vat, andthe exposure the start- time and end-layersand the start- for and each end-layers print step for are each defined. print Tostep minimize are defined. the cross-contaminationTo minimize the cross- of dicontaminationfferent resins, of a rinsedifferent step resins, can also a berinse integrated step can in also the be print integrated recipe.In inthis the case,print typically recipe. In the this middle case, vattypically is filled the with middle a suitable vat is filled organic with solvent, a suitable like isopropanol,organic solvent, which like is isopropanol, compatible withwhich the is PMMAcompatible vat. with Thethe PMMA MMSL devicevat. works in print-cycles, where the completion of a print-cycle corresponds to the photocuringThe MMSL ofdevice a layer works of a definedin print-cycles, thickness. where Figure the6 completionshows a flowchart of a print-cycle describing corresponds a print-cycle to forthe aphotocuring multi-material of a print. layer Theof a print-cycledefined thickness. begins withFigure the 6 loweringshows a flowchart of the build-platform describing a into print-cycle the vat, tofor a a depth multi-material equal to one print. layer-thickness The print-cycle (adjustable begins with between the lowering 10 µm and of100 the µbuild-platformm) less than that into of the vat, last layer.to a depth The programequal to thenone layer-thickness waits for the resin (adjustable to reflow inbetween between 10 the µm build-platform and 100 µm) less and than the vat that window. of the Thelast layer. current The slice program image isthen then waits exposed for the for resin a user-defined to reflow in period, between polymerizing the build-platform the resin and layer. the Forvat thewindow. first layer, The acurrent longer slice exposure image time is then should exposed be used for to a ensureuser-defined the adhesion period, of polymerizing the solidified the material resin onlayer. the For build-platform. the first layer, After a longer one illuminationexposure time is shou finished,ld be the used vat-platform to ensure the moves adhesion along of the the y-direction solidified awaymaterial from on thethe vatbuild-platform. window (axis After orientations one illuminati definedon is in finished, Figure2 ).the This vat-platform lateral movement moves along step the is importanty-direction to away protect from the the silicone vat wind onow the (axis vat window orientations from defined damage, in asFigure a direct 2). This upward lateral movement movement of step is important to protect the silicone on the vat window from damage, as a direct upward movement of the build-platform would result in an upward suction and possible delamination of the silicone from the vat’s body. After decoupling with the silicone, the build-platform can move upwards by a defined amount, followed by a y-stage movement to re-establish the line-of-sight between the build-platform, the vat window and the projector, enabling the next exposure cycle. In the case of multi-material prints, the print-cycle algorithm additionally checks if the previous layer

Micromachines 2020, 11, 532 7 of 17 the build-platform would result in an upward suction and possible delamination of the silicone from the vat’s body. After decoupling with the silicone, the build-platform can move upwards by a defined amount,Micromachines followed 2020, 11 by, x FOR a y-stage PEER REVIEW movement to re-establish the line-of-sight between the build-platform,7 of 17 the vat window and the projector, enabling the next exposure cycle. In the case of multi-material prints, thewas print-cycle the last one algorithm for the current additionally print step checks and if can the previoushome the layer build-platform, was the last moving one for theit out current of the print way stepfor a and change can home of vats the or build-platform, a rinse cycle.moving After all it outlayers of theare way finished, for achange the x-, ofy-, vats and or z-motors a rinse cycle. are Aftersequentially all layers homed, are finished, the projector the x-,y-, is andturned z-motors off and are the sequentially build-platform homed, can the be projector manually is turneddetached; off andfinally the the build-platform printed structure can be manuallycan then detached;be manual finallyly removed the printed from structurethe build-platform can then be and manually post- removedprocessed from if necessary. the build-platform and post-processed if necessary.

FigureFigure 6.6. Print-cycle flowchartflowchart of thethe MMSLMMSL device.device. 4. Resin Development and Characterization 4. Resin Development and Characterization Previous work has shown that resins developed in our research group are compatible with Previous work has shown that resins developed in our research group are compatible with the the B9Creator and its illumination source [37]. Building up on this, experiments showed that B9Creator and its illumination source [37]. Building up on this, experiments showed that a new base a new base resin comprising BAEDA and EGDMA as comonomers, and in combination with TPO resin comprising BAEDA and EGDMA as comonomers, and in combination with TPO as a as a photoinitiator, had the best performance in terms of print quality with the B9Creator. For photoinitiator, had the best performance in terms of print quality with the B9Creator. For repeatable repeatable and reliable printing, the resin viscosity under ambient conditions (25 ◦C) should remain and reliable printing, the resin viscosity under ambient conditions (25 °C) should remain below below 0.2 Pas for the given setup without the use of an active surface scraper. The shear rate and 0.2 Pas for the given setup without the use of an active surface scraper. The shear rate and temperature-dependent viscosity of the BAEDA–EGDMA resins were measured by cone-and-plate temperature-dependent viscosity of the BAEDA–EGDMA resins were measured by cone-and-plate rheometry and are shown in Figure7. The notation describes the by-wt.%-ratio of the two monomers, rheometry and are shown in Figure 7. The notation describes the by-wt.%-ratio of the two monomers, i.e., BE-5050 means 50 wt.% BAEDA and 50 wt.% EGDMA. The two monomers individually exhibit i.e., BE-5050 means 50 wt.% BAEDA and 50 wt.% EGDMA. The two monomers individually exhibit a simple Newtonian flow profile (Figure7a), with the expected decrease in viscosities with increasing a simple Newtonian flow profile (Figure 7a), with the expected decrease in viscosities with increasing temperature (Figure7b), in correlation with the Andrade equation [38]. temperature (Figure 7b), in correlation with the Andrade equation [38].

Figure 7. Viscosities of all investigated resin mixtures as function of (a) shear rate and (b) temperature.

Micromachines 2020, 11, x FOR PEER REVIEW 7 of 17

was the last one for the current print step and can home the build-platform, moving it out of the way for a change of vats or a rinse cycle. After all layers are finished, the x-, y-, and z-motors are sequentially homed, the projector is turned off and the build-platform can be manually detached; finally the printed structure can then be manually removed from the build-platform and post- processed if necessary.

Figure 6. Print-cycle flowchart of the MMSL device.

4. Resin Development and Characterization Previous work has shown that resins developed in our research group are compatible with the B9Creator and its illumination source [37]. Building up on this, experiments showed that a new base resin comprising BAEDA and EGDMA as comonomers, and in combination with TPO as a photoinitiator, had the best performance in terms of print quality with the B9Creator. For repeatable and reliable printing, the resin viscosity under ambient conditions (25 °C) should remain below 0.2 Pas for the given setup without the use of an active surface scraper. The shear rate and temperature-dependent viscosity of the BAEDA–EGDMA resins were measured by cone-and-plate rheometry and are shown in Figure 7. The notation describes the by-wt.%-ratio of the two monomers, i.e., BE-5050 means 50 wt.% BAEDA and 50 wt.% EGDMA. The two monomers individually exhibit a simple Newtonian flow profile (Figure 7a), with the expected decrease in viscosities with increasing Micromachines 2020, 11, 532 8 of 17 temperature (Figure 7b), in correlation with the Andrade equation [38].

Micromachines 2020, 11, x FOR PEER REVIEW 8 of 17 Figure 7. Viscosities of all investigated resin mixtures as function of (a) shear rate and (b) temperature. 5. Single-MaterialFigure 7. Viscosities Print of Mode all investigated Using the resin Base mixtures Resin as BE-5050 function of (a) shear rate and (b) temperature. 5. Single-Material Print Mode Using the Base Resin BE-5050 Prior to its application in the MMSL, the B9Creator was used to validate the printability of

BE-5050,Prior which to its was application selected as in the the base MMSL, resin the due B9Creator to its good was curing used behavior, to validate even the with printability the presence of ofBE-5050, oxygen. which To optimize was selected the asphotopolymerization the base resin due to behavior, its good curing the mechanical behavior, evenproperties with the of presenceprinted tensileof oxygen. test Tospecimens optimize the(five photopolymerization samples per material behavior, composition) the mechanical were measured properties as of printeda function tensile of photoinitiatortest specimens content, (five samples which per was material set to between composition) 0.1 wt.% were and measured 3.0 wt.% as a(Figure function 8). ofIncreasing photoinitiator TPO contentscontent, whichforce up was the set ultimate to between tensile 0.1 wt.%strength and (UTS) 3.0 wt.% significantly (Figure8). (Figure Increasing 8a), TPO whilst contents the maximum force up failure-strainthe ultimate tensilehas its strengthmaximum (UTS) at 1.0 significantly wt.%. TPO. (Figure Consequently,8a), whilst the the TPO maximum concentration failure-strain was set has to 1its wt.% maximum for further at 1.0 experiments. wt.%. TPO. Both Consequently, of the used the monomers TPO concentration possess two was cu setrable to function 1 wt.% foral groups, further henceexperiments. after curing, Both ofa thermoset the used monomers is synthesized. possess Incr twoeasing curable the functionalphotoinitiator groups, content hence causes after a curing, rapid formationa thermoset of is a synthesized. highly crosslinked Increasing system the photoinitiatorwith increasing content ultimate causes tensile a rapid strength formation and ofa reduced a highly maximumcrosslinked strain. system with increasing ultimate tensile strength and a reduced maximum strain.

Figure 8. ImpactImpact of of the the photoinitiator photoinitiator content content on on the the me mechanicalchanical properties properties in BE-5050 in BE-5050 base base resin. resin. (a) Ultimate(a) Ultimate tensile tensile strength strength UTS; UTS; (b) (Maximumb) Maximum strain strain at break. at break.

The usageusage ofof fluorescent fluorescent dyes, dyes, dissolved dissolved in the in basethe resin,base resin, helps tohelps visualize to visualize the normally the normally colorless colorlessand transparent and transparent printed parts. printed The parts. demonstrators The demonstrators shown in shown Figure in9 wereFigure printed 9 were using printed the using BE-5050 the BE-5050resin containing resin containing 5 wt.% of 5 thewt.% fluorescent of the fluorescen europiumt europium (III) complex. (III) complex. The B9Creator The B9Creator was used was to printused tothe print Y-bend the Y-bend (Figure (Figures9a,b), adapted 9a,b), adapted from an from optical an optical application application (beam (beam splitter), splitter), at a at resolution a resolution of 50 50 50 µm3 to investigate the polymerizability of the fluorescent resin. Figure9c,d show the logo of 50× × 50× × 50 µm³ to investigate the polymerizability of the fluorescent resin. Figures 9c and d show theof the logo microsystem of the microsystem engineering engineering department department at University at University of Freiburg, of Freiburg, printed printed by the MMSLby the MMSL device. device. Due to the absorption characteristics of the material, longer exposure times (30 s per layer, compared to 3 to 5 s for the unfilled resins) were required for sufficient polymerization. Using the MMSL device for the first time in the single-material mode at a voxel size of 30 × 30 × 30 µm³, the per- layer exposure time was reduced to 12 s due to the close proximity of the projector to the vat. Figures 9a and 9c show the test structures under ambient light and Figures 9b and 9d under near UV-light.

Figure 9. Single-material print with fluorescent dye-doped BE-5050. Printed with B9Creator: (a) Under ambient light; (b) Under a near UV LED. Printed with the MMSL device: (c) Under ambient light; (d) Under a near UV LED.

Micromachines 2020, 11, x FOR PEER REVIEW 8 of 17

5. Single-Material Print Mode Using the Base Resin BE-5050 Prior to its application in the MMSL, the B9Creator was used to validate the printability of BE-5050, which was selected as the base resin due to its good curing behavior, even with the presence of oxygen. To optimize the photopolymerization behavior, the mechanical properties of printed tensile test specimens (five samples per material composition) were measured as a function of photoinitiator content, which was set to between 0.1 wt.% and 3.0 wt.% (Figure 8). Increasing TPO contents force up the ultimate tensile strength (UTS) significantly (Figure 8a), whilst the maximum failure-strain has its maximum at 1.0 wt.%. TPO. Consequently, the TPO concentration was set to 1 wt.% for further experiments. Both of the used monomers possess two curable functional groups, hence after curing, a thermoset is synthesized. Increasing the photoinitiator content causes a rapid formation of a highly crosslinked system with increasing ultimate tensile strength and a reduced maximum strain.

Figure 8. Impact of the photoinitiator content on the mechanical properties in BE-5050 base resin. (a) Ultimate tensile strength UTS; (b) Maximum strain at break.

The usage of fluorescent dyes, dissolved in the base resin, helps to visualize the normally colorless and transparent printed parts. The demonstrators shown in Figure 9 were printed using the BE-5050 resin containing 5 wt.% of the fluorescent europium (III) complex. The B9Creator was used Micromachinesto print the 2020Y-bend, 11, 532 (Figures 9a,b), adapted from an optical application (beam splitter), at a resolution9 of 17 of 50 × 50 × 50 µm³ to investigate the polymerizability of the fluorescent resin. Figures 9c and d show the logo of the microsystem engineering department at University of Freiburg, printed by the MMSL Duedevice. to theDue absorption to the absorption characteristics characteristics of the material, of the material, longer exposure longer exposure times (30 stimes per layer, (30 s comparedper layer, tocompared 3 to 5 s for to the3 to unfilled 5 s for resins)the unfilled were requiredresins) were for surequiredfficient polymerization.for sufficient polymerization. Using the MMSL Using device the for the first time in the single-material mode at a voxel size of 30 30 30 µm3, the per-layer exposure MMSL device for the first time in the single-material mode at a voxel× × size of 30 × 30 × 30 µm³, the per- timelayer was exposure reduced time to 12was s duereduced to the to close 12 s proximitydue to the ofclose the proximity projector to of the the vat. projector Figure to9a,c the show vat. theFigures test structures9a and 9c show under the ambient test structures light and under Figure ambient9b,d under light near and UV-light. Figures 9b and 9d under near UV-light.

Figure 9.9. Single-materialSingle-material print print with with fluorescent fluorescent dye-doped dye-doped BE-5050. BE-5050. Printed Printed with B9Creator:with B9Creator: (a) Under (a) ambientUnder ambient light; ( blight;) Under (b) Under a near a UV near LED. UV Printed LED. Printed with the with MMSL the MMSL device: device: (c) Under (c) Under ambient ambient light; Micromachines(light;d) Under (d 2020) Under a, near11, x a UVFOR near LED. PEER UV LED.REVIEW 9 of 17

6. Dual Dual Material Material Print Print Mode Mode Applying Applying the the MMSL Device

6.1. Feasibility Feasibility Study Study The first first feasibility studies of the new MMSL device device used, used, as as described described in in the the previous previous section, section, fluorescentfluorescent dyes as markers for the different different appliedapplied resins. Figure 10 shows test patterns on a base plate. The The base base plate plate was made from pure BE-505 BE-50500 and the test patterns used the same resin doped with the fluorescentfluorescent dyesdyes Lumogen Lumogen V705 V705 (Figure (Figures 10 a,c)10a,c) and and F305 F305 (Figure (Figures 10b,d). 10b,d). Especially Especially under under near nearUV-radiation, UV-radiation, it is clearly it is clearly visible visible that onlythat theonly test the patterns test patterns on top on of top the of base the platebase plate were were doped doped with withthe fluorescent the fluorescent dye. dye. The The MMSL MMSL device device was was shown shown to workto work reliably reliably with with two two di ffdifferenterent materials materials at a minimum lateral resolution of 30 30 µm2, with freestanding structures down to 400 400 µm2 with at a minimum lateral resolution of× 30 × 30 µm², with freestanding structures down to× 400 × 400 µm² withan aspect an aspect ratio ratio of 1 to of 5. 1 to 5.

Figure 10. Microstructure demonstrators printed with the MMSL device, using a BE-8020 substrate with BE-5050 containing a fluorescent fluorescent composite of the Lumogen V570 ( a,c) and Lumogen F305 dyes (b,d); ( a,b) Under ambient light; (c,d) Under near UV light.light.

6.2. Systematic Investigations 6.2. Systematic Investigations To systematically investigateinvestigate the the print print parameters, parameters, precision precision and and accuracy, accuracy, a benchmark a benchmark artifact artifact for forthe evaluationthe evaluation of the of achievable the achievable MMSL MMSL geometric geometric and structural and structural accuracy accuracy performance performance was designed, was designed,following following recommendations recommendations from the literaturefrom the lite (Figurerature 11 (Figure)[39,40 11)]. The[39,40]. shown The benchmarkshown benchmark artifact contains structural features with a height of 2.5 mm on a 40 2 25 mm3 base plate. The geometric artifact contains structural features with a height of 2.5 mm× on× a 40 × 2 × 25 mm³ base plate. The features of the cuboids range from 100 µm up to 2 mm, the columns have sizes from 100 100 µm2 up geometric features of the cuboids range from 100 µm up to 2 mm, the columns have sizes× from 100 × to 2 2 mm2 and the trenches have a depth of 100 µm. The central star-like structure possesses an 100 µm²× up to 2 × 2 mm² and the trenches have a depth of 100 µm. The central star-like structure possessesouter diameter an outer of 12diameter mm and of an12 innermm and diameter an inner of diameter 1 mm. The of 1 small mm. lettersThe small in Figure letters 11inb Figure describes 11b describesthe measuring the measuring points for points sample for height sample measurements. height measurements.

Figure 11. Benchmark artifact for the validation of the MMSL print quality; (a) Rendered 3D overview; (b) 2D projection with measuring points for structural feature heights.

6.2.1. Single-Material Print With ensure a reproducible printing quality, the printing behavior of the three different above- mentioned resins (BE-5050, BE-5050 doped with Lumogen F305 or Lumogen V570) were investigated systematically. The layer thickness was set to 50 µm; the curing time was varied between 1–4 s. All samples were post-processed after printing according to the listed parameters (Table 1).

Micromachines 2020, 11, x FOR PEER REVIEW 9 of 17

6. Dual Material Print Mode Applying the MMSL Device

6.1. Feasibility Study The first feasibility studies of the new MMSL device used, as described in the previous section, fluorescent dyes as markers for the different applied resins. Figure 10 shows test patterns on a base plate. The base plate was made from pure BE-5050 and the test patterns used the same resin doped with the fluorescent dyes Lumogen V705 (Figures 10a,c) and F305 (Figures 10b,d). Especially under near UV-radiation, it is clearly visible that only the test patterns on top of the base plate were doped with the fluorescent dye. The MMSL device was shown to work reliably with two different materials at a minimum lateral resolution of 30 × 30 µm², with freestanding structures down to 400 × 400 µm² with an aspect ratio of 1 to 5.

Figure 10. Microstructure demonstrators printed with the MMSL device, using a BE-8020 substrate with BE-5050 containing a fluorescent composite of the Lumogen V570 (a,c) and Lumogen F305 dyes (b,d); (a,b) Under ambient light; (c,d) Under near UV light.

6.2. Systematic Investigations To systematically investigate the print parameters, precision and accuracy, a benchmark artifact for the evaluation of the achievable MMSL geometric and structural accuracy performance was designed, following recommendations from the literature (Figure 11) [39,40]. The shown benchmark artifact contains structural features with a height of 2.5 mm on a 40 × 2 × 25 mm³ base plate. The geometric features of the cuboids range from 100 µm up to 2 mm, the columns have sizes from 100 × 100 µm² up to 2 × 2 mm² and the trenches have a depth of 100 µm. The central star-like structure Micromachinespossesses an2020 outer, 11, 532diameter of 12 mm and an inner diameter of 1 mm. The small letters in Figure10 of11b 17 describes the measuring points for sample height measurements.

Figure 11. Benchmark artifact for the validation of the MMSL print quality; quality; ( a)) Rendered Rendered 3D 3D overview; overview; ((bb)) 2D2D projectionprojection withwith measuringmeasuring pointspoints forfor structuralstructural featurefeature heights.heights. 6.2.1. Single-Material Print 6.2.1. Single-Material Print With ensure a reproducible printing quality, the printing behavior of the three different With ensure a reproducible printing quality, the printing behavior of the three different above- above-mentioned resins (BE-5050, BE-5050 doped with Lumogen F305 or Lumogen V570) were mentioned resins (BE-5050, BE-5050 doped with Lumogen F305 or Lumogen V570) were investigated investigated systematically. The layer thickness was set to 50 µm; the curing time was varied between systematically. The layer thickness was set to 50 µm; the curing time was varied between 1–4 s. All 1–4 s. All samples were post-processed after printing according to the listed parameters (Table1). samples were post-processed after printing according to the listed parameters (Table 1).

Table 1. Post processing parameters after MMSL printing.

Step Parameter Rinsing in ultrasonic bath 5 min, isopropanol Flood illumination wavelength 385 nm Flood illumination intensity 200 mW/cm2 Flood illumination time 100 s Flood illumination atmosphere N2

For the different compositions and the respective absorption behaviors of the dopants, a slight difference in the best layer curing time was found. The visual appearance and the edge sharpness of the printed structures was selected as the main quality criterion. In the case of BE-5050, the best structure quality could be achieved with a layer curing time of 2 s, in which case only the smallest 100 100 µm2 columns were not repeatably producible. The curing defects in the middle of the star-like × structure can be attributed to an insufficient monomer removal prior to the final flood illumination (Figure 12a). For BE-5050 doped with F305, a slightly longer curing time of 2.75 s was sufficient to form sharp edges (Figure 12b). As in the pure base resin BE-5050, the smallest columns showed poor structure quality. The addition of Lumogen V570 caused a significantly reduced print quality (Figure 12c), where all edges and corners were rounded, and polymeric residue could be found between the individual structures, even at the best curing time of 2 s, and feature sizes below 200 200 µm2 could not be realized. The non-unique curing × times can be attributed to the addition of the fluorescent dyes with their additional light absorption in the visible and near UV-range. When compared to the measured geometric features at the best curing layer time, all results remain close to the theoretical value (x-, y-directions) or stay constant (density), with exception of the average structure height, which differs significantly from the theoretical value of 4.5 mm (Table2). This may be attributed to the poor spindle positioning of the z-axis and its reduced positional accuracy, originating from the spindle drive taken from the Makerbot Replicator 2X. Beyond the described feature properties, it is noticeable that the substrates often show surface defects. Micro-cracks within SLA-printed specimens arise from internal tensions during the polymerization process. These can be optimized for each resin variant by varying the light intensity and exposure time during printing, as well as during the post-print flood illumination. The systematic optimization of each resin variant (with or without a fluorescent dye), in terms of print parameters, lies outside Micromachines 2020, 11, x FOR PEER REVIEW 10 of 17

Table 1. Post processing parameters after MMSL printing.

Step Parameter Rinsing in ultrasonic bath 5 min, isopropanol Flood illumination wavelength 385 nm Flood illumination intensity 200 mW/cm² Flood illumination time 100 s Flood illumination atmosphere N2

For the different compositions and the respective absorption behaviors of the dopants, a slight difference in the best layer curing time was found. The visual appearance and the edge sharpness of the printed structures was selected as the main quality criterion. In the case of BE-5050, the best structureMicromachines quality2020, 11 could, 532 be achieved with a layer curing time of 2 s, in which case only the smallest11 of 17 100 × 100 µm² columns were not repeatably producible. The curing defects in the middle of the star- like structure can be attributed to an insufficient monomer removal prior to the final flood illuminationthe scope of this (Figure manuscript, 12a). For which BE-5050 is a proof-of-conceptdoped with F305, for a multi-materialslightly longer printing. curing time The of same 2.75 is s valid was sufficientfor the usage to form of commercial sharp edges resins. (Figure 12b).

(a) (b)

(c)

Figure 12. Microscopic images of the test patterns’ best layer curing time: ( a) BE-5050; ( b) BE-5050 Lumogen F305; ( c) BE-5050 Lumogen V570.

Table 2. Sample features at best layer curing time estimated by visual inspection (mean values).

Item BE-5050 BE-5050 F305 BE-5050 V570 \ \ Polymerization time per layer (s) 2 2.75 2 Height (mm) 3.67 0.20 2.92 0.28 3.80 0.31 ± ± ± Density (g/cm3) 1.21 0.00 1.22 0.01 1.21 0.00 ± ± ± Length (mm) 38.90 39.14 39.35 Width (mm) 24.52 25.04 24.93 Smallest structural detail (µm) 100–200 100–200 400

To overcome the subjective evaluation of the structural appearance, mechanical testing should deliver reliable data for the selected best layer curing time. Table3 lists the resulting Young’s modulus and maximum tensile strength for the investigated curing times. Despite the significant experimental error, the curing time with the best structure quality delivers the highest maximum tensile strength Micromachines 2020, 11, 532 12 of 17

as well. In general, an increase in the curing time should enhance the mechanical properties, for instance, due to a better intermixture accompanied with an enhanced polymer chain entanglement. Unfortunately, a clear correlation between curing time and mechanical properties cannot be found here, which may be attributed to the layer-by-layer fabrication process generating defects, as is prominently known for FFF printing [31]. A closer look at the fracture images after tensile testing verifies this possible explanation for the observed data scattering (Figure 13). A lamellar arrangement can be seen, in particular for the BE-5050 and BE5050 V750 systems, which reduces the mechanical stability due \ the presence of voids.

Table 3. Mechanical properties of all investigated systems as function of the layer curing time.

Item BE-5050 BE-5050 F305 BE-5050 V570 \ \ Curing time (s) 1 2 3 2 2.75 3.5 1 2 2.5 Young’s module 236 474 459 614 578 781 261 348 304 (MPa) 124 78 53 193 152 144 113 61 51 ± ± ± ± ± ± ± ± ± Max. tensile 16 30 30 26 40 35 26 36 24 strength (MPa) 4 11 5 13 12 14 5 7 8 Micromachines 2020, 11, x FOR PEER± REVIEW± ± ± ± ± ± ± ± 12 of 17

(a) (b) (c)

FigureFigure 13. 13.Fracture Fracture images images after after tensile tensile testing. testing. (a ()a BE-5050;) BE-5050; (b (b) BE-5050) BE-5050\F305;F305; (b ()b BE-5050) BE-5050\V570.V570. \ \ 6.2.2.6.2.2. Multi-Material Multi-Material Print Print TheThe same same artifact artifact structure, structure, as shown as shown in Figure in Figure 11, is taken11, is fortaken the evaluationfor the evaluation of the multi-material of the multi- print.material The print. following The stepsfollowing describe steps the describe multi-material the multi-material printing sequence: printing sequence: (a) Material positioning: The two outer vats (Figure4) are filled with the two materials to be printed a) Material positioning: The two outer vats (Figure 4) are filled with the two materials to be (base plate resin and feature resin), while the middle vat is filled with isopropanol for rinsing. printed (base plate resin and feature resin), while the middle vat is filled with isopropanol (b) The material intended to be the base plate is printed using the steps described for for rinsing. single-material printing. b) The material intended to be the base plate is printed using the steps described for single- (c) The base plate is rinsed in isopropanol before moving to the vat with the feature material. material printing. (d) Thec) featuresThe base are plate printed is rinsed using in single-material isopropanol before print moving parameters. to the vat with the feature material. (e) Thed) multi-materialThe features are print printed is delaminated using single-material from the build-platform print parameters. and post-processed, as described ine) Table The1. multi-material print is delaminated from the build-platform and post-processed, as It hasdescribed to be accentuated in Table that 1. only two different printing materials have been used, one for the base plate and a second one for the structural features. For the first attempt, BE-5050 was selected as the base It has to be accentuated that only two different printing materials have been used, one for the plate material, in combination with the two doped systems. In all cases, five identically processed base plate and a second one for the structural features. For the first attempt, BE-5050 was selected as samples were printed, validating the printing reproducibility. Figure 14 shows the central star-like the base plate material, in combination with the two doped systems. In all cases, five identically feature of five identical samples of BE-5050 – BE-5050 F305 (Figure 14a) and BE-5050 – BE5050 V570 processed samples were printed, validating the printing\ reproducibility. Figure 14 shows the\ central (Figure 14b), demonstrating the reproducibility. Apart from a little polymerized residue in between star-like feature of five identical samples of BE-5050 – BE-5050\F305 (Figure 14a) and BE-5050 – the structures on the bottom, which may attributed to poor monomer removal during rinsing and BE5050\V570 (Figure 14b), demonstrating the reproducibility. Apart from a little polymerized cleaning, microstructured details of 250–300 µm (BE-5050 – BE-5050 F305) and 250–400 µm (BE-5050 residue in between the structures on the bottom, which may attributed\ to poor monomer removal – BE-5050 V570) could be realized in a repeatable manner. Both combinations possess a smooth during rinsing\ and cleaning, microstructured details of 250–300 µm (BE-5050 – BE-5050\F305) and surface and sharp edges. Dimensional measurements on the printed features and specimen properties 250–400 µm (BE-5050 – BE-5050\V570) could be realized in a repeatable manner. Both combinations showed (Table4) similar results for both material combinations. As for single-material printing, possess a smooth surface and sharp edges. Dimensional measurements on the printed features and specimen properties showed (Table 4) similar results for both material combinations. As for single- material printing, the accuracy of the sample was fine, with the exception of the total structure height, which should be attributed to construction and calibration issues, especially to the positions of the different vats on the platform relative to each other. In all cases, a good adhesion of the second component on the first one was observed. The artifact was additionally printed in an inverted arrangement, i.e., with one of the doped resins as the base plate and the other doped resin, or pure BE-5050, as the feature material. These showed worse print quality, as shown in Figure 15. With BE-5050\F305 as the base plate, the use of pure BE-5050 led to large monomer residue between the structures (Figure 15a). A very poor adhesion of BE-5050\V570 on the base plate can be seen, if used as the second system (Figure 15b), although the quality of the structures is better. The use of BE-5050\V570 as the substrate delivered good, reliable structural features and sharp edges (Figures 15 c,d) and only small adhesion problems were observed. A closer look at the accuracy of the printed structures once again showed a good agreement with the x-, y-dimensions, and a significantly larger deviation of the nominal structure height (Table 5), which should be attributed to inaccuracies of the vat positioning in the z-direction. The smallest listed accessible structural feature was derived from the visual inspection of the structures regarding surface quality, edge sharpness and completeness.

Micromachines 2020, 11, 532 13 of 17 the accuracy of the sample was fine, with the exception of the total structure height, which should be attributed to construction and calibration issues, especially to the positions of the different vats on the platform relative to each other. In all cases, a good adhesion of the second component on the first oneMicromachines was observed. 2020, 11, x FOR PEER REVIEW 13 of 17

(a) (b)

Figure 14.14. MicroscopicMicroscopic images images of of the MMSL print appl applyingying BE-5050 as base plate material (five(five samples): ( a) BE-5050\F305 BE-5050 F305 asas featurefeature material;material; ((bb)) BE-5050BE-5050\V570V570as asfeature feature material. material. \ \ Table 4. Sample features of MMSL print with BE-5050 as base plate in combination with BE-5050 F305 Table 4. Sample features of MMSL print with BE-5050 as base plate in combination \with or BE-5050 V570 (average values). BE-5050\F305\ or BE-5050\V570 (average values). Item BE-5050 F305 BE-5050 V570 Item BE-5050\ \F305 BE-5050\V570\ Height (mm) 3.42 0.44 3.20 0.56 Height (mm) 3.42± ± 0.44 3.20 ± 0.56± Density (g/cm3) 1.20 0.00 1.20 0.01 Density (g/cm³) 1.20± ± 0.00 1.20 ± 0.01± Length (mm) 39.39 0.17 39.26 0.29 ± ± Width (mm)Length (mm) 24.91 39.390.07 ± 0.17 39.26 24.81 ± 0.290.04 ± ± Smallest structuralWidth detail (mm) (µm) 24.91 250 ± 0.07 24.81 ± 0.04 250 Smallest structural detail (µm) 250 250

The artifact was additionally printed in an inverted arrangement, i.e., with one of the doped resins as the base plate and the other doped resin, or pure BE-5050, as the feature material. These showed worse print quality, as shown in Figure 15. With BE-5050 F305 as the base plate, the use of \ pure BE-5050 led to large monomer residue between the structures (Figure 15a). A very poor adhesion of BE-5050 V570 on the base plate can be seen, if used as the second system (Figure 15b), although \ the quality of the structures is better. The use of BE-5050 V570 as the substrate delivered good, reliable \ structural features and sharp edges (Figure 15c,d) and only small adhesion problems were observed. (a) (b) (c) (d) A closer look at the accuracy of the printed structures once again showed a good agreement with the x-,Figure y-dimensions, 15. Representative and a significantlymicroscopic images larger of deviation different material of the nominal combinations: structure (a) BE-5050\F305 height (Table 5), whichas should base plate be attributedwith BE-5050 to inaccuracieson top; (b) BE-5050\F305 of the vat positioning as base plate inthe with z-direction. BE-5050\V570 The on smallest top; (c) listed accessibleBE-5050\V570 structural as featurebase plate was with derived BE-5050 from on top; the visual(d) BE-5050\V570 inspection as of base the structuresplate with BE-5050\F305 regarding surface quality,on top. edge sharpness and completeness.

Table 5. Sample features of MMSL print with different material combinations (average values).

Item Base: BE-5050\F305 BE-5050\F305 BE-5050\V570 BE-5050\V570 Top: BE-5050 BE-5050\V570 BE-5050 BE-5050\F305 Height (mm) - 2..84 ± 0.42 2.10 ± 0.24 3.56 ± 0.17 1.89 ± 0.52 Length (mm) - 39.13 ± 0.40 39.50 ± 0.26 39.11 ± 0.10 39.76 ± 0.30 Width (mm) - 25.02. ±0.42 25.13 ± 0.12 24.96 ± 0.29 25.47± 0.11 Smallest structural - 300 300 200 200 detail (µm)

Micromachines 2020, 11, x FOR PEER REVIEW 13 of 17

(a) (b)

Figure 14. Microscopic images of the MMSL print applying BE-5050 as base plate material (five samples): (a) BE-5050\F305 as feature material; (b) BE-5050\V570 as feature material.

Table 4. Sample features of MMSL print with BE-5050 as base plate in combination with BE-5050\F305 or BE-5050\V570 (average values).

Item BE-5050\F305 BE-5050\V570 Height (mm) 3.42 ± 0.44 3.20 ± 0.56 Density (g/cm³) 1.20 ± 0.00 1.20 ± 0.01 Length (mm) 39.39 ± 0.17 39.26 ± 0.29 Micromachines 2020, 11, 532 Width (mm) 24.91 ± 0.07 24.81 ± 0.04 14 of 17 Smallest structural detail (µm) 250 250

(a) (b) (c) (d)

Figure 15. Representative microscopic microscopic images images of of different different material material combinations: combinations: (a ()a BE-5050\F305) BE-5050 F305 \ as base plate with BE-5050 BE-5050 on on top; top; (b (b) )BE-5050\F305 BE-5050 F305 as as base base plate plate with with BE-5050\V570 BE-5050 V570 on on top; top; (c) ( c) \ \ BE-5050BE-5050\V570V570 as base plate with with BE-5050 BE-5050 on on top; top; (d (d) )BE-5050\V570 BE-5050 V570 as as base base plate plate with with BE-5050\F305 BE-5050 F305 \ \ \ on top.

Table 5. Sample features features of of MMSL MMSL print print with with different different material material combinations combinations (average (average values). values). BE-5050 F305 BE-5050 F305 BE-5050 V570 BE-5050 V570 ItemItem Base:Top:Base: BE-5050\F305 BE-5050\F305 BE-5050\V570 BE-5050\V570 BE-5050\ BE-5050\V570 BE-5050\ BE-5050\F305 Top: BE-5050 BE-5050\V570\ BE-5050 BE-5050\F305\ Height (mm) - 2.84 0.42 2.10 0.24 3.56 0.17 1.89 0.52 Height (mm) - 2..84 ± ±0.42 2.10 ± 0.24± 3.56 ± 0.17± 1.89 ± 0.52± Length (mm) - 39.13 0.40 39.50 0.26 39.11 0.10 39.76 0.30 Length (mm) - 39.13 ± 0.40± 39.50 ± 0.26± 39.11 ± 0.10± 39.76 ± 0.30± Width (mm) - 25.02 0.42 25.13 0.12 24.96 0.29 25.47 0.11 ± ± ± ± SmallestWidth structural (mm) - 25.02. ±0.42 25.13 ± 0.12 24.96 ± 0.29 25.47± 0.11 - 300 300 200 200 Smallest structuralµ detail ( m) - 300 300 200 200 Micromachinesdetail 2020 (µm), 11 , x FOR PEER REVIEW 14 of 17

6.2.3. Best Accessible Resolution and Smallest Feature Size Whilst FiguresFigures 12 12–15–15 give give an an overview overview about about the generalthe general printability printability and reproducibility and reproducibility of single- of andsingle- multi-material and multi-material prints, prints, a more a detailedmore detailed view isview necessary is necessary to evaluate to evaluate the structural the structural quality quality and accessibleand accessible print print resolution. resolution. Figure Figure 16 shows 16 shows the surfacethe surface quality quality of printed of printed columns columns and and cuboids cuboids for the single-material print (BE-5050 F305). The smallest measured lateral (xy) features are around 120 µm for the single-material print (BE-5050\F305).\ The smallest measured lateral (xy) features are around (Figure120 µm 16 (Figurea) and 16a) around and 90 aroundµm (Figure 90 µm 16 (Figureb). Unfortunately, 16b). Unfortunately, it was not it possible was not to possible print these to print features these at thisfeatures size inat athis reproducible size in a reproducible manner, hence manner, the smallest hence reproduciblethe smallest reproducible measured lateral measured (xy) features lateral were (xy) aroundfeatures 200 wereµm. around In all cases,200 µm. a good In all retention cases, a good of the rete edgesntion can of be the observed, edges can at thebe observed, top face, aat grid-like the top patternface, a originatinggrid-like pattern from the originating DMD mirrors from of thethe DLPDMD projector mirrors can of be the seen. DLP Concerning projector multi-materialcan be seen. printing,Concerning Figure multi-material 17 shows the printing, surface qualityFigure 17 of printedshows the columns surface and quality cuboids of forprinted the combination columns and of BE-5050 as the base and BE-5050 F305 as the feature material. The smallest measured features are cuboids for the combination of BE-5050\ as the base and BE-5050\F305 as the feature material. The aroundsmallest 250 measuredµm. features are around 250 µm.

(a) (b)

Figure 16.16. Single-material print: Micros Microscopiccopic images images of printed mi micro-featurescro-features with BE-5050BE-5050\F305F305 \ resin; ((aa)) columns;columns; ((bb)) cuboids.cuboids.

(a) (b)

Figure 17. Multi-material print: Microscopic images of printed micro-features with BE-5050 as base plate material and BE-5050\F305 as feature material; (a) columns; (b) cuboids.

6.3. Process Evaluation The aforementioned results show that the new MMSL equipment is able to combine two different curable resins in an acceptable manner. With respect to accuracy and reproducibility, two points have to be considered. The data shown in Tables 2, 4 and 5, for the geometric features and densities, are the mean values over five samples printed under identical conditions. The standard deviation for these results is acceptable. The reproducibility of the feature quality as a function of the printing parameters is not unique. In the case of the single-material printing surface quality, the presence of defects and rounded edges varies within one set of samples printed using identical

Micromachines 2020, 11, x FOR PEER REVIEW 14 of 17

6.2.3. Best Accessible Resolution and Smallest Feature Size Whilst Figures 12–15 give an overview about the general printability and reproducibility of single- and multi-material prints, a more detailed view is necessary to evaluate the structural quality and accessible print resolution. Figure 16 shows the surface quality of printed columns and cuboids for the single-material print (BE-5050\F305). The smallest measured lateral (xy) features are around 120 µm (Figure 16a) and around 90 µm (Figure 16b). Unfortunately, it was not possible to print these features at this size in a reproducible manner, hence the smallest reproducible measured lateral (xy) features were around 200 µm. In all cases, a good retention of the edges can be observed, at the top face, a grid-like pattern originating from the DMD mirrors of the DLP projector can be seen. Concerning multi-material printing, Figure 17 shows the surface quality of printed columns and cuboids for the combination of BE-5050 as the base and BE-5050\F305 as the feature material. The smallest measured features are around 250 µm.

(a) (b)

MicromachinesFigure 16.2020 Single-material, 11, 532 print: Microscopic images of printed micro-features with BE-5050\F30515 of 17 resin; (a) columns; (b) cuboids.

(a) (b)

FigureFigure 17. 17. Multi-materialMulti-material print: print: Microscopic Microscopic images images of of printed printed micro-features micro-features with with BE-5050 BE-5050 as as base base plateplate material material and and BE-5050\F305 BE-5050 F305 as as feature feature material; material; ( (aa)) columns; columns; ( (bb)) cuboids. cuboids. \ 6.3.6.3. Process Process Evaluation Evaluation TheThe aforementioned aforementioned resultsresults showshow that that the the new new MMSL MMSL equipment equipment is able is toable combine to combine two diff erenttwo differentcurable resinscurable in resins an acceptable in an acceptable manner. Withmanner. respect With to respect accuracy to andaccuracy reproducibility, and reproducibility, two points two have pointsto be considered.have to be considered. The data shown The data in Tables shown2,4 in and Tables5, for 2, the 4 geometricand 5, for featuresthe geometric and densities, features areand densities,the mean are values the overmean five values samples over printed five samples under identical printed conditions.under identical The standard conditions. deviation The standard for these deviationresults is for acceptable. these results The reproducibilityis acceptable. The of thereproducibility feature quality of asthe a feature function quality of the as printing a function parameters of the printingis not unique. parameters In the is case not of unique. the single-material In the case of printing the single-material surface quality, printing the presence surface of quality, defects the and presencerounded of edges defects varies and within rounded one edges set of varies samples wi printedthin one using set of identical samplesparameters, printed using especially identical in the case of the doped systems BE-5050 F305 and BE-5050 V570. With respect to accuracy and precision, \ \ especially in the case of structural details significantly below 500 µm, some construction-related improvements can be made. Firstly, the spindle used for the z-direction control has to be substituted by a more precise one, i.e., one with a lower pitch, so that the stepper motor can deliver the desired positional accuracy (see Section 3.2). Secondly, the mounting of the vats on the build-platform, and the build-platform calibration relative to the vat window, should be improved, enabling plane–parallel operation and a more precise deposition of the second material on the first one. The majority of observed print defects can be traced back to misalignment issues.

7. Conclusions and Outlook Within this work, a comprehensive approach for the realization of a low-cost multi-material stereolithography 3D printing device was presented, ranging from construction, software control, material development and characterization, as well as the first printing results with two different curable resins with similar chemistry. It was possible to print two-component test parts with the smallest structural features in the 200–300 µm range. The feasibility of low-cost multi-material printing using parts from defunct machines and inexpensive components was successfully demonstrated. If one wants to realize smaller structures in a better quality, accuracy and printing reliability, a significant process improvement is necessary. This can be achieved, e.g., by the assembly of high performance positioning elements. Looking ahead, the MMSL printer is a versatile device capable of being tuned to work with other resin systems and fabricating structures with enhanced mechanical and functional properties.

Author Contributions: Conceptualization, B.K., M.F. and T.H.; methodology, B.K. and A.R.-F.; software, B.K. and A.R.-F.; validation, B.K., M.-V.S. and M.F.; formal analysis, T.H.; investigation, B.K. and M.F.; resources, M.F.; data curation, B.K. and M.-V.S.; writing—original draft preparation, T.H.; writing—review and editing, T.H.; supervision, T.H. and M.F.; project administration, B.K.; funding acquisition, T.H. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Internationale Graduiertenakademie (IGA) at the University of Freiburg for the GenMik II research grant. The article processing charge was funded by the German Research Foundation (DFG) and the KIT in the funding program Open Access Publishing. Micromachines 2020, 11, 532 16 of 17

Acknowledgments: The authors gratefully acknowledge Aditya Bhat for the initial characterization of the resin system, as well as well as Jürgen Wilde at IMTEK for the granted access to the mechanical testing equipment. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Hull, C.H. Apparatus for Production of Three-Dimensional Objects by Stereolithography. U.S. Patent 4575330, 3 November 1886. 2. Gao, W.; Zhang, Y.; Ramanujan, D.; Ramani, K.; Chen, Y.; Williams, C.B.; Wang, C.C.L.; Shin, Y.C.; Zhang, S.; Zavattieri, P.D. The status, challenges, and future of additive manufacturing in engineering. Comput. Aided Des. 2015, 69, 65–89. [CrossRef] 3. Attaran, M. The rise of 3-D printing: The advantages of additive manufacturing over traditional manufacturing. Bus. Horiz. 2017, 60, 677–688. [CrossRef] 4. Ngo, T.D.; Kashani, A.; Imbalzano, G.; Nguyen, K.T.Q.; Hui, D. Additive manufacturing (3D printing): A review of materials, methods, applications and challenges. Compos. Part B Eng. 2018, 143, 172–196. [CrossRef] 5. Tofail, S.A.M.; Koumoulos, E.P.; Bandyopadhyay, A.; Bose, S.; O’Donoghue, L.; Charitidis, C. Additive manufacturing: Scientific and technological challenges, market uptake and opportunities. Mater. Today 2018, 21, 22–37. [CrossRef] 6. Singh, S.; Ramakrishna, S.; Singh, R. Material issues in additive manufacturing: A review. J. Manuf. Process. 2017, 25, 185–200. [CrossRef] 7. Oropallo, W.; Piegl, L.A. Ten challenges in 3D printing. Eng. Comput. 2016, 32, 135–148. [CrossRef] 8. Khatri, B.; Lappe, K.; Habedank, M.; Mueller, T.; Megnin, C.; Hanemann, T. Fused Deposition Modeling of ABS-Barium Titanate Composites: A Simple Route towards Tailored Dielectric Devices. Polymers 2018, 10, 666. [CrossRef] 9. Khatri, B.; Lappe, K.; Noetzel, D.; Pursche, K.; Hanemann, T. A 3D-Printable Polymer-Metal Soft-Magnetic Functional Composite-Development and Characterization. Materials 2018, 11, 189. [CrossRef] 10. Wei, X.; Liu, Y.; Zhao, D.; Mao, X.; Jiang, W.; Ge, S.S. Net-shaped barium and strontium ferrites by 3D printing with enhanced magnetic performance from milled powders. J. Magn. Magn. Mater. 2020, 493, 165664. [CrossRef] 11. Noetzel, D.; Eickhoff, R.; Hanemann, T. Fused Filament Fabrication of Small Ceramic Components. Materials 2018, 11, 1463. [CrossRef] 12. Gonzalez-Gutierrez, J.; Arbeiter, F.; Schlauf, T.; Kukla, C.; Holzer, C. Tensile properties of sintered 17-4PH stainless steel fabricated by material extrusion additive manufacturing. Mater. Lett. 2019, 248, 165–168. [CrossRef] 13. Gonzalez-Gutierrez, J.; Duretek, I.; Holzer, C.; Arbeiter, F.; Kukla, C. Filler Content and Properties of Highly Filled Filaments for Fused Filament Fabrication of Magnets. In Proceedings of the ANTEC Conference, Anaheim, CA, USA, 8–10 May 2017. 14. Abel, J.; Scheithauer, U.; Janics, T.; Hampel, S.; Cano, S.; Muller-Kohn, A.; Gunther, A.; Kukla, C.; Moritz, T. Fused Filament Fabrication (FFF) of Metal-Ceramic Components. J. Vis. Exp. (JoVE) 2019, 143, e57693. [CrossRef][PubMed] 15. Qiu, D.; Langrana, N.A. Void eliminating toolpath for extrusion-based multi-material layered manufacturing. Rapid Prototyp. J. 2002, 8, 38–45. [CrossRef] 16. Matsuzaki, R.; Kanatani, T.; Todoroki, A. Multi-material additive manufacturing of polymers and metals using fused filament fabrication and electroforming. Addit. Manuf. 2019, 29, 100812. [CrossRef] 17. Ambrosi, A.; Webster, R.D.; Pumera, M. Electrochemically driven multi-material 3D-printing. Appl. Mater. Today 2020, 18, 100530. [CrossRef] 18. Bandyopadhyay, A.; Heer, B. Additive manufacturing of multi-material structures. Mater. Sci. Eng. R Rep. 2018, 129, 1–16. [CrossRef] 19. LITHOZ. Available online: www.lithoz.com (accessed on 2 April 2020). 20. ADMATEC. Available online: www.admateceurope.com (accessed on 2 April 2020). Micromachines 2020, 11, 532 17 of 17

21. Inamdar, A.; Magana, M.; Medina, F.; Grajeda, Y.; Wicker, R.B. Development of an automated multiple material stereolithography machine. In Proceedings of the SFF Symposium Proceedings, Austin, TX, USA, 14–16 August 2006; pp. 624–635. 22. Choi, J.-W.; Kim, H.-C.; Wicker, R.B. Multi-material stereolithography. J. Mater. Process. Technol. 2011, 211, 318–328. [CrossRef] 23. Choi, J.-W.; MacDonald, E.; Wicker, R. Multi-material microstereolithography. Int. J. Adv. Manuf. Technol. 2010, 49, 543–551. [CrossRef] 24. Wicker, R.B.; MacDonald, E.W. Multi-material, multi-technology stereolithography. Virtual Phys. Prototyp. 2012, 7, 181–194. [CrossRef] 25. Hohnholz, A.; Obata, K.; Albrecht, D.; Koch, J.; Hohenhoff, G.; Suttmann, O.; Kaierle, S.; Overmeyer, L. Multimaterial bathless stereolithography using aerosol jet printing and UV laser based polymerization. J. Laser Appl. 2019, 31, 0022301. [CrossRef] 26. Roach, D.J.; Hamel, C.M.; Dunn, C.K.; Johnson, M.V.; Kuang, X.; Qi, H.J. The m4 3D printer: A multi-material multi-method additive manufacturing platform for future 3D printed structures. Addit. Manuf. 2019, 29, 100819. [CrossRef] 27. Zhou, C.; Chen, Y.; Yang, Z.; Koshnevis, B. In Proceedings of the Development of a Multi-Material Mask-Image-Projection-Based Stereolithography for the Fabrication of Digital Materials. Solid Freeform Fabrication Proceedings, Austin, TX, USA, 08–10 May 2011. pp. 65–80. Available online: http://utw10945. utweb.utexas.edu/Manuscripts (accessed on 18 May 2020). 28. Kim, Y.T.; Castro, K.; Bhattacharjee, N.; Folch, A. Digital Manufacturing of Selective Porous Barriers in Microchannels Using Multi-Material Stereolithography. Micromachines 2018, 9, 125. [CrossRef][PubMed] 29. Muguruza, A.; Bo, J.; Gómez, A.; Minguella-Canela, J.; Fernandes, J.; Ramos, F.; Xuriguera, E.; Varea, A.; Cirera, A. Development of a multi-material additive manufacturing process for electronic devices. Procedia Manuf. 2017, 13, 746–753. [CrossRef] 30. Pantelakis, S.; Lehmhus, D.; Busse, M.; von Hehl, A.; Jägle, E.; Koubias, S. State of the Art and Emerging Trends in Additive Manufacturing: From Multi-Material processes to 3D printed Electronics. MATEC Web Conf. 2018, 188, 03013. [CrossRef] 31. Hanemann, T.; Syperek, D.; Noetzel, D. 3D Printing of ABS Barium Ferrite Composites. Materials 2020, 13, 1481. [CrossRef] 32. Singh, R.; Kumar, R.; Farina, I.; Colangelo, F.; Feo, L.; Fraternali, F. Multi-Material Additive Manufacturing of Sustainable Innovative Materials and Structures. Polymers 2019, 11, 62. [CrossRef] 33. Magdassi, S. The Chemistry of Inkjet Inks; World Scientific Publishing: Singapore, 2010. 34. Graf, D.; Burchard, S.; Crespo, J.; Megnin, C.; Gutsch, S.; Zacharias, M.; Hanemann, T. Influence of Al2O3Nanoparticle Addition on a UV Cured Polyacrylate for 3D Inkjet Printing. Polymers 2019, 11, 633. [CrossRef] 35. Gleißner, U.; Hanemann, T. Tailoring the optical and rheological properties of an epoxy acrylate based host-guest system. Opt. Eng. 2014, 83, 087106. [CrossRef] 36. Gleißner, U.; Hanemann, T.; Megnin, C.; Wieland, F. The influence of photo initiators on refractive index and glass transition temperature of optically and rheologically adjusted acrylate based polymers. Polym. Adv. Technol. 2016, 27, 1294–1300. [CrossRef] 37. Gleißner, U.; Khatri, B.; Megnin, C.; Sherman, S.; Xiao, Y.; Hofmann, M.; Günther, A.; Rahlves, M.; Roth, B.; Zappe, H.; et al. Optically and rheologically tailored polymers for applications in integrated optics. Sens. Actuators A Phys. 2016, 241, 224–230. [CrossRef] 38. Andrade, E.D.C. A Theory of the viscosity of liquids—Part I. Lond. Edinb. Dub. Philos. Mag. J. Sci. 1934, 17, 497–511. [CrossRef] 39. Rebaioli, L.; Fassi, I. A review on benchmark artifacts for evaluating the geometrical performance of additive manufacturing processes. Int. J. Adv. Manuf. Technol. 2017, 93, 2571–2598. [CrossRef] 40. Moylan, S.; Slotwinski, J.; Cooke, A.; Jurrens, K.; Donmez, M.A. An Additive Manufacturing Test Artifact. J. Res. Natl. Inst. Stand. Technol. 2014, 119, 429–459. [CrossRef][PubMed]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).