Additive Manufacturing of Sub-Micron to Sub-Mm Metal Structures with Hollow AFM Cantilevers

micromachines

Article Additive Manufacturing of Sub-Micron to Sub-mm Metal Structures with Hollow AFM Cantilevers

Giorgio Ercolano 1,*, Cathelijn van Nisselroy 2, Thibaut Merle 1,János Vörös 2, Dmitry Momotenko 2, Wabe W. Koelmans 1 and Tomaso Zambelli 2,*

1 Exaddon AG, Sägereistrasse 25, 8152 Glattbrugg, Switzerland; [email protected] (T.M.); [email protected] (W.W.K.) 2 Laboratory of Biosensors and Bioelectronics, ETH Zürich, Gloriastrasse 35, 8092 Zurich, Switzerland; [email protected] (C.v.N.); [email protected] (J.V.); [email protected] (D.M.) * Correspondence: [email protected] (G.E.); [email protected] (T.Z.)

Received: 12 November 2019; Accepted: 12 December 2019; Published: 18 December 2019

Abstract: We describe our force-controlled 3D printing method for layer-by-layer additive micromanufacturing (µAM) of metal microstructures. Hollow atomic force microscopy cantilevers are utilized to locally dispense metal ions in a standard 3-electrode electrochemical cell, enabling a confined electroplating reaction. The deflection feedback signal enables the live monitoring of the voxel growth and the consequent automation of the printing protocol in a layer-by-layer fashion for the fabrication of arbitrary-shaped geometries. In a second step, we investigated the effect of the free parameters (aperture diameter, applied pressure, and applied plating potential) on the voxel size, which enabled us to tune the voxel dimensions on-the-fly, as well as to produce objects spanning at least two orders of magnitude in each direction. As a concrete example, we printed two different replicas of Michelangelo’s David. Copper was used as metal, but the process can in principle be extended to all metals that are macroscopically electroplated in a standard way.

Keywords: additive micromanufacturing; electrochemical microprinting; copper microcoils; FluidFM; David

1. Metal Additive Manufacturing at the Micro Scale At the macro- and mesoscale (down to 100 µm), additive manufacturing (AM) of metal objects is recurrently performed with robust methods like selective laser and electron beam melting, both relying on the local fusion of metal particles to obtain a solid metal with the desired shape [1–5]. Typical voxel sizes are reported in the range of 100 to 400 µm [6] As alternatives, one can take into consideration the well-established localized electrochemical deposition (LED) [7,8] and laser chemical vapor deposition (LCVD) [9] methods, whose minimum feature sizes are continuously improved down to 10 µm [10,11] and correspondingly interpreted [12,13]. Restricting the depiction only to the fabrication of metal objects at the micron scale (i.e., with details smaller than 10 µm, µAM), three main strategies are being established [14]: fabrication of templates by micro-stereolithography to be successively metallized either by coating (positive templates) or by electroplating (negative templates) [15–18], transfer of metallic materials to be eventually sintered in a second step, or in situ metal synthesis. As an example of the “transfer” strategy, we can refer to direct ink writing [19–21], electrohydrodynamic printing [22–24], laser-assisted electrophoretic deposition [25], laser-induced forward transfer [26], melt droplets [27]. On the other hand, focused ion/electron beam methods (metal precursor dissociation) [28–30], as well as laser-induced photoreduction [31–33], together with electrochemical reduction (meniscus-conﬁned) [34,35], local dispensing [36,37], electrohydrodynamic

Micromachines 2020, 11, 6; doi:10.3390/mi11010006 www.mdpi.com/journal/micromachines Micromachines 2020, 11, 6 2 of 14 redox printing [38], can be catalogued in the “in situ metal synthesis” ensemble. More recently, other methods have appeared like pyrolysis of metal-containing resins structured with two-photon lithography [39], as well as implosion fabrication [40]. Electrochemical additive manufacturing at the micro scale. Since in this contribution we are presenting results obtained via local electrochemical (ec) reduction, we would like to focus on the techniques of this category. Their common underlying idea is to fill a printing nozzle (in most cases a pipette) with a solution containing a metal salt, to bring it as close as possible to a biased conductive substrate (generally connected as working electrode, WE), and to eject the solution inducing a confined metal plating because of the vicinity of the nozzle with the working electrode. Based on this idea, four different protocols have been developed; meniscus-confined electroplating, scanning ion conductance electroplating, electrohydrodynamic redox printing, force-controlled electroplating. In the case of meniscus-confined electroplating (MCEP), which is performed in an atmosphere with controlled humidity, a glass pipette containing the metal ion solution and equipped with a counter electrode (CE) is moved toward the WE until a meniscus is formed between them. Inside the meniscus, the ec cell reduction takes place, forming a narrowed metal deposit that can be interpreted as a “voxel” [34]. In this way, metal wires were produced [41], eventually branching out in 3D structures [35]. The wires obtained show a diameter even smaller than the aperture diameter which is typically in the range between 100 and 250 nm. As the metals, copper and platinum were electroplated. Recent MCEP developments are focusing on implementing a feedback for the pipette approach by taking advantage of the ec current flowing once that the meniscus is established [42], as well as on the parallelization to enlarge the printed area by utilizing nozzle arrays [43,44]. For scanning ion conductance electroplating (SICM-EP), a double-barrel glass capillary is immersed in an ec cell, each barrel containing a CE. One barrel is filled with the copper salt solution, whereas the other one with a salt solution (i.e., same as in the bulk of electrolyte and without any metal ions). This second barrel without metal ions is responsible for the SICM feedback during the approach toward the WE, while the first one serves as source of metal ions to be reduced in the volume between pipette apex and WE [37]. High aspect-ratio structures, together with overhanging ones, could be manufactured, whose section was found to be 10–20 times larger than that of the aperture (30–50 nm). In a recent publication [45], Yoshioka and coauthors utilized a similar double-channel pipette configuration, whereby one channel was assigned to the SICM feedback for the apex-substrate separation, while the other was filled with a gold colloidal solution (3 nm) for electrophoretic deposition of the Au nanoparticles, obtaining high-aspect ratio micropillars with diameter in the sub-micron range (~800 nm diameter with a nozzle aperture of ~200 nm). Electrohydrodynamic redox printing (EHD-RP) also uses double-barrel capillary glasses, each barrel containing solvated metal ions generated via in situ ec dissolution of a metal electrode (copper and silver) [38]. Ion-loaded solvent droplets are then expelled by electrohydrodynamic forces and reduced to a metal deposit on the WE (~400 nm diameter with a nozzle aperture of ~200 nm). The biggest novelty of this method is inherent to the possibility of manufacturing multimaterial microstructures. Force-controlled electroplating (FCEP). At the core of the method lies the FluidFM technology [46,47], which combines atomic force microscopy ([48], AFM) and microfluidics by means of AFM cantilevers with an embedded microchannel and an aperture at their apex. When used for µAM, these hollow AFM probes are termed ion tips (Exaddon AG, Glattbrugg, Switzerland). Upon immersion in a standard three-electrode ec cell, the ion tip could be used as nozzle for local supply of precursor ions (copper), thus confining their ec reduction on the cathodically polarized WE [49]. This elicited a novel protocol for 3D µAM of metal structures [36]. By setting an initial separation of a few hundred nanometers between the probe aperture and the substrate, the locally plated metal can freely grow in the vertical direction, creating a voxel and thus giving access to printing in the third dimension. Moreover, the protocol could be automated by relying on the force-sensing capability: by detecting the moment when the growing voxel touches the probe apex, the deflection signal can be used as a trigger to define when the voxel deposition has ended. Subsequently, the probe is moved to the next coordinate where Micromachines 2020, 11, 6 3 of 14 the following voxel is planned to be deposited. Therefore, complex microstructures could be fabricated in a true layer-by-layer fashion (Figure1). MicromachinesAs metal, 2019 copper, 10, x is mainly chosen for ec printing because of its coulombic efficiency (minimum3 of 13 faradaic losses due to side reactions) beside its importance for electronic applications. In a recent comprehensivecomprehensive study,study, thethe mechanicallymechanically properties were compared for for two two benchmark benchmark structures structures producedproduced with with mostmost methodsmethods mentionedmentioned inin thisthis introductionintroduction [[50].50].

Figure 1. Automated force-controlled electrochemical (ec) 3D printing process schematized [36] the Figure 1. Automated force-controlled electrochemical (ec) 3D printing process schematized [36] the simple case of two pillars side-by-side: we emphasize the layer-by-layer fabrication (layers I-III are simple case of two pillars side-by-side: we emphasize the layer-by-layer fabrication (layers I-III are labeled), i.e., the pillars are printed in parallel and not in series. (a) The ion tip filled with CuSO4 labeled), i.e., the pillars are printed in parallel and not in series. a) The ion tip filled with CuSO4 solution is positioned over the first pillar at a set separation (e.g., 500 nm) where the metal voxel is solution is positioned over the first pillar at a set separation (e.g., 500 nm) where the metal voxel is to to be deposited. Local electroplating is switched on at a given overpressure, leading to local pillar be deposited. Local electroplating is switched on at a given overpressure, leading to local pillar growth (voxel III). (b) When the growing voxel touches the pyramidal apex, a cantilever deflection is growth (voxel III). b) When the growing voxel touches the pyramidal apex, a cantilever deflection is detected on the photodiode via the moving laser beam. The inset graph shows the temporal evolution detected on the photodiode via the moving laser beam. The inset graph shows the temporal evolution of the deflection signal (defl.) for a voxel touching event (yellow segment). (c) As soon as this touching of the deflection signal (defl.) for a voxel touching event (yellow segment). c) As soon as this touching event is recognized by the software, the probe is moved to the next position, i.e., on top of the second event is recognized by the software, the probe is moved to the next position, i.e., on top of the second pillar again with a typical separation of 500 nm, and the voxel ec deposition is started. Reprinted with pillar again with a typical separation of 500 nm, and the voxel ec deposition is started. Reprinted with permission from Ref. [36]. Copyright 2017 John Wiley & Sons Inc. permission from Ref. 36. Copyright 2017 John Wiley & Sons Inc. We conceived this contribution to describe the FCEP in detail revisiting our two main We conceived this contribution to describe the FCEP in detail revisiting our two main publications [36,51] but also presenting complementary unpublished results. publications [36,51] but also presenting complementary unpublished results. 2. The Force-Controlled Microprinting Tool 2. The Force-Controlled Microprinting Tool Microchanneled AFM probes. AFM cantilevers were produced with an embedded microchannel Microchanneled AFM probes. AFM cantilevers were produced with an embedded microchannel connecting the hollow pyramidal tip on one side and a macro reservoir on the other side (ion tips, connecting the hollow pyramidal tip on one side and a macro reservoir on the other side (ion tips, Exaddon AG). Such microchannels were obtained either by etching a sacrificial polysilicon layer Exaddon AG). Such microchannels were obtained either by etching a sacrificial polysilicon layer between two layers of Si3N4 according to the batch processes reported in [52,53]. The standard section between two layers of Si3N4 according to the batch processes reported in [52,53]. The standard section of the microchannel is of 20 µm 1 µm, whereas the pyramid is 10 µm 10 µm 7 µm with a circular of the microchannel is of 20 µm× × 1 µm, whereas the pyramid is 10 µm × 10 µm ×× 7 µm with a circular aperture at the apex either of 300 nm or of 500 nm diameter (Figure2). Alternatively, for apex apertures aperture at the apex either of 300 nm or of 500 nm diameter (Figure 2). Alternatively, for apex of different sizes, we opened so called “closed” probes (i.e., having no aperture at the apex) by means apertures of different sizes, we opened so called “closed” probes (i.e., having no aperture at the apex) of a focused ion beam (FIB) milling. In this case, the probes were mounted in a custom probe holder by means of a focused ion beam (FIB) milling. In this case, the probes were mounted in a custom andprobe coated holder with and an 18-nm-thickcoated with carbon an 18-nm-thick layer using carbon a CCU-010 layer Carbon using Coatera CCU-010 (Safematic Carbon GmbH, Coater Bad Ragaz,(Safematic Switzerland) GmbH, Bad before Ragaz, milling Switzerland) by a FIB-scanning before milling electron by microscopea FIB-scanning (SEM) electron Nvision microscope 40 device (Zeiss,(SEM) Oberkochen,Nvision 40 device Germany) (Zeiss, with Oberkochen, the SmartSEM Germany) software with (Zeiss). the SmartSEM The milled software face of the (Zeiss). pyramidal The probemilled was face aligned of the pyramidal in parallel probe with was the FIB-beamaligned in equippedparallel with with the a FIB-beam gallium ionequipped source. with Subsequently, a gallium theion probes source. were Subsequently, milled with the an accelerationprobes were voltage milled ofwith 30 kVan atacceleration milling currents voltage varying of 30 kV from at 10milling pA up tocurrents 80 pA, varying depending from on 10 the pA desired up to opening.80 pA, depend The activeing on milling the desired process opening. was followed The active in live milling SEM modeprocess and was was followed only terminated in live SEM when mode the and desired was only opening terminated size was when reached. the desired After the opening milling size process, was thereached. probes After were the glued milling onto process, a dedicated the probes printing were holder glued (Exaddon onto a dedicated AG). AFM printing approaches holder were (Exaddon always performedAG). AFM in approaches contact mode were in thealways standard performed optical beamin contact deflection mode method. in the standard optical beam deflection method.

Micromachines 2020, 11, 6 4 of 14 Micromachines 2019, 10, x 4 of 13

Figure 2. ScanningScanning electron microscope (SEM)(SEM) micrographs micrographs of of the the apertures apertures at at the the apex apex of of the the pyramid pyramid of ofthe the ion ion tips. tips. Top Top view view of a (aof) 300a (a nm) 300 and nm a ( band) 500 a nm(b) diameter500 nm diameter aperture, obtainedaperture, by obtained contact lithography.by contact lithography.Top view ofa Top (c) 100 view nm, of a a ( d(c)) 1 100µm nm, and a a ( de)) 21 µµmm sideand squarea (e) 2 aperture,µm side obtainedsquare aperture, by focused obtained ion beam by focused(FIB) milling ion ofbeam closed (FIB) pyramidal milling probes.of closed The pyramidal choice of theprobes. smallest The aperture choice dimensionof the smallest of 100 aperture nm was dimensionconditioned of to 100 avoid nm potential was conditioned complications to avoid in filling potential the probes complications with electrolyte in filling [54 the], whereas probes 2withµm waselectrolyte the maximum [54], whereas size possible 2 µm was due the to maximum the limit set size by possible the presence due to of the the limit apex set of theby the inner presence pyramid of the(discernable apex of the in (innere)). Reprinted pyramid with (discernable permission in e from). Reprinted Ref. [51]. with Copyright permission 2019 Johnfrom WileyRef. 51. & Copyright Sons Inc. 2019 John Wiley & Sons Inc. ec cell. The first version of the electrochemical cell consisted of gold 15-nm-thin films on glass withec 3 cell. nm TiThe adhesion first version layer of as the WE, electrochemical an Ag wire as cell quasi-reference consisted of gold electrode 15-nm-thin (RE), andfilms a on Pt glass wire withas CE 3 innm a Ti simple adhesion droplet layer of as H WE,2SO 4anat Ag pH wire 3 [36 as]. quasi-reference The second version electrode of the (RE), printing and a chamberPt wire as (i.e., CE inthe a three-electrodesimple droplet cell) of H was2SO redesigned4 at pH 3 [36]. in aThe modular second fashion version with of thethe partsprinting made chamber out of solid (i.e., Teflon the three- [51]. The working electrode is connected to the printing substrate of 12 mm 12 mm; furthermore, a graphite electrode cell) was redesigned in a modular fashion with the parts made× out of solid Teflon [51]. The workingcounter electrode electrode and is connected a Ag/AgCl to wire the printing are used subs as reference.trate of 12 As mm WE, × 12 100-nm-thick mm; furthermore, films of a sputtered graphite Cu on a 1.3 cm 1.3 cm silicon substrate with 13 nm of Ti adhesion layer were used with a graphite CE counter electrode× and a Ag/AgCl wire are used as reference. As WE, 100-nm-thick films of sputtered Cuand on a Aga 1.3/AgCl cm × wire 1.3 ascm RE. silicon substrate with 13 nm of Ti adhesion layer were used with a graphite CE andThe a supportingAg/AgCl wire electrolyte as RE. was a 0.5 M H2SO4 with the addition of HCl to a concentration of 0.5 mM.The Thesupporting plating solutionelectrolyte was was a 0.8a 0.5 M M CuSO H2SO4 in4 with 0.5 M the H 2additionSO4 solution. of HCl to a concentration of 0.5 mM. SetupThe plating. The prototypesolution was printer a 0.8 wasM CuSO obtained4 in 0.5 assembling M H2SO4 solution. an AFM head (Nanowizard I by JPK, Berlin,Setup Germany),. The prototype a pressure printer controller was obtained (MFCS-4C assembling by Fluigent, an Villejuif,AFM head France), (Nanowizard and a potentiostat I by JPK, Berlin,(PalmSens, Germany), Utrecht, a Thepressure Netherlands) controller to (MFCS-4C be synchronized by Fluigent, together Villejuif, [36]. Automation France), and was a potentiostat achieved by (PalmSens,a custom-made Utrecht, LabVIEW The Netherlands) program which to be controlled synchronized the probe together position [36]. andAutomation read the was deflection achieved signal by afrom custom-made the AFM via LabVIEW two data program acquisition which cards controlled (NI-USB the 6009 probe and position NI-USB and 6343, read National the deflection Instruments, signal fromEnnetbaden, the AFM Switzerland). via two data However,acquisition for cards sizes (NI-USB in the z 6009 direction and NI-USB exceeding 6343, the National piezo range Instruments, of 15 µm, Ennetbaden,further z increments Switzerland). could only However, be accomplished for sizes in with the thez direction AFM stepper exceeding motors, the whereby piezo range their of operation 15 µm, furtherintroduced z increments x-y offsets could which only had tobe beaccomplished accounted for with by printingthe AFM an stepper additional motors, pillar whereby close to their each operationstructure. introduced The tip end x–y of this offsets continually which had increasing to be acco referenceunted for pillar by hadprinting to be an found additional after each pillar stepper close tomovement each structure. by automatically The tip end performing of this continually a rough AFMincreasing scan inreference the pillar pillar area had at a zto value be found corresponding after each stepperto the height movement of the by current automatically pillar. The performing x-y error introduceda rough AFM was scan then in determined the pillar area and at successfully a z value correspondingcompensated, butto the at theheight cost of of the important current pillar. time delays. The x–y To error get ridintroduced of this deficiency, was then thedetermined printing and tool wassuccessfully completely compensated, redesigned but [51 at]. Here,the cost we of introduce important a time completely delays. redesigned To get rid of system this deficiency, as far as both the printing tool was completely redesigned [51]. Here, we introduce a completely redesigned system as far as both the hardware and the software are concerned. A controller (Exaddon AG) was conceived

Micromachines 2020, 11, 6 5 of 14

Micromachines 2019, 10, x 5 of 13 the hardware and the software are concerned. A controller (Exaddon AG) was conceived with a low with a low latency and a fast feedback loop in order to react to a touching event (Figure 3) within 1 latency and a fast feedback loop in order to react to a touching event (Figure3) within 1 ms. In this ms. In this way, possible clogging of the opening is minimized, thus increasing the probe lifetime way, possible clogging of the opening is minimized, thus increasing the probe lifetime while higher while higher printing rates are also permitted. On the other side, the controller buffer is able to printing rates are also permitted. On the other side, the controller buﬀer is able to contain the next 16 contain the next 16 voxel coordinates to avoid any delay by a lack of voxel coordinates in the voxel coordinates to avoid any delay by a lack of voxel coordinates in the controller for stable and controller for stable and long-term operation up to tens of hours. Further components of the hardware long-termof the new operation system are up two to tenshigh-resolution of hours. Furtheroptical systems components (top and of the bottom hardware view) offor the loading new systemof the arenozzle, two high-resolutionfor printer adjustment optical and systems calibration, (top and as bottomwell as view)for imaging for loading of the ofprinted the nozzle, structures. for printer The adjustmentmaximum andvolume calibration, of the printing as well as space for imaging is 200 mm of the × printed70 mm structures.× 60 mm, Theassured maximum by a three-axis volume of the printing space is 200 mm 70 mm 60 mm, assured by a three-axis positioning stage (Exaddon positioning stage (Exaddon AG)× consisti× ng of three linear motors with an x-y positioning accuracy AG)(full-range, consisting repeatability) of three linear of 250 motors nm, as with well anas with x-y positioning 5 nm positioning accuracy accuracy (full-range, in z at repeatability)a 0.1 nm sensor of 250resolution. nm, as well Theas corresponding with 5 nm positioning repositioning accuracy time infromz at a a voxel 0.1 nm to sensorthe following resolution. one Thetakes corresponding place in less repositioningthan 50 ms (for time a Δ fromz of 500 a voxel nm). to the following one takes place in less than 50 ms (for a ∆z of 500 nm).

FigureFigure 3.3. SchematicSchematic of thethe force-controlledforce-controlled ec µAMµAM system. On On the the left, left, a a system system computer computer sends sends commandscommands to to the the system system control control unit unit on theon right.the right. The The system system control control unit governsunit governs the printing the printing process usingprocess an using embedded an embedded controller. controller. The ion tipThe is ion mounted tip is mounted on the z-stage on the in z-stage the printing in the printing head and head is moved and insideis moved the inside ec deposition the ec deposition cell by the cell z-and by the x-y z- stages.and x-y Astages. microﬂuidics A microfluidics control control system system with 1 with mbar precision1 mbar precision regulates theregulates electrolyte the ﬂowelectrolyte through flow the cantileverthrough the aperture. cantilever Adapted aperture. with permissionAdapted with from Ref.permission [51]. Copyright from Ref. 2019 51. Copyright John Wiley 2019 & Sons John Inc. Wiley & Sons Inc.

3.3. TuningTuning thethe VoxelVoxel Size Size atat DiDifferentﬀerent Scales:Scales: fromfrom 0.50.5 toto 20 µµmm

TheThe ececµ µAMAM system system is is a a multiparameter multiparameter instrument, instrument, whereby whereby the the value value of theof the probe probe aperture apertureDeq , ofDeq the, of applied the applied pressure pressurep, of p the, of electrodepositionthe electrodeposition potential potentialE as E wellas well as as the the voxel voxel height height∆ zΔcanz can be preciselybe precisely adjusted. adjusted. InIn a a previous previous work work [51 [51],], we we investigated investigat theed etheﬀect effect of the of ion the tip ion aperture tip apertureDeq and Dpeqon and the pdiameter on the ddiameterof a pillar d printedof a pillar at printed ﬁxed potential at fixed (potentialE = 0.5 V(𝐸 vs.= − Ag-AgCl)0.5 V vs Ag-AgCl) and geometry and geometry (i.e., distance (i.e., distance between − twobetween consecutive two consecutive voxels ∆z ,voxels here referred Δz, here to referred as voxel to height) as voxel as height) in Figure as4 ina. Figure The framework 4a. The framework consists in measuringconsists in the measuring pillar diameter the pillard from diameter SEM images 𝑑 from (Figure SEM4a, images all the d(Figurevalues are4a, compiledall the d invalues Figure are4b) . andcompiled the vertical in Figure speed 4b)z andfrom the the vertical printer speed logged 𝑧 z-axis from the position. printer These logged two z-axis experimental position. These values two can . thenexperimental be combined values to calculatecan then be the combined volumetric to depositioncalculate the speed volumetricV: deposition speed 𝑉 : .𝑉 π𝑑 𝑧. (1) V = d2z (1) 4 as function of opening size d and p (Figure 4c). as function of opening size d and p (Figure4c).

Micromachines 2020, 11, 6 6 of 14 Micromachines 2019, 10, x 6 of 13

Figure 4. (a) SEM image of a field of pillars printed at different pressure and deposition potential values Figure 4. (a) SEM image of a field of pillars printed at different pressure and deposition potential with a 300 nm aperture ion tip, while a SEM image of a single pillar with overlaid the results of the values with a 300 nm aperture ion tip, while a SEM image of a single pillar with overlaid the results diameter d analysis is shown in the inset. (b) Pillar diameters d obtained as a function of the pressure of the diameter d analysis is shown in the inset. (b) Pillar diameters. d obtained as a function of the applied during printing and, (c) volumetric deposition speed z as a function of the pressure. Panels (b) pressure applied during printing and, (c) volumetric deposition speed 𝑧 as a function of the pressure. and (c) reprinted with permission from Ref. [51]. Copyright 2019 John Wiley & Sons Inc. Panels b) and c) reprinted with permission from Ref. 51. Copyright 2019 John Wiley & Sons Inc. In this work, the same protocol is used to investigate the effect of the deposition potential E. ArraysIn this work, of pillars the weresame printed protocol at is pressures used to varyinginvestigate from the 10 effect mbar of to the 210 deposition mbar with potential a pressure E step. of 5 mbar.Arrays Each of pillars array was were printed printed at aat di pressuresfferent deposition varying potential,from 10 mbar the potential to 210 mbar values with were a pressure changed fromstep of0.5 5 mbar. V to Each0.42 Varray with was a 20 printed mV di ffaterence a differen betweent deposition each array; potential, the deposition the potential potential values was were set − − againstchanged an from Ag/ AgCl−0.5 V pellet to −0.42 as REV with and a all 20 the mV given difference potential between values each are referredarray; the to deposition in this reference. potential was Figureset against5 shows an aAg/AgCl set of selected pellet SEMas RE images and all showing the given pillars potential printed values at selected are referred pressure to values.in this Inreference. agreement with [51], increasing p while keeping E constant results in an increase of the diameter d. A largerFigure flow 5 shows of ions a requires set of selected a larger SEM area images for the showing ions contained pillars inprinted it to be at completelyselected pressure consumed values. by theIn agreement reduction reactionwith [51], so increasing that the pillars p while grow keeping with a E wider constant cross-section. results in Byan contrast,increase anof the increase diameter in E (i.e.,d. A morelarger negative flow of ions deposition requires potential) a larger area has anforopposite the ions contained effect: pillars in it printed to be completely at the same consumed pressure butby the at increasingreduction Ereactionhave a so smaller that the cross-section. pillars grow with a wider cross-section. By contrast, an increase in E (i.e., more negative deposition potential) has an opposite effect: pillars printed at the same pressure but at increasing E have a smaller cross-section.

Micromachines 2020, 11, 6 7 of 14 Micromachines 2019, 10, x 7 of 13

Figure 5. SEM images of pillars deposited with a nozzle aperture of 300 nm with 10, 50, 100, 150 and 200Figure mbar 5. appliedSEM images pressure of pillars and deposited0.42, 0.44, with0.46, a nozzl0.48e aperture and 0.50 of 300 V deposition nm with 10, potential. 50, 100, All150 theand − − − − − images200 mbar are takenapplied at thepressure same magnificationand −0.42, −0.44, (the − scale0.46, bar−0.48 corresponds and −0.50 toV 2depositionµm). potential. All the images are taken at the same magnification (the scale bar corresponds to 2 µm). This result highlights that electrochemical printing is fully controlled by the electrodeposition potentialThis and result pressure, highlights while that it is electrochemical monitored by measuring printing is the fully retraction controlled speed by of the the electrodeposition nozzle, which is automatedpotential and with pressure, the force while feedback. it is monitored In principle, by the measuring arrangement the retraction of the ec cell speed resembles of the nozzle, the so-called which wall-jetis automated electrode with system, the force a toolfeedback. in hydrodynamic In principle, the voltammetry arrangement that of utilizesthe ec cell a jet resembles of solution the so- of electroactivecalled wall-jet species, electrode which system, is pushed a tool fromin hydrodynam a circular nozzleic voltammetry to hit the that working utilizes (collector) a jet of solution electrode of perpendicularlyelectroactive species, [55]. The which diff usionis pushed limited from current a circIulard is well-describednozzle to hit the in theworking wall-jet (collector) arrangement electrode and isperpendicularly approximated by[55]. analytical The diffusion expressions limited that current relate I itd is to well-described the geometrical in parameters the wall-jet of arrangement the system (nozzle,and is approximateda, and working by electrode analytical radius, expressionsR) and the that parameters relate it to of the fluidgeometrical flow: parameters of the system (nozzle, a, and working electrode radius, R) and the parameters of the fluid flow: 2/3 5/12 1/2 3/4 3/4 Id = 1.6 k n F c0 D /υ− /d− /R /V / (2) 𝐼 1.6 𝑘 𝑛 𝐹 𝑐 𝐷 𝜐 𝑑 𝑅 𝑉 (2) wherewherek ,k,n n, ,F F, ,c 0c,0,D D, ,ν νand andV Vspecify specify geometricalgeometrical proportionalityproportionality factor,factor, numbernumber ofof electronselectrons inin ecec reaction,reaction, Faraday Faraday constant, constant, concentration concentration of of electroactive electroactive species, species, their their di diffusffusionion coe coefficient,fficient, dynamic dynamic viscosityviscosity of of the the medium medium and and volumetric volumetric flow flow rate, rate, respectively. respectively. This This expression expression can can be be simplified simplified by by arrangingarranging all all the the constants constants into into a a proportionality proportionality factor factorα α(di (differentfferent for for each each voltage voltage value): value):

/ / I 𝐼𝛼𝑅= αR3/4 p 3𝑝/4 (3)(3) The deposition current could not be measured directly because the active Cu2+ reducing area is 2+ in theThe order deposition of tens currentof µm2, couldwhile notthe betotal measured working directly electrode because area is the cm active2, resulting Cu inreducing the deposition area is µ 2 2 incurrent the order being of tenscompletely of m , whilemasked the by total the working large background electrode area current is cm ,on resulting the working in the deposition electrode. currentHowever, being using completely the vertical masked growth by the speed large background(Figure 6a) along current with on theexperimentally working electrode. measured However, pillar usingdiameters the vertical (Figure growth 6b), it speedbecomes (Figure possible6a) alongto deriv withe further experimentally insights measuredof the printing pillar process. diameters For example, the vertical copper growth rate (using the given Cu molar mass and density, MCu and ρCu), can be estimated from the eq. 3 by a simple transformation:

Micromachines 2019, 10, x 8 of 13

𝑀 𝑉𝛼 𝑅/ 𝑝/ (4) 𝜋𝜌 This relationship is complicated by the fact that the radius of the growing pillar R, which collects metal ions upon electroplating, is in fact a function of the applied pressure R = R(p) (Figure 6b). For better fitting of the vertical growth data using the experimental values of pillar diameters Rexperiment(p), we introduced an empirical scaling coefficient that allows estimation of the radius of the area where ions are collected Rfit(p): Micromachines 2020, 11, 6 / 8 of 14 𝑅p e R𝑝 (5) The result of this estimation suggests that the area where the current is collected upon plating at (Figurehigher pressures6b), it becomes (p > 60 mbar) possible is somehow to derive larger further than insights the printed of the metal printing pillar. process. Volumetric For deposition example, therate vertical𝑉 also copperdeviates growth from ratethe theoretical (using the givenprediction Cu molar, highlighting mass and that density, at higherMCu andpressuresρCu), can some be estimatedportion of fromthe metal the Equation ions is electroplated (3) by a simple around transformation: and not directly on the pillar. These results are in good agreement with electron microscopy images (Figure 5) that depict circular «dunes» around the MCu 5/4 3/4 pillars, with sizes increasing at larger pressureV = α magnitudes.R− p (4) πρCu

Figure 6. (a) Vertical plating speeds (dots), (b) pillar diameters and, (c) volumetric deposition speeds Figure 6. (a) Vertical plating speeds (dots), (b) pillar diameters and, (c) volumetric deposition speeds as a function of the pressure applied during printing at different deposition potentials. Dots are the as a function of the pressure applied during printing at different deposition potentials. Dots are the experimental data, whereas lines the fitted data. Because of the semi-empirical fitting, the lines have a experimental data, whereas lines the fitted data. Because of the semi-empirical fitting, the lines have characteristic shape. a characteristic shape. This relationship is complicated by the fact that the radius of the growing pillar R, which collects metal ions upon electroplating, is in fact a function of the applied pressure R = R(p) (Figure6b). For better

fitting of the vertical growth data using the experimental values of pillar diameters Rexperiment(p), we introduced an empirical scaling coefficient that allows estimation of the radius of the area where ions are collected Rfit(p): p/1000 Rfit(p) = e Rexperiment(p) (5) The result of this estimation suggests that the area where the current is collected upon plating at higher pressures (p > 60 mbar) is somehow larger than the printed metal pillar. Volumetric deposition . rate V also deviates from the theoretical prediction, highlighting that at higher pressures some portion Micromachines 2020, 11, 6 9 of 14 of the metal ions is electroplated around and not directly on the pillar. These results are in good agreement with electron microscopy images (Figure5) that depict circular «dunes» around the pillars, Micromachines 2019, 10, x 9 of 13 with sizes increasing at larger pressure magnitudes.

4. Arbitrary-Shaped Arbitrary-Shaped 3D 3D Structures Structures in in a a Laye Layer-by-Layerr-by-Layer Fashion: Michelangelo’s Michelangelo’s David The confinedconfined electrodepositionelectrodeposition is is not not limited limited at strand at strand printing printing of pillars of pillars and coils. and Thecoils. same The system same systemcan also can be usingalso be to using produce to produce complex complex 3D object 3D by object taking by advantage taking advantage of the merging of the of merging voxels placed of voxels side placedby side. side To demonstrateby side. To demonstrate the capabilities the of capabilities this process, of athis 3D process, model of a Michelangelo’s 3D model of Michelangelo’s David (author Davidof the original(author of model the original ‘gabrielmda’, model distributed‘gabrielmda’, under distributed CC-0 license under on CC-0 blendswap.com) license on blendswap.com) was modified wasand slicedmodified using and the open-sourcesliced using 3D the modeling open-source software 3D Blender modeling 2.8 (Blendersoftware Foundation, Blender 2.8 Amsterdam, (Blender Foundation,the Netherlands). Amsterdam The voxels, the Netherlands). defining the structureThe voxels were defining routed the in structure a layer-by-layer were routed fashion in a using layer- a by-layercustom clustering fashion using algorithm. a custom The sameclustering model algo wasrithm. then reproducedThe same model at a 10,000:1 was then and areproduced 70,000:1 scale. at a 10,000:1The and two a replicas70,000:1 ofscale. Michelangelo’s David [56], at the two completely different scales shown in FigureThe7, representtwo replicas the currentof Michelangelo’s state-of-the-art David in confined [56], at the electrodeposition two completely 3D different printing. scales The 700 shownµm-tall in Figure1:10,000 7, replicarepresent is shown the current in Figure state-of-the-art7a–c; the structure in confined is defined electrodeposition by 130,000voxels 3D printing. and the The total 700 print µm- talltime 1:10000 required replica is 16 his using shown a pressure in Figure of 7a–c; 50 mbar the and structure a deposition is defined potential by 130000 of 0.46 voxels V. The and 100- theµm-tall, total − print1:70,000 time version required of the is 16 same h using model a pressure is shown of in 50 Figure mbar7a,b and inset a deposition (to allow apotential direct comparison of -0.46 V. betweenThe 100- µm-tall,the two objects)1:70000 andversion 7d; thisof the structure same model is defined is shown by 25,000 in Figure voxels 7a,b and inset the (to printing allow a time direct required comparison is 2 h betweenusing a pressure the two ofobjects) 15 mbar and and 7d; a potentialthis structure of 0.5 is V.defined by 25000 voxels and the printing time required is 2 h using a pressure of 15 mbar and− a potential of −0.5 V.

Figure 7. SEMSEM images images of of a a1:10,000 1:10,000 and and a a1:70000 1:70,000 replica replica of ofthe the David David (Michelangelo). (Michelangelo). (a), ((ab),) (andb) and (c) are(c) arethe the1:10000 1:10,000 scaled, scaled, 700 µm-tall 700 µm-tall replica replica imaged imaged from from different diﬀerent angles, angles, the insets the insets in (a) in and (a) (b), and and (b), theand panel the panel (d) are (d )the are 1:70,000 the 1:70,000 scaled, scaled, 100 µm-tall 100 µm-tall replicas. replicas.

The two different diﬀerent pressure and potential values were utilized to optimize the print time and resolution required: the the larger larger pressure pressure and lowe lowerr deposition deposition potential potential used used on on the the 700 700 µmµm structure allowed forfor thethe deﬁnition definition of of a geometrya geometry with with larger larger voxels voxels (~4 µ (~4m diameter, µm diameter, as from as Figure from 6Figureb), thereby 6b), therebyreducing reducing the total the voxel total number voxel andnumber print and time print required, time required, while keeping while the keeping resolution the resolution to an acceptable to an acceptablevalue for the value scale for of the object.scale of By the contrast, object. forBy thecontrast, smaller for 100 theµ msmaller replica 100 a higher µm replica resolution a higher was resolution was required to target smaller voxels (~1.6 µm diameter, as from Figure 6b), therefore lower pressure and higher overpotential were employed.

Micromachines 2020, 11, 6 10 of 14 required to target smaller voxels (~1.6 µm diameter, as from Figure6b), therefore lower pressure and higherMicromachines overpotential 2019, 10, x were employed. 10 of 13

5. Intertwined Coils with Different Different Section Sizes Herein we also also demonstrate demonstrate how how it it is is possible possible to to tune tune the the voxel voxel area area by by adjusting adjusting p on-the-flyp on-the-fly to tofabricate fabricate a structure a structure composed composed of of parts parts with with diff differenterent sections. sections. For For this this purpose, purpose, we we designed a structure ofof four four intertwined intertwined coils, coils, each each wire wire having having its own its own diameter. diameter. If fabricated If fabricated with a with single a nozzle,single suchnozzle, an such entangled an entangled object can object be produced can be produced only in a layer-by-layeronly in a layer-by-layer fashion, on fashion, the condition on the that condition each of thethat four each voxels of the offour each voxels layer of is each plated layer at a is di platfferented at value a different of pressure. value Aof probepressure. with A a probe 500 nm with aperture a 500 wasnm aperture operated was at 20, operated 40, 60, andat 20, 80 40, mbar 60, and in succession 80 mbar in (Figure succession4b) obtaining (Figure 4b) the obtaining structure presentedthe structure in Figurepresented8, whereby in Figure wires 8, whereby are presented, wires are artificially presented, colored artificially for better colored visualization. for better visualization.

Figure 8.8. SEMSEM imagesimages (colored) (colored) of of four four intertwined intertwined coils co printedils printed with with a 500 a nm 500 nozzle, nm nozzle, each coil each printed coil atprinted a diff erentat a different pressure pressure to highlight to highlight the flexibility the fl andexibility the ability and the to changeability to and change control and the control voxel sizethe withinvoxel size the within same structure the same during structure the printingduring the process; printing 3056 process; voxels were3056 printedvoxels were in 25 printed min to obtainin 25 min the 180to obtainµm tall the structures. 180 µm tall (a) structures. and (b) are ( twoa) and diff (erentb) are views two different of the same views object. of the Reprinted same object. with permission Reprinted fromwith permission Ref. [51]. Copyright from Ref. 2019 51. Co Johnpyright Wiley 2019 & Sons John Inc. Wiley & Sons Inc.

The mainmain didifficultyfficulty ofof growing growing multiple multiple threads threads in in parallel parallel each each at diat ffdifferenterent pressure pressure values values is the is timethe time response response required required to establish to establish a stable a stable pressure; pressure; specifically, specifically, whether whether or not or the not pressure the pressure and thusand thethus cloud the cloud of Cu of2+ Cuions2+ ions can changecan change quickly quickly enough enough when when the nozzlethe nozzle is moved is moved from from one one strand strand to the to nextthe next (which (which takes takes just ajust few a milliseconds). few milliseconds). As evincible As evincible from Figure from8 Figure, each printed8, each wireprinted has indeedwire has a diindeedfferent a different diameter, diameter, which is which homogeneous is homogeneous along all alon its lengthg all its (5,length 6, 7 and(5, 6, 8 7µ andm, respectively).8 µm, respectively). From FigureFrom Figure4b it would 4b it would be expected be expected that the that diameters the diamet at thoseers at particular those particular pressure pressure values are values 2, 3, 3.7are and,2, 3, 4.43.7 µand,m. Experimentally, 4.4 µm. Experimentally, the diameters the of diameters the strands of of thethe intertwinedstrands of helicesthe intertwined are consistently helices larger are thanconsistently those measured larger than for thethose individual measured pillar. for the This individual bigger voxel pillar. size This is probably bigger duevoxel to size the “additional”is probably platingdue to the taking “additional” place during plating the taking lateral place motions during from the alateral strand motions to the next.from a Such strand lateral to the motions next. Such are absentlateral motions in the printing are absent of pillars in the which printing occurs of pillars only inwhich the z occurs direction. only in the z direction.

6. Conclusions and Outlook In conclusion,conclusion, wewe havehave introducedintroduced a a new new one-step one-step method method to to do do additive additive micromanufacturing micromanufacturing of metallicof metallic structures structures using using microchanneled microchanneled AFM AFM probes. probes. The keyThe noveltykey novelty of this of approach this approach is that is the that voxel the growthvoxel growth can be can detected be detected in situ in by situ monitoring by monitoring the real-time the real-time deﬂection deflection of the hollowof the hollow cantilever. cantilever. This is This is essential to avoid clogging of the probe aperture and to automate the process. Furthermore, it removes the need to calibrate parameters such as growth speed. Finally, the fact that the manufacturing takes place in a supporting electrolyte allows for the reliable voxel-wise printing on arbitrary positions without the restrictions of methods relying on meniscus formation. Combining these advantages, an ec 3D metal printing is achieved in a true voxel-by-voxel layer-by-layer fashion,

Micromachines 2020, 11, 6 11 of 14 essential to avoid clogging of the probe aperture and to automate the process. Furthermore, it removes the need to calibrate parameters such as growth speed. Finally, the fact that the manufacturing takes place in a supporting electrolyte allows for the reliable voxel-wise printing on arbitrary positions without the restrictions of methods relying on meniscus formation. Combining these advantages, an ec 3D metal printing is achieved in a true voxel-by-voxel layer-by-layer fashion, enabling the fabrication of arbitrarily shaped geometries. This is a breakthrough in metal 3D printing on the micrometer scale, enabling a new range of printable shapes and structures. In a second step, the force-controlled ec µAM was completely redesigned to optimize the ec printing of copper by reducing the response and positioning times and increasing the motor range. We gained further understanding of the deposition process by investigating the pillar growth systematically varying the ec deposition potential, the applied pressure and the probe aperture at a fixed the voxel height (∆z = 500 nm). The influence of the applied pressure and the probe aperture on the lateral diameter of the electroplated copper voxels was also investigated and rationalized. In this way, the voxel area could be spanned over two orders of magnitude using the same probe aperture. This multi-scale capability was exploited to fabricate copper microstructures of two different replicas of Michelangelo’s David. Lastly, a 50 µm 50 µm 170 µm object consisting of four helical wires was obtained by × × changing the pressure on a voxel-by-voxel basis so that each printed wire had a different diameter, yet homogeneous along its entire length. A significant throughput increase is now imaginable with the possibility of tuning the voxel size depending on the feature dimensions within the same design. Copper is the most compliant metal for electrodeposition, nonetheless this protocol can be applied in a one-to-one way to other metals.

Author Contributions: Conceptualization, all authors; investigation, G.E., T.M., W.W.K., and C.v.N.; writing—original draft preparation, G.E., D.M., and T.Z.; writing—review and editing, all authors; supervision, G.E., W.W.K., and T.Z.; funding acquisition, T.Z., and D.M. All authors have read and agreed to the published version of the manuscript. Funding: This work was partially funded by the Swiss Innovation Agency Innosuisse (Project Nr. PNFM-NM 18511.1). D.M. acknowledges the Swiss National Science Foundation Ambizione Grant PZ00P2_174217/1. Acknowledgments: We thank Stephen Wheeler (LBB) for technical assistance, as well as Joakim Reuteler (ETH Zurich ScopeM facility) for his backing during the FIB-SEM sessions. We also acknowledge Cytosurge AG for their support. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Murr, L.E.; Gaytan, S.M.; Ramirez, D.A.; Martinez, E.; Hernandez, J.; Amato, K.N.; Shindo, P.W.; Medina, F.R.; Wicker, R.B. Metal Fabrication by Additive Manufacturing Using Laser and Electron Beam Melting Technologies. J. Mater. Sci. Technol. 2012, 28, 1–14. [CrossRef] 2. Frazier, W.E. Metal Additive Manufacturing: A Review. J. Mater. Eng. Perform. 2014, 23, 1917–1928. [CrossRef] 3. Herzog, D.; Seyda, V.; Wycisk, E.; Emmelmann, C. Additive manufacturing of metals. Acta Mater. 2016, 117, 371–392. [CrossRef] 4. Körner, C. Additive manufacturing of metallic components by selective electron beam melting—A review. Int. Mater. Rev. 2016, 61, 361–377. [CrossRef] 5. Buchanan, C.; Gardner, L. Metal 3D printing in construction: A review of methods, research, applications, opportunities and challenges. Eng. Struct. 2019, 180, 332–348. [CrossRef] 6. Galati, M.; Iuliano, L. A literature review of powder-based electron beam melting focusing on numerical simulations. Addit. Manuf. 2018, 19, 1–20. [CrossRef] 7. Jansson, A.; Thornell, G.; Johansson, S. High Resolution 3D Microstructures Made by Localized Electrodeposition of Nickel. J. Electrochem. Soc. 2000, 147, 1810–1817. [CrossRef] 8. Habib, M.A.; Rahman, M. Performance of electrodes fabricated by localized electrochemical deposition (LECD) in micro-EDM operation on diﬀerent workpiece materials. J. Manuf. Process. 2016, 24, 78–89. [CrossRef] Micromachines 2020, 11, 6 12 of 14

9. Duty, C.; Jean, D.; Lackey, W.J. Laser chemical vapour deposition: Materials, modelling, and process control. Int. Mater. Rev. 2001, 46, 271–287. [CrossRef] 10. Said, R.A. Localized electro-deposition (LED): The march toward process development. Nanotechnology 2004, 15, S649. [CrossRef] 11. Braun, T.M.; Schwartz, D.T. The Emerging Role of Electrodeposition in Additive Manufacturing. Electrochem. Soc. Interface 2016, 25, 69–73. [CrossRef] 12. Brant, A.; Sundaram, M. Molecular dynamics study of direct localized overpotential deposition for nanoscale electrochemical additive manufacturing process. Precis. Eng. 2019, 56, 412–421. [CrossRef] 13. Drahushuk, L.W.; Govind Rajan, A.; Strano, M.S. Fundamental scaling laws for the direct-write chemical vapor deposition of nanoscale features: Modeling mass transport around a translating nanonozzle. Nanoscale 2019, 11, 2925–2937. [CrossRef][PubMed] 14. Hirt, L.; Reiser, A.; Spolenak, R.; Zambelli, T. Additive Manufacturing of Metal Structures at the Micrometer Scale. Adv. Mater. 2017, 29, 1604211. [CrossRef][PubMed] 15. Fischer, J.; Wegener, M. Three-dimensional optical laser lithography beyond the diffraction limit: 3D optical lithography off limits. Laser Photonics Rev. 2013, 7, 22–44. [CrossRef] 16. Hohmann, J.K.; Renner, M.; Waller, E.H.; von Freymann, G. Three-Dimensional µ-Printing: An Enabling Technology. Adv. Opt. Mater. 2015, 3, 1488–1507. [CrossRef] 17. Sugioka, K.; Cheng, Y. Femtosecond laser three-dimensional micro- and nanofabrication. Appl. Phys. Rev. 2014, 1, 041303. [CrossRef] 18. Wang, D.; Wen, C.; Chang, Y.; Lin, W.; Chen, S.-C. Ultrafast laser-enabled 3D metal printing: A solution to fabricate arbitrary submicron metal structures. Precis. Eng. 2018, 52, 106–111. [CrossRef] 19. Ahn, B.Y.; Duoss, E.B.; Motala, M.J.; Guo, X.; Park, S.-I.; Xiong, Y.; Yoon, J.; Nuzzo, R.G.; Rogers, J.A.; Lewis, J.A. Omnidirectional printing of flexible, stretchable, and spanning silver microelectrodes. Science 2009, 323, 1590–1593. [CrossRef] 20. Zhou, N.; Liu, C.; Lewis, J.A.; Ham, D. Gigahertz Electromagnetic Structures via Direct Ink Writing for Radio-Frequency Oscillator and Transmitter Applications. Adv. Mater. 2017, 29, 1605198. [CrossRef] 21. Lee, S.; Kim, J.H.; Wajahat, M.; Jeong, H.; Chang, W.S.; Cho, S.H.; Kim, J.T.; Seol, S.K. Three-dimensional Printing of Silver Microarchitectures Using Newtonian Nanoparticle Inks. ACS Appl. Mater. Interfaces 2017, 9, 18918–18924. [CrossRef][PubMed] 22. Park, J.-U.; Hardy, M.; Kang, S.J.; Barton, K.; Adair, K.; Mukhopadhyay, D.K.; Lee, C.Y.; Strano, M.S.; Alleyne, A.G.; Georgiadis, J.G.; et al. High-resolution electrohydrodynamic jet printing. Nat. Mater. 2007, 6, 782–789. [CrossRef][PubMed] 23. Galliker, P.;Schneider, J.; Eghlidi, H.; Kress, S.; Sandoghdar, V.;Poulikakos, D. Direct printing of nanostructures by electrostatic autofocussing of ink nanodroplets. Nat. Commun. 2012, 3, 890. [CrossRef][PubMed] 24. Onses, M.S.; Sutanto, E.; Ferreira, P.M.; Alleyne, A.G.; Rogers, J.A. Mechanisms, Capabilities, and Applications of High-Resolution Electrohydrodynamic Jet Printing. Small 2015, 11, 4237–4266. [CrossRef] 25. Takai, T.; Nakao, H.; Iwata, F. Three-dimensional microfabrication using local electrophoresis deposition and a laser trapping technique. Opt. Express 2014, 22, 28109–28117. [CrossRef] 26. Wang, J.; Auyeung, R.C.Y.; Kim, H.; Charipar, N.A.; Piqué, A. Three-dimensional printing of interconnects by laser direct-write of silver nanopastes. Adv. Mater. 2010, 22, 4462–4466. [CrossRef] 27. Zenou, M.; Sáar, A.; Kotler, Z. Laser Transfer of Metals and Metal Alloys for Digital Microfabrication of 3D Objects. Small 2015, 11, 4082–4089. [CrossRef] 28. Utke, I.; Hoffmann, P.; Melngailis, J. Gas-assisted focused electron beam and ion beam processing and fabrication. J. Vac. Sci. Technol. B Microelectron. Nanom. Struct. Process. Meas. Phenom. 2008, 26, 1197–1276. [CrossRef] 29. Huth, M.; Porrati, F.; Dobrovolskiy, O.V. Focused electron beam induced deposition meets materials science. Microelectron. Eng. 2018, 185, 9–28. [CrossRef] 30. Winkler, R.; Lewis, B.B.; Fowlkes, J.D.; Rack, P.D.; Plank, H. High-Fidelity 3D-Nanoprinting via Focused Electron Beams: Growth Fundamentals. ACS Appl. Nano Mater. 2018, 1, 1014–1027. [CrossRef] 31. Hada, H.; Yonezawa, Y.; Yoshida, A.; Kurakake, A. Photoreduction of silver ion in aqueous and alcoholic solutions. J. Phys. Chem. 1976, 80, 2728–2731. [CrossRef] Micromachines 2020, 11, 6 13 of 14

32. Tanaka, T.; Ishikawa, A.; Kawata, S. Two-photon-induced reduction of metal ions for fabricating three-dimensional electrically conductive metallic microstructure. Appl. Phys. Lett. 2006, 88, 081107. [CrossRef] 33. Ishikawa, A.; Tanaka, T.; Kawata, S. Improvement in the reduction of silver ions in aqueous solution using two-photon sensitive dye. Appl. Phys. Lett. 2006, 89, 113102. [CrossRef] 34. Suryavanshi, A.P.; Yu, M.-F. Probe-based electrochemical fabrication of freestanding Cu nanowire array. Appl. Phys. Lett. 2006, 88, 083103. [CrossRef] 35. Seol, S.K.; Kim, D.; Lee, S.; Kim, J.H.; Chang, W.S.; Kim, J.T. Electrodeposition-based 3D Printing of Metallic Microarchitectures with Controlled Internal Structures. Small 2015, 11, 3896–3902. [CrossRef] 36. Hirt, L.; Ihle, S.; Pan, Z.; Dorwling-Carter, L.; Reiser, A.; Wheeler, J.M.; Spolenak, R.; Vörös, J.; Zambelli, T. Template-Free 3D Microprinting of Metals Using a Force-Controlled Nanopipette for Layer-by-Layer Electrodeposition. Adv. Mater. 2016, 28, 2311–2315. [CrossRef][PubMed] 37. Momotenko, D.; Page, A.; Adobes-Vidal, M.; Unwin, P.R. Write-Read 3D Patterning with a Dual-Channel Nanopipette. ACS Nano 2016, 10, 8871–8878. [CrossRef] 38. Reiser, A.; Lindén, M.; Rohner, P.; Marchand, A.; Galinski, H.; Sologubenko, A.S.; Wheeler, J.M.; Zenobi, R.; Poulikakos, D.; Spolenak, R. Multi-metal electrohydrodynamic redox 3D printing at the submicron scale. Nat. Commun. 2019, 10, 1853. [CrossRef] 39. Vyatskikh, A.; Delalande, S.; Kudo, A.; Zhang, X.; Portela, C.M.; Greer, J.R. Additive manufacturing of 3D nano-architected metals. Nat. Commun. 2018, 9, 593. [CrossRef] 40. Oran, D.; Rodriques, S.G.; Gao, R.; Asano, S.; Skylar-Scott, M.A.; Chen, F.; Tillberg, P.W.; Marblestone, A.H.; Boyden, E.S. 3D nanofabrication by volumetric deposition and controlled shrinkage of patterned scaffolds. Science 2018, 362, 1281–1285. [CrossRef] 41. Hu, J.; Yu, M.-F. Meniscus-confined three-dimensional electrodeposition for direct writing of wire bonds. Science 2010, 329, 313–316. [CrossRef][PubMed] 42. Aaronson, B.D.B.; Güell, A.G.; McKelvey, K.; Momotenko, D.; Unwin, P.R. Scanning Electrochemical Cell Microscopy. Mapping, Measuring, and Modifying Surfaces and Interfaces at the Nanoscale. In Nanoelectrochemistry; CRC Press: Boca Raton, FL, USA, 2015; pp. 655–693. 43. Lin, Y.-P.; Zhang, Y.; Yu, M.-F. Parallel Process 3D Metal Microprinting. Adv. Mater. Technol. 2019, 4, 1800393. [CrossRef] 44. Chen, M.; Xu, Z.; Kim, J.H.; Seol, S.K.; Kim, J.T. Meniscus-on-Demand Parallel 3D Nanoprinting. ACS Nano 2018, 12, 4172–4177. [CrossRef][PubMed] 45. Yoshioka, M.; Mizutani, Y.; Ushiki, T.; Nakazawa, K.; Iwata, F. Micropillar fabrication based on local electrophoretic deposition using a scanning ion conductance microscope with a theta nanopipette. Jpn. J. Appl. Phys. 2019, 58, 046503. [CrossRef] 46. Meister, A.; Gabi, M.; Behr, P.; Studer, P.; Vörös, J.; Niedermann, P.; Bitterli, J.; Polesel-Maris, J.; Liley, M.; Heinzelmann, H.; et al. FluidFM: Combining atomic force microscopy and nanofluidics in a universal liquid delivery system for single cell applications and beyond. Nano Lett. 2009, 9, 2501–2507. [CrossRef][PubMed] 47. Guillaume-Gentil, O.; Potthoff, E.; Ossola, D.; Franz, C.M.; Zambelli, T.; Vorholt, J.A. Force-controlled manipulation of single cells: From AFM to FluidFM. Trends Biotechnol. 2014, 32, 381–388. [CrossRef] 48. Binnig, G.; Quate, C.F.; Gerber, C. Atomic force microscope. Phys. Rev. Lett. 1986, 56, 930–933. [CrossRef] 49. Hirt, L.; Grüter, R.R.; Berthelot, T.; Cornut, R.; Vörös, J.; Zambelli, T. Local surface modification via confined electrochemical deposition with FluidFM. RSC Adv. 2015, 5, 84517–84522. [CrossRef] 50. Reiser, A.; Koch, L.; Dunn, K.A.; Matsuura, T.; Iwata, F.; Fogel, O.; Kotler, Z.; Charipar, K.; Piqué, A.; Rohner, P.; et al. Metals by micro-scale additive manufaturing: Comparison of microstructure and mechanical properties. arXiv 2019, arXiv:1912.03581. 51. Ercolano, G.; Zambelli, T.; van Nisselroy, C.; Momotenko, D.; Vörös, J.; Merle, T.; Koelmans, W. Multiscale Additive Manufacturing of Metal Microstructures. Adv. Eng. Mater. 2019.[CrossRef] 52. Deladi, S.; Tas, N.R.; Berenschot, J.W.; Krijnen, G.J.M.; de Boer, M.J.; de Boer, J.H.; Péter, M.; Elwenspoek, M.C. Micromachined fountain pen for atomic force microscope-based nanopatterning. Appl. Phys. Lett. 2004, 85, 5361–5363. [CrossRef] 53. Berenschot, E.J.W.; Burouni, N.; Schurink, B.; van Honschoten, J.W.; Sanders, R.G.P.; Truckenmuller, R.; Jansen, H.V.; Elwenspoek, M.C.; van Apeldoorn, A.A.; Tas, N.R. 3D nanofabrication of fluidic components by corner lithography. Small 2012, 8, 3823–3831. [CrossRef][PubMed] Micromachines 2020, 11, 6 14 of 14

54. Dorwling-Carter, L.; Aramesh, M.; Forró, C.; Tiefenauer, R.F.; Shorubalko, I.; Vörös, J.; Zambelli, T. Simultaneous scanning ion conductance and atomic force microscopy with a nanopore: Effect of the aperture edge on the ion current images. J. Appl. Phys. 2018, 124, 174902. [CrossRef] 55. Yamada, J.; Matsuda, H. Limiting diffusion currents in hydrodynamic voltammetry: III. Wall jet electrodes. J. Electroanal. Chem. Interfacial Electrochem. 1973, 44, 189–198. [CrossRef] 56. Galleria Dell’Accademia Di Firenze (Italy). Available online: http://www.galleriaaccademiafirenze. beniculturali.it (accessed on 12 November 2019).

© 2019 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/).